FOOD AND DRUG ADMINISTRATION









                              Volume I





                          Tuesday, May 3, 2005


                               8:30 a.m.



                CDER Advisory Committee Conference Room

                           5630 Fishers Lane

                          Rockville, Maryland



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


      Charles Cooney, Ph.D., Chair


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


      Committee Members:


      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.

      Jurgen Venitz, M.D., Ph.D.


      Special Government Employees:


      Carol Gloff, Ph.D.

      Arthur H. Kibbe, Ph.D.

      Thomas P. Layloff, Jr., Ph.D.

      Marvin C. Meyer, Ph.D.


      FDA Participants:


      Gary Buchler, R.Ph.

      Lucinda Buhse, Ph.D.

      Ajaz Hussain, Ph.D.

      Mehul Mehta, Ph.D.

      Vibhakar Shah, Ph.D.

      Helen Winkle

      Lawrence Yu, Ph.D.



                            C O N T E N T S



      Call to Order

                 Charles Cooney, Ph.D., Chair                    5


      Conflict of Interest Statement

                Hilda Scharen, M.S., Executive Secretary         5


      Introduction to Meeting--OPS Update

                Helen Winkle                                     7


      Opening Remarks

                Charles Cooney, Ph.D.                           16


      Establishing Drug Release or Dissolution



         Topic Introduction

                Ajaz Hussain, Ph.D.                             18


         Dissolution Measurement System: Current State

           and Opportunities for Improvement

                Lucinda Buhse, Ph.D.                            45


         Questions by Committee Members                         76


         Overview of Guidance Documents

           and Decision Process:


         Biopharmaceutics Section

                Mehul Mehta, Ph.D.                              95


         Questions by Committee Members                        128


         Establishing Dissolution Specifications:

           Current Practice

                Vibhakar Shah, Ph.D.                           138


         Questions by Committee Members                        156


      Open Public Hearing:

                Will Brown, USP                                162


      Questions by Committee Members                           171



                      C O N T E N T S (Continued)



      Establishing Drug Release or Dissolution

      Specifications: (Continued)


         Factors Impacting Drug Dissolution and

           Absorption:  Current State of Science

                Lawrence Yu, Ph.D.                             179


         Questions by Committee Members                        198


         Summary of Tactical Plan

                Ajaz Hussain, Ph.D.                            208


         Committee Discussion and Recommendations              229


      Clinical Pharmacology Subcommittee Report

         (via teleconference)

                Jurgen Venitz, M.D., Ph.D.                     284


      Questions by Committee Members                           301



                         P R O C E E D I N G S


                             Call to Order


                DR. COONEY:  I would like to welcome


      everyone to this morning's meeting.  We have an


      opportunity for an on-time start.  I am Charles


      Cooney, the new chair of this committee.  I am


      delighted to welcome everyone here, both the


      committee members as well as the guests.  We have,


      not surprisingly, a full agenda this morning and we


      will begin with addressing the conflict of




                     Conflict of Interest Statement


                MS. SCHAREN:  Good morning.  The Food and


      Drug Administration has prepared general matters


      waivers for the following special government


      employees, Charles Cooney, Patrick DeLuca, Carol


      Gloff, Arthur Kibbe, Michael Korczynski, Thomas


      Layloff, Marvin Meyer, Kenneth Morris, Nozer


      Singpurwalla and Jurgen Venitz who are


      participating in today's meeting of the


      Pharmaceutical Science Advisory Committee to, one,


      receive an update from the Clinical Pharmacology



      Subcommittee and, two, discuss and provide comments


      on the general topic of establishing drug release


      or dissolution specifications.


                This 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 broad 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


      representatives, 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 one


      particular company.  Dr. Fackler is employed by


      Teva Pharmaceuticals and Dr. Migliaccio is employed


      by Pfizer.


                In the event that the discussions involve


      any other products or firms, not already on the


      agenda, for which FDA participants have a financial


      interest, the participant's involvement and


      exclusion will be noted for the record.  With


      respect to all other participants, we ask in in the


      interest of fairness that they address any current


      or previous financial involvement with any firm


      whose product they may wish to comment upon.  Thank




                DR. COONEY:  Thank you.  Now Helen Winkle


      will provide an update.


                  Introduction to Meeting--OPS Update


                MS. WINKLE:  Good morning, everyone.  I


      would like to welcome all the members of the



      advisory committee and to especially welcome Dr.


      Charles Cooney as our new chair of the advisory


      committee.  We, at FDA, have worked with Dr. Cooney


      as a member of the committee and have really felt


      that he has provided a lot of input into the


      committee's activities, and feel that working with


      him in the next two years as chair is going to be a


      very important step for all of us.


                Before I talk about the agenda for this


      session of the advisory committee, I would like to


      talk a little bit about our current focus at the


      agency or what we are calling a paradigm shift.  I


      think it is important for all of us to understand


      clearly the changes that we are making in the


      agency and the role of the advisory committee in


      assisting us in making these changes.  Based on


      recent initiatives in FDA, including the


      Pharmaceutical cGMP Initiative for the 21st


      Century, the PAT Initiative and the Critical Path


      Initiative, you can see the shift in FDA's thinking


      about regulating product quality.


                Specifically, there is a focus in these



      initiatives to place more responsibility on


      industry to ensure the quality of their


      pharmaceutical products rather than rely solely on


      regulatory scrutiny to maintain that quality.  This


      is really the paradigm shift, a sharing of


      responsibility for drug quality with emphasis


      placed on industry to understand their processes


      and the underlying science of those processes.


                Why would we want to make that change?


      There is no evidence that the products out there on


      the market are bad products.  There is no evidence


      that the agency has done a bad job in serving as a


      surrogate for ensuring good quality products for


      the consumer.  And, there is no evidence that


      industry is not focused on quality as an important


      attribute to manufacturing products.  However,


      times are changing.  As we enter the 21st century


      we have an excellent opportunity to begin to


      prepare for how we will handle pharmaceutical


      regulation in the future.  The time is ripe for us


      at FDA to invest in that future and to ensure that


      the direction we are going in is adequate to handle



      the changing world of pharmaceutical development


      and manufacturing while we continue to be able to


      serve the consumer.  It is the right time too to


      ensure that our regulatory involvement does not


      hinder the innovation and continuous improvement in


      manufacturing and ensuring the quality of


      pharmaceutical products.


                So, FDA has begun a journey towards this


      paradigm shift.  I want to say it is a long


      journey.  It started several years ago but we have


      a long way to go, and we have numerous challenges


      along that way.  However, with these challenges


      come opportunities and I think this is the


      important thing for us and the advisory committee


      to remember, that we need to take advantage of


      these opportunities.  It is important not only to


      take advantage of the opportunities to help us


      improve on how we regulate product quality, but


      also to ensure that we provide for modernization


      both at FDA and within industry for the 21st




                The guiding principles of the



      Pharmaceutical cGMP for the 21st Century, which


      include risk-based orientation, science-based


      policies and standards, integrated quality systems


      orientation, international cooperation and strong


      public health protection, serve to help us in


      developing the pathway to restructure the oversight


      of the pharmaceutical quality.  As each of you


      knows, there are a number of forks in that path and


      you, as members of this advisory committee, are


      really here to help us determine the right path in


      the road to go from a scientific perspective, and


      to help and advise us on how to fill the gaps which


      exist in the FDA.  These include gaps in


      organization, gaps in science and gaps in policy.


                The committee has already participated in


      discussions on a number of scientific issues which


      have helped in formulating a strategy for


      addressing many of the questions that have emerged


      as a result of this paradigm shift.  We have


      already discussed a number of issues which have


      significance as we develop our future regulatory


      paradigm, including such issues as polymorphism,



      bio-inequivalence of generic products, and we have


      worked together to support such initiatives as the


      Process Analytical Technology Initiative.  The


      committee has also been extremely helpful in moving


      toward this new paradigm with other discussions


      that we have had on various topics.  However,


      again, the journey has only just begun.


                The agenda for the next two days was


      developed to provide an opportunity to discuss two


      other scientific topics which are important to us


      to better understand and manage in order to move us


      steadily along the path of change.  The first topic


      is on establishing drug release or dissolution


      specifications.  Obviously, how we set


      specifications is important to the future.  As we


      move to the desired state of pharmaceutical quality


      we want to ensure that specifications are based on


      mechanistic understanding of how product and


      process factors impact product performance.  We are


      currently in the process of developing a tactical


      plan for setting dissolution specifications.  As


      you will hear from the presentations today, we have



      developed the fundamentals for this plan which


      include a systems view of setting specifications,


      ensuring that all factors which affect dissolution


      are considered; basically ensuring that we connect


      all the dots in CMC to ensure a more comprehensive


      and systematic way of setting specifications.


                FDA recently held a specifications


      workshop in co-sponsorship with the Product Quality


      Research Institute.  The workshop indicated a need


      for additional efforts to move toward better


      setting of specifications in general.  Some of the


      specific points that were brought out at the


      workshop included a lack of globalization on how


      specifications are set; a need to better define


      what we should do versus what we can do; a need to


      better define the role of the compendia in the new


      paradigm; and a need to revisit the decision trees


      in ICH Q6A.


                Our discussion at the advisory committee


      today is not designed to address all the workshop


      issues and concerns on setting specifications.  The


      discussion today will, however, help us finalize



      the tactical plan for setting dissolution


      specification and will lay the foundation for our


      thinking in setting specifications for CMC and


      addressing the specific issues that were identified


      at the workshop.


                We would appreciate the committee's


      comments and suggestions as to what data is needed


      to support our plans.  This would include looking


      at statistical methodology, etc., and how we might


      improve on our thinking in our tactical plan and


      specifications setting in general.


                At this meeting we will also discuss, as


      our second big topic, quality by design and


      pharmaceutical equivalence.  As you will remember,


      at the last meeting we set the stage in our


      discussions on bioinequivalence and bioequivalence


      testing of locally acting GI drugs.  At this


      meeting our goal is to modernize our general


      thinking about pharmaceutical equivalence and to


      explore how quality by design can be leveraged to


      ensure more rational approaches to decision-making


      so that we can move from a reactive environment to



      a proactive regulatory scheme of assessing




                We will discuss a number of relevant


      topics, including biopharm. classification system;


      using product development information to address


      highly variable drugs; and we will revisit the


      concept of decision trees for ensuring a rational


      approach to determining bioequivalence for topical


      drug products.  We look forward to the committee's


      feedback on these extremely important topics, and


      that discussion is for tomorrow.


                There are a number of other topics we will


      cover at this meeting, including an update from the


      working group on parametric tolerance interval test


      for dose content uniformity.  Bob O'Neill will give


      that update.  And an update from the Clinical


      Pharmacology Subcommittee.  We will also discuss


      with the committee our perceived need to establish


      a working group for the review and assessment of


      OPS' research programs.  Our goal in looking at our


      research programs is to ensure a common approach to


      all laboratory work and to ensure that our research



      aligns with our overall mission.


                Now that we have two laboratories in OPS,


      a biotech. laboratory and a lab focused on small


      molecules, it is extremely important that this


      alignment takes place and we really look forward to


      your input on how we can better align these two




                As you can see, we have a full agenda but


      I think the topics to be discussed are really of


      great interest to us as we move down the path to


      the desired state for pharmaceutical quality, and I


      look forward to a very interesting discussion on


      each of these topics.  Thank you.


                            Opening Remarks


                DR. COONEY:  Thank you, Helen.  I would


      like to just add a couple of comments, if I may, to


      get us started.  I am delighted to have the


      opportunity to work with the FDA and to work with


      this committee during the coming two years.  It is


      a particularly exciting time because as we look


      forward, as Helen has indicated, there are very


      important new initiatives on the table with the



      cGMP Initiative and the Critical Path Initiative


      and these, indeed, lay a foundation that we all


      need to work within.  In fact, it is an exciting


      opportunity to work within those initiatives to


      look at how we can better address some of the


      challenges going forward.


                Certainly, as we look forward there are


      more challenges than there have been in the past.


      We are facing a world of increasing molecular


      complexity; a world of increasing demands by


      consumers; a world in which we have increasing


      complexity not just in the molecules but in the


      delivery formats of these products and the role of


      this committee is very important in helping to


      provide advice to the FDA and to the division to


      deal with these problems.  I must say, I applaud


      the forward-looking and the proactive stance that


      is being taken on these issues.


                We have some challenging goals today and


      tomorrow, not the least of which, of course, is to


      stay on time.  But the reason that the challenge of


      staying on time is a challenge is because of the



      very high content of material that we need to deal


      with.  So, I will do my best to try and keep us


      within the proper boundaries.


                Again, I look forward to working with


      everyone.  This will be a very interesting two days


      and I see it as an important step in what will be a


      continuing series of discussions and activities and


      recommendations that we will need to work on with


      the FDA.  With that, the first presentation and


      discussion this morning will be by Ajaz Hussain,


      and we will begin by digging in to establishing


      drug release and dissolution specifications.  Ajaz,




                Establishing Drug Release or Dissolution


                   Specifications Topic Introduction


                DR. HUSSAIN:  Thank you, Dr. Cooney.  I


      think topic one is entitled quality by design


      approach for quality control and assurance of


      dissolution rate.  In the background packet, as


      well as in the presentations, I have tried to keep


      the terminology dissolution rate all along to


      illustrate the one challenge which we will not be



      discussing today, and that is the metrics for


      dissolution rate itself.  We express dissolution


      rate as a Q factor which tends to be confounded


      with the variability of the assay itself.  So, that


      is not the topic for discussion today but I just


      wanted to alert you on why the word "rate" keeps


      coming back and back again.  So.


                Topic one:  Our goal is to seek your


      recommendations and endorsement of the proposed


      regulatory tactical plan.  With this tactical plan


      we hope to start moving towards putting together a


      set of regulatory tools and policies that will help


      us define elements and details of the elements


      necessary to realize the goals of quality by




                The question that we are posing to you is


      are the tactical steps outlined consistent with the


      goals that we have shared with you?  What initial


      steps and/or changes would you recommend to improve


      this plan?  What additional scientific evidence do


      you feel would be necessary to support the


      development and implementation of this plan? 



      General considerations for identifying and


      developing statistical procedures and, in


      particular, I want to emphasize for this discussion


      today that we have left out a discussion on


      statistical procedures because our experience has


      been that if we start with that as a topic it leads


      to protracted debate, and you will hear one report


      on that debate from Bob O'Neill tomorrow, the


      debate on parametric tolerance interval that has


      been going on for years and hasn't come to any


      resolution.  We feel that if you approach it from


      scientific foundations first, statistics is simply


      a tool to implement the scientific decision


      framework.  So, that is the reason we have kept


      that out of discussion for today.  Clearly, we are


      seeking your specific recommendations and other


      recommendations that you may have including how we


      should prioritize our work to develop this tactical


      plan to a full proposal, which we hope to bring to


      you at a subsequent meeting.


                What are the proposed steps?  The proposed


      steps are to develop an alternate regulatory



      approach to demonstrate the suitability of


      dissolution measurements system; introduce and


      utilize the concept of reproducibility and


      repeatability study using the actual pharmaceutical


      product for which we set specification.  Here, the


      proposal is to consider using the pivotal clinical


      lot or the bio. lot as a basis for identifying how


      sensitive, or lack of it, it is to a dissolution


      test method and estimate the variability in the


      method and, therefore, of the product.


                So, the first two steps are sort of


      together.  But the next three steps are also sort


      of in one clump.  We want to move towards a


      system-based decision tree for establishing


      dissolution rate specification.  Within that


      framework I think we would like to utilize


      opportunities to utilize the PAT approach for


      controlling dissolution rate and development of


      real-time quality assurance strategies.  Also, a


      decision tree for design-based concepts articulated


      in the draft ICH Q8 guideline, which is in your


      background packet.


                So, those are the decision trees which we


      would like to develop.  At the same time, when we


      come back with the full proposal to you, we would



      like to bring to you a side-by-side comparison of


      new and generic drugs because we think this is an


      opportunity for both sides, and some of the


      frustrations the generic industry feels can be


      addressed with this proposal, and I will explain


      that towards the end of the day, and explain why


      the level of quality assurance or quality control


      confidence in the proposed approach will be higher


      than what is achieved in the current system.  There


      is no doubt in my mind but, clearly, you have to


      agree with that.


                We also seek today your recommendations on


      how we should approach the statistical aspect of


      this and then what will help you to discuss this


      proposal when it comes to you.  So please give us


      your recommendations on how we should prepare a


      detailed proposal for a subsequent ACPS meeting.


                The other step that I think is important


      and is very timely, because this Friday or this



      Saturday I leave for Brussels for our ICH meeting,


      is that we do intend to seek the discussion here,


      and utilize the discussion here, to seek


      harmonization of the approach we are proposing


      under the ICH, especially the ICH Q8 Part 2, and we


      will start developing that guideline in Brussels


      next week.


                ICH Q8 draft guideline essentially brought


      a basis for getting and considering pharmaceutical


      development information in a structured way for


      pharmaceutical decision-making in the CMC arena at


      FDA.  The guideline was constructed with this


      figure, on your right, in mind.  You have to focus


      your design efforts on the intended use of the


      product, the patient population and so forth, that


      leads to a product design and that product design


      dictates the design specification, which are


      customer requirements, and these requirements, some


      of them if they are critical, become regulatory


      specifications.  Then the product design and design


      specification dictates or leads to design of a


      manufacturing process to reliably and predictably



      deliver those specifications back to deliver the


      intended use.


                In a systematic way, if you approach


      pharmaceutical development in a structured way, you


      get some benefit, we believe.  You achieve a higher


      degree of process understanding and give regulators


      high confidence of low risk of releasing a poor


      quality product; high efficiency through continuous


      learning and improvement.  And, I think it helps us


      to address some of the gaps we have in the current


      system.  I have tried to illustrate the current


      process within FDA and the manufacturing and R&D


      process within industry.  Research and development


      will develop the products, transfer them to


      manufacturing and then we have, by law, a separate


      quality unit to maintain quality assurance.  You


      have all the specification results and you have


      products which don't have all the specification




                In approaching and assessing the quality


      we bring a team approach, a multi-disciplinary team


      approach which includes pharmacology, toxicology,



      CMC review, clin. pharm., bio. pharm. review and


      the clinical assessment.  And the decision


      collectively is a risk-benefit decision that leads


      to an approval of a product.  The approved product


      is then transferred--it is called technology


      transfer--and the process is validated.


                The validation process includes


      qualification criteria and so forth, but there is


      an element of that which is process qualification.


      Process qualification is essentially, in my


      opinion, the interaction between materials and


      equipment and environment that you really have to


      study.  In the current state that essentially is


      judged on your ability to repeat it three times.


      Since the pharmaceutical development information is


      not available for CMC reviewers, the quality of


      that and the understanding containing that is not


      well understood in the regulatory sense.  So, we


      are losing an opportunity to make more rational




                Now, we have a divide, an organizational


      divide within the agency between, say, review and



      GMP inspection.  The cGMP process is helping


      bridge.  The PAT is an example of how we have


      bridged it.  Our experience or learning from the


      initiative clearly identified a need for a quality


      system orientation.  I would be wrong if I said


      that I really did not understand what this really


      meant or really didn't care about what quality


      system issues were because I was looking from


      several years ago.  But I think I have gained a


      much deeper understanding of the importance of the


      quality system orientation.


                So, here is a representation of that from


      a paper that we published on innovation and


      continuous improvement in pharmaceutical


      manufacturing.  Say what you do, do what you say,


      prove it and improve it are the elements that make


      up a quality system.  Consider the way what you do


      is your application to FDA.  So, that is a CMC


      review assessment process.  Now, do what you say


      can be considered as are you able to manufacture to


      the commitments that you have placed in your


      application?  Now, there is a gap since our



      reviewers don't have an idea that they really do


      what they say.  That is a GMP function so there is


      uncertainty there.  And prove it.  How do you prove


      it?  I think metrics for proving it could be


      process capability and the recalls, and this and


      that.  If you are unable to prove it you need to


      have a collective action and a preventive action.


                Our experience has suggested that in most


      cases root cause is unknown or a poor analyst is


      blamed.  So, we actually don't get to a root cause


      generally.  Does the current system support or


      facilitate getting to the root cause?  I think that


      is the question.  In many ways I think say what you


      do and do what you say, if you take that ratio is


      process understanding and your ability to prove.


      So, in many ways I think you have to think about




                Now, a modern quality system has a


      dimension of improvement, continuous improvement


      and innovation.  The dotted line simply says that


      is an option that should be available for industry


      to do.  It is dotted because that is not a



      requirement per se, but the rest are all


      requirements.  So, I think we are trying to address


      some of these gaps along the way.


                Now, the definition of continuous


      improvement is interesting and it really sets the


      stage for this discussion.  I have taken the


      definition from QS-9000 to illustrate the challenge


      we face for continuous improvement.  For those


      product characteristics and process parameters that


      can be validated using variable data, that is


      continuous data, continuous improvement means


      optimizing the characteristics and parameters at a


      target value and reducing variation around the


      target value.  So, in a sense, you need a target


      value and you need to have an estimate of variation


      to start thinking of continuous improvement.  In


      our specification setting often we don't even have


      a target value.  So.  And forget variation.  So.


                But the second bullet is more important.


      For those product characteristics and process


      parameters that can be only evaluated using


      attribute data, pass/fail, continuous improvement



      is not possible until characteristics are


      conforming.  If attribute data results do not equal


      zero defect it is by definition a non-conforming


      product.  Improvements made in these situations are


      by definition corrective actions, not continuous


      improvement.  And, we have clearly distinguished


      between corrective action, which is a risky


      scenario, and continuous improvement, which can be


      managed differently.


                Continuous improvement in processes that


      have demonstrated stability, acceptability,


      capability and performance--continuous improvement


      really is only possible for those products that


      have demonstrated stability.  Process validation


      today does not give us the assurance that the


      process is stable.  So, that is another element.


      Acceptable capability, we don't have an estimate of


      the capability value.


                Now, the reason for finding this out is


      that I think we don't use compendial methods as


      release specifications.  Actually, the compendium


      approach to specifications is right.  That is the



      way they should be.  There is nothing wrong with


      this specification criteria for the market


      standard.  It is perfectly all right.  But it is,


      as Janet Woodcock says in her paper, different from


      release specification and that is the


      distinguishing feature that I think is the problem


      here.  If you use market standard as release


      specification, then you have all the elements that


      hold back continuous improvement.  So, you really


      need to distinguish between standards and


      specifications.  Unfortunately, in the current


      paradigm specifications equals standard.  So, what


      we are moving towards is a control strategy that


      will allow you to have your market standard but


      then have a control philosophy that allows you a


      risk-based decision process.


                A recent proposal from USP I think is a


      step in the right direction.  It is essentially a


      similar proposal to the parametric tolerance


      interval test to take dissolution specification


      criteria towards more of a tolerance interval


      approach.  But as you will hear tomorrow from the



      parametric tolerance interval discussion, you


      cannot approach it as hypothesis testing for every


      product batch, and that is one of the discussions


      that we will have.  And, there are many challenges


      before we even can get to that, and that is a part


      of this discussion.  We believe one has to start


      with a pharmaceutical science discussion before


      developing appropriate statistical tools.


                One other challenge for continuous


      improvement is the mind set--and this is a major


      challenge not only within the U.S. but


      globally--that corrective actions is the only way


      to force improvement of quality on industry.  This


      is direct current paste from the paper that we


      issued.  Some would argue that corrective actions


      provide the necessary constancy of purpose for


      improvement, necessary since manufacturing is a


      stepchild of industry because the difference


      between cost of manufacturing and price of drugs is


      large.  Keeping the system of corrective action


      provides the leverage for ensuring improvement, to


      ensure the cGMP.


                That is a fundamental challenge.  How do


      we achieve that?  If you improve your manufacturing


      process by reducing variability your regulatory



      acceptance criteria will be narrow.  So, that takes


      things into a way for continuous improvement.  So,


      that is another challenge that we will start




                The argument has some validity but it is


      based on an assumption that current practices,


      including measurement systems and product


      specification, provide efficient means for


      identifying, understanding and then reducing


      variability.  For quality assurance in the 21st


      century we need a sound basis to verify such


      assumptions in the current system.


                To emphasize this point further, we


      discussed the case of dissolution and that is what


      we present to you today.  Let me illustrate an


      example, a real case example.  This is an example


      of an approved and validated manufacturing process


      at a major pharmaceutical company.  I will read the


      middle portion of this.  This is the warning



      letter:  There is no assurance that the production


      in process control procedures established--this is


      controlled-release product--to produce a product


      that has the quality it is purported to have or


      represented to possess.  How did we approve it?


      How was it validated?  So, this is after the fact.


      The duration of each coating cycle is determined by


      the pan operator but is based on visual


      determination that the coating solutions are evenly


      distributed before proceeding to the next step.  It


      is noted that literally 50 percent of the batches


      are thrown out every year because of dissolution


      failures, and then you have partial release


      occurring too.  Doesn't this undermine the entire


      credibility of our system?  And, this was


      catastrophic for the company.


                Now, inability to resolve our


      specification observations I think undermines the


      credibility of our decision system.  It raises


      questions of adequacy of the current decision


      system.  It increases the risk of releasing an


      unacceptable quality product to the consumer, and



      contributes to low efficiency.


                Now, corrective action, preventive


      action--there are some challenges.  There are


      difficult questions faced by manufacturing groups


      and regulators since we have a calibrated system


      that we use for dissolution and a calibrated system


      is a tablet similar to any other tablet that we


      use, and the quality is an issue there.  If you


      choose to use a calibrated tablet for gauge R&R


      study, reproducibility and repeatability study,


      what you see there is that the calibrator


      variability and its manufacturing process is


      confounded within that system.  I am not going to


      go through the equations but it is simple algebra.


                In addition, we have another challenge.


      The challenge is that the assumption of independent


      variable cannot be really verified because the


      hydrodynamics of the vessels are such--I see our


      colleagues from Health Canada here who have been


      criticizing this for a long time.  Thank you for


      coming, sir.  So, how representative is the


      suitability for that product is an issue.


                But the need for improvement is not


      limited.  We need to be confident of our analysis,


      of surveillance samples, consumer complaints, other



      investigations.  One of the frustrating jobs that I


      have is where we get consumer complaints; we do


      investigations; we do dissolution--no answers.  I


      mean, you really don't get to the root cause.


                I think the basic philosophy that Walter


      Shewhart sort of proclaimed years ago is very


      important.  Pure and applied science have gradually


      pushed further and further the requirements of


      accuracy and precision.  However, applied science,


      particularly in the mass production of


      interchangeable parts, is even more exacting than


      pure science in certain matters of accuracy and


      precision.  That is the basis of this discussion.


                Is the current approach to calibration


      adequate?  Dr. Cindy Buhse will share with you her


      challenges--as one of the premier labs, probably


      the world standard for dissolution at FDA and


      elsewhere, and Tom Layloff had started some of


      these processes and he is here too--dissolution



      testing of the USP wants to require diligent


      attention to details, mechanical and chemical.


      Dosage forms can respond definitely to small


      variations; large differences in dissolution


      results are possible unless all parameters are


      carefully controlled.  Differences in


      reproducibility can often be traced to improper


      mechanical calibration or degassing.  Much of that


      is mechanical.  When you only have suitability


      criteria just based on a tablet, it hides some of


      this variability.


                We had a rude awakening to this ourselves.


      This is really when I started realizing the


      confounding nature of the problem that we have.


      Just to illustrate how frustrating this experience


      was, our marines were contracting malaria when they


      were in Liberia and we were asked to see whether


      this was a quality problem.  We faced significant


      challenges in analysis because I had insisted that


      two labs would do this because this was a grave


      situation.  Unexpected inter-laboratory differences


      highlighted limitations of current calibration. 



      Here is just a quote from our DPA lab:  We are at a


      loss to explain the difference between DPA and the


      Philadelphia district office initial results.  Then


      we started tracing it back.  It had to be


      mechanical differences and degassing.


                Well, I think that is not the only issue.


      I think the bigger issue that we are confronted


      with is that we need to better understand the


      sources of variability in product performance and


      quality so as to establish the most appropriate


      design specifications for the product that support


      continuous improvement and address the increasing


      complexity of product designs.


                This is another concern.  We are moving


      towards drug eluting, towards nano materials,


      towards other complex devices and, yet, we don't


      have good measurement systems for these products.


      We want measurement systems for products intended


      for non-oral administration and non-oral drug


      delivery systems; develop and implement globally


      harmonized proactive regulatory decision system,


      including Q6A and Q8.


                I just want to sort of lay the foundation


      for other aspects that Mehul and Vibhakar will


      share with you.  Pharmaceutical development and



      dissolution specification without pharmaceutical


      development information creates more challenges.


      Decisions focus only on dissolution test data.


      Tests are often used for both in-process control


      and final product testing.  Decision


      characteristics focused only on the mean value will


      deal with variability indirectly.  Variability


      managed indirectly using "disconnecting test


      conditions" and acceptance criteria leads to


      deterministic interpretation of specifications and


      ignores background variability and, as Dr. Woodcock


      has said, we need to move towards a probabilistic


      decision system.  Specifications are standards and


      standards don't give any room for uncertainty or


      risk-based decisions.  If you don't meet the


      standards, you are off the market.  It is as simple


      as that.  So.  And you have event trees as opposed


      to decision trees.  It is difficult to resolve


      specification observations which could be related



      to how we set specifications, and post-approval


      changes and optimization in continuous improvement


      is difficult.


                This is simply an illustration of the gap


      that we base all of our decisions on test-to-test


      comparison, in vivo to in vitro, and there is an


      opportunity to use the design information to make


      rational decisions.  Just to illustrate this, again


      this is from Health Canada which has been very


      proactive and pushing this agenda and I am sorry we


      just didn't react more quickly--here is an


      illustration of the false-positive and


      false-negatives that you get.  The reference


      product dissolves 95 percent in 15 minutes, and the


      reference AUC, Cmax.  But if you look at product F,


      it dissolves very slowly in vitro but, yet, in vivo


      it meets the criteria--it is almost identical to




                So, there is a formulation attribute that


      does this.  For example, if you have a large amount


      of organic or insoluble excipient it is a


      hydrodynamic effect.  That doesn't happen in vivo. 



      The in vivo media, the surface tension, the


      hydrodynamics are completely different.  So, you


      tend to see this but you also get false-positives


      and false-negatives.  If I look at product C, it


      has only 62 percent dissolution compared to product


      F and has half the Cmax.


                There are other differences in how we


      approach specification setting.  The difference


      between the U.S. and Japan--we included a paper of


      the Japanese perspective on this in your background


      packet.  Because of the new restrictions I took the


      names off.  I had to go back and erase those.  This


      is a published paper so I was surprised that I


      needed to take the names off.


                The point here is this, all are basic


      drugs and this is a rule of thumb that has been


      known for 30 years, if you have a drug with PK


      between 4-6 the best media to illustrate in vivo


      performance is that of the PK value.  That is where


      the dissolution is slower.  So, the Japanese have


      been in that direction.  All our specifications use


      0.1 normal, here.  Is that important?  Well, the



      Japanese think so because they are very concerned


      with hypoacidity in the subjects.  If I really look


      at it, with antacids and H2 blockers most of us are


      hypoacidic too.  So, is this a gap that we need to


      fill is the question that I think we will address


      as we go along.  So, you can see the dramatic


      difference in dissolution as pH 1.2 to pH 7.2 and


      the resulting blood levels.


                So, in a sense, the opportunity we are


      trying to realize is ICH Q6A actually had it quite


      nicely captured in this quote:  The quality of a


      drug substance and drug product is determined by


      the design development, in-process controls, GMP


      controls, process validation and by specifications


      applied throughout development and manufacture.


      So, you have the goal; you have the decision


      characteristics; and you have the life cycle.  The


      design development was the missing element in our


      decision characteristics.  Now we have an


      opportunity to use it more effectively.


                This is how ICH Q8 captured that


      opportunity, to bring the development and design



      information not only to ask the right question but


      also to realize the opportunities of flexibility


      that might bring.  So, design and development


      should impact positively on how we set


      specifications in process controls and have more


      confidence in process validation and GMP controls.


                With that as a background and the reason


      for this topic for discussion, in many ways the


      tactical plan is an attempt now to go back ten


      years and to see how we can do better with our new


      information that could come through the PAT process


      and the ICH Q8 process.  In many ways we are


      reexamining the SUPAC guideline, the dissolution


      guideline for '97, the biopharmaceutics


      classification assumed in ICH Q6A.  The vector for


      the desired state is that we are adding another


      layer of variability assessment, identification


      assessment and utilization of variability in our


      decision-making.  So, the basic fundamental is that


      the quality of decisions can only be better so the


      current system is the minimum level of quality that


      we achieve.


                So, for the discussion today Cindy Buhse


      will share with you her proposal on measurement


      system, how mechanical calibration will be better



      and that is what we want to use.  Mehul Mehta will


      share with you the general overview of our decision


      process in our guidances.  Lawrence Yu is one of


      the leading experts I think in sort of modeling


      dissolution and in vivo absorption.  So, I have


      asked him to share a perspective on the current


      state of science.  Then I will come back and


      outline the steps of the proposal.  I have a number


      of slides in your packet but I will not be using


      those slides.  I will be using only the first 16


      slides to give you ample opportunity for


      discussion.  Those are backup slides.  If there are


      questions I will come back to them.


                In your background packet I specifically


      identified one person by name for his


      contributions, and that is Dr. Vinotcha [ph.].  I


      think the work he has done in particular--the


      reason I am pointing him out today is because he


      has decided to retire and I want to recognize his



      contribution to dissolution.  He has brought it to


      this level and I think taking it beyond that, and I


      thank him for that and he is here today.  A number


      of people are there from DPA who are experts in


      this and I will recognize them at some other point.


                With that, I will stop and invite Cindy to


      share her thoughts with you.  Any questions before


      I leave?


                DR. COONEY:  Thank you, Ajaz.  We will


      certainly have time for extensive discussion later


      but I think, particularly since we are right on


      schedule, if anyone has any questions for Ajaz


      right now, particularly for clarification of any of


      the points he has made, this would be a very


      appropriate time to take a moment for this.  Ken?


                DR. MORRIS:  Yes, just one quick point on


      an early slide where you were talking about the


      development process, it actually goes from the


      intended use through to development.  I would just


      say for clarification, because this is something


      that I get quite a lot, what we really want to get


      across I think is the idea that when you have the



      intended use and the characteristics you really


      select your process first.


                DR. HUSSAIN:  Yes.


                DR. MORRIS:  And then come back to the


      formulation.  So, it doesn't necessarily change the


      order but it adds a level because that is a


      constant source of confusion, particularly when you


      are talking about building in dissolution




                    Dissolution Measurement System:


            Current State and Opportunities for Improvement


                DR. BUHSE:  Thank you, Ajaz.  It is going


      to be my job to tell everybody a little bit about


      dissolution.  Some of you, I know, are very


      familiar with it but some of you may never have


      experienced it or seen it done and it is kind of a


      very different way of testing so I am going to show


      you a little bit about the different apparatus you


      can choose to do dissolution testing; talk a little


      bit about how we currently determine instrument


      suitability in terms of calibration, both


      mechanical and chemical; and also validation of



      dissolution of test methods and what we typically


      see in our lab when we take a look at method


      validation packages.  Then I am going to show you


      some sources of variability within dissolution,


      show you examples of how some formulations are


      sensitive to some parameters and some formulations


      are sensitive to others and we really need to


      understand for your particular formulation where


      your sources of variability are coming from.  Then


      I will just briefly talk about some opportunities


      for improvement, many of which Ajaz already alluded


      to in his talk.


                If you go to USP, there are seven


      different dissolution apparatus listed.  They are


      all up here.  You can see that the ones I am going


      to talk about today mostly are apparatus 1 and 2


      because those are the two that are used the most by


      most pharmaceutical companies.  We do see some of


      the other apparatus occasionally.  Apparatus 3,


      reciprocating cylinder, can also be set up for


      apparatus 7 so those are actually the same piece of


      equipment.  The flow-through cell is used more in



      Europe than it is in the United States.  We don't


      see much with that here.  Then, apparatus 5 and 6


      are used for transdermal delivery systems and they


      are actually a modification of apparatus 1 and 2.


                What I am going to talk about most today


      is apparatus 1 and 2, which is actually the same


      piece of equipment and what you are doing is you


      are changing the shaft on the different vessels to


      change it from apparatus 1 to apparatus 2.


      Actually shown in the picture there is apparatus 2


      and you can see there are paddles above each one of


      the about 900 ml vessels there.  The way


      dissolution works is that you are actually testing


      6 tablets at once.  I think Ajaz showed that in the


      specifications there usually is a specification


      which says 6 tablets have to have a certain


      dissolution value and if one of those 6 fails you


      go to 12 tablets and then you go to 24.  So, you


      start with just 6 and if everything goes right,


      then you will be done after the 6 tablets.


                So, you essentially have 6 different


      pieces of apparatus here because each one of those



      vessels acts independently.  You would fill each


      one with whatever media it is that you want to test


      in, whether it is 0.1 normal HCL or water or


      simulated intestinal fluid.  There are all sorts of


      ranges of media that people use.  So, you put


      500-900 ml in these vessels and then for apparatus


      2 you just lower the paddle down and start it going


      at whatever rpm you decide.  Certainly, that is


      another variable you can manipulate.  Then you drop


      your tablet or capsule in and then you take a


      sample out of the media at whatever time point your


      specification is.  If your specification may be 80


      percent dissolved after an hour, then after an hour


      you would withdraw a small portion of the media and


      then you would determine how much the drug has


      dissolved.  Usually the determinative step there is


      HPLC.  So, you do that for all 6 of these vessels


      and then, hopefully, everything dissolves in the


      right amount of time and you will be done.


                the basket--similar.  You just change the


      shaft and you put a basket on and you actually put


      the drug in the basket and then you lower the



      basket and start it spinning and you go through the


      same procedure.


                Just so you can see what it looks like,


      this shows you what apparatus 3 looks like, which


      you can also turn into apparatus 7 by changing the


      holders.  You actually would put the tablet or


      capsule inside each one of those up at the top.


      What it does, it comes up and down inside each one


      of these little vessels down at the bottom.  What


      you can do with this apparatus is you can change


      the media so in every row you can put a different


      medium if you want.  So, if you want to start your


      capsule dissolving at 0.1 normal HCL and move it to


      simulated intestinal fluid, in the first row you


      could put acid.  In the next row you could put


      intestinal fluid.  In the next row you could put


      whatever you want.  Then you can move this


      apparatus up, you know make it go up and down for


      an hour in one and then move to the next and go up


      and down.  So, that is how you could do it with


      apparatus 3.


                This is apparatus 4, and I think I



      mentioned we don't see a lot of this one.  This is


      a flow-through cell.  You can see over there, on


      the far side, that is what the actual cell looks


      like.  So, if you had a capsule or tablet that


      didn't completely disintegrate you could put it in


      this cell and actually flow through, somewhat like


      actually happens in humans--flow through a media


      and change it as you go.  You can either recycle it


      around or you can actually have a one pass through


      media as well and then analyze the media as it is


      coming out to see how much drug is dissolved.  For


      this one there is also a bunch of different cells,


      different geometries that you could put in this.  I


      kind of show examples of that there.


                Most of what I am going to talk about


      today is apparatus 1 and 2, and that is because


      that is the majority of what we see in methods that


      are given to us for method validation.  When they


      use apparatus 1 or 2 they use the USP criteria for


      setting up the equipment and for calibrating the


      equipment, and I will go over what those parameters


      are.  Then, as I think Ajaz said, most tablets and



      capsules have a one point acceptance criteria.  For


      immediate-release products we see anywhere from 2


      to 4 time points, maybe 1 hour, 4 hours, 8 hours,


      24 hours depending on the product.


                The first thing you are going to do if you


      have one of these apparatus, you are going to run a


      test method.  You need to ensure that you have


      instrument suitability.  The first point I have up


      there is which one of these 7 instruments you are


      going to use.  What we find is that most people use


      1 and 2.  Most people believe that that is what the


      FDA wants to see.  I have been to many different


      dissolution conferences and, you know, consultants


      and companies will get up there and say if at all


      possible use apparatus 1 or 2 because that is what


      the FDA wants.  I have heard many people say that


      so a lot of people try to use 1 and 2.


                Then, once you have chosen your


      instrument, you need to make sure it is set up


      properly for mechanical calibration.  You can see


      by the picture that if your shaft is not quite


      centered, or if your vessel is not quite seated



      right, your rpm aren't calibrated, etc., you can


      imagine that you can get different hydrodynamics


      from vessel to vessel or from time to time.  You


      need to really carefully make sure that everything


      is set up properly.  Then, once you have everything


      set up properly, you can then run a calibrator


      tablet provided by the USP to see if you get within


      the range that the calibrator tablet says you


      should get.  Then that gives you some measure of


      confidence that perhaps you have set this thing up


      properly with mechanical calibration.  I think Ajaz


      has mentioned that the calibrator tablets actually


      are U.S. phenomena and they are not used either in


      the European or Japanese pharmacopeias.


                Once you have instruments all set up, then


      you can certainly do method development/method


      validation, and I will talk a little bit about what


      we see and what is actually given to us, as the


      agency, when it comes to validating the dissolution




                Here is an example of some of the


      mechanical calibration parameters out of the USP. 



      Some of them have specific values.  For instance,


      the shaft has to be 2 mm from the centerline, which


      means you actually have a 4 mm spread because you


      can have one direction and then it spins around to


      the other.  You can see there are other parameters


      which don't really have any hard numbers associated


      with them, such as the wobble--no significant


      wobble and that is kind of nebulous there, or no


      significant vibration.  So, those are the some of


      the USP criteria for setting up the basket and


      paddle methods.


                The actual calibrator tablets--actually,


      our lab in St. Louis had a lot to do with


      calibrator tablets coming into being.  It is


      certainly the current 10 mg one that is used today.


      But they came around in the 1970s and there are two


      different calibrator tablets.  One is


      disintegrating and one is non-disintegrating.  So,


      one pretty much falls apart when it goes into the


      dissolution apparatus; the other stays together as


      a tablet throughout the calibration procedure.


                In 1997 a 50 mg prednisone tablet, the



      disintegrating one, was replaced with a 10 mg


      tablet which was manufactured at the University of


      Maryland, here, and was based on the formulation of


      a product that our lab had found was sensitive to a


      lot of the parameters of calibration, including


      degassing and mechanical calibration, so we thought


      it would be a good calibrator tablet.


                Actually, last year the working group at


      the USP was actually looking for a replacement for


      the 10 mg tablet.  It does have quite a bit of


      variability associated with it and some stability


      issues so they would like to see if they can find


      something else.


                So, if you are actually calibrating your


      apparatus what you would do, if you use your


      equipment for both basket and paddle which is what


      we do in our lab--a lot of pharmaceutical companies


      will have one that will always stay paddle and


      another will always stay basket but we go back and


      forth.  If you are using the same instrument for


      both paddle and basket, what you would do is you


      would do 4 different calibration runs.  You would



      do both calibrators with the paddle installed and


      then you would turn around and do both calibrators


      with the basket installed to make sure that your


      instrument is set up properly.


                How often do you do these?  In our lab we


      do it every 6 months.  We do the calibration using


      the prednisone 10 mg tablet.  Here is the actual


      data on the current lots of calibrator tablets.


      The O lot, which has been in effect now for almost


      two years I think--you can see there are different


      dissolution criteria depending on whether you are


      running it in the basket or the paddle method.  You


      see there is a fairly wide range.  You can see that


      for the basket as long as you are anywhere between


      53-77 percent for each vessel you are going to pass


      calibration.  So, you have your 6 vessels and this


      one, over here, can be 53 and this one, over here,


      could be 77 but you are still going to pass


      calibration.  Actually, late last year they changed


      the ranges of the prednisone tablet because there


      were stability issues and a lot of failures in the


      market, and you can see that the range is even



      wider now, 51-81 percent.


                I have also included up there the values


      we get in our lab for at least the prednisone


      tablet.  For the basket method we get 72.6.  You


      can see we run very much on the high end of that


      range.  In fact, we do quite often fall out on the


      high end.  You can see we tend to run on the low


      end of the range on the paddle method for these


      calibrator tablets.


                The salicylic acid tablet has a much


      narrower range.  It is also much less sensitive to


      many of the parameters that you set for dissolution


      testing so it is not sensitive to degassing; it is


      not sensitive to mechanical calibration setup.


                The problem often with running these


      calibrator tablets is if you do get an out of


      specification value, then what do you do?  You


      check your mechanical calibration.  It can be


      difficult to decide whether the issue is the actual


      calibrator tablet itself or the issue is some way


      that you set up the instrumentation.


                The other problem with the calibrator



      tablet is that you can see it has a fairly wide


      range.  It can often interfere with a continuous


      improvement process.  If your vessels can be


      anywhere from 51-81 percent and you are still


      passing, what does that say when you are running


      your own product and you want to try to narrow down


      the variability of your product?  You don't have


      much room here I guess to try to keep everything




                I am going to talk just a little bit about


      development and validation.  We don't see a lot of


      development data but we do see the validation data


      in our lab.  Obviously, when you are developing a


      dissolution method you have to decide about all


      these different parameters, a lot of which I have


      alluded to, and you want to develop a method that


      is going to be discriminatory.  You want to be able


      to tell between good product and bad product.  You


      want the method to be repeatable.  You would like


      the method to give you the same results no matter


      which lab you are running it in.  I think Ajaz said


      we had some trouble with the malaria drug in trying



      to get two different labs get the same results.


      You have to decide which instrument to use.  Like I


      said, most people try to pick 1 and 2 if at all


      possible; then what media to run it in.  A lot of


      the test methods we get either are in 0.1 normal or


      HCL; a lot of them are just plain water.  Then you


      have to decide whether degassing is going to be


      important or not for your product; and decide


      whether or not you need sinkers.  Some products


      don't automatically go to the bottom of the vessel


      if you are using the paddle method.  You can buy


      commercial sinkers, which are these little devices


      that you put the tablet in that will actually make


      it fall to the bottom, or you can just wrap a wire


      around, which is what is in the USP, to make it go


      down to the bottom.


                Once you have decided all these


      parameters, you still need a determinative step,


      and that is what the main focus of validation is


      for most companies.  So, when we get validation


      packages in from companies on their dissolution


      test methods, their validation really focuses on



      the determinative step.  They do a lot of work on


      varying the parameters on the HPLC method but less


      data do we see on varying the parameters on the


      actual dissolution method.  So, we see more on the


      determinative step and less on the actual


      parameters that are associated with the dissolution




                You can see that there are a lot of places


      here where variability can be introduced, and


      certainly when developing a product if you want to


      have a test method that is going to allow you to


      continuously improve your product you really need


      to understand what all the sources of variability


      are going to be.


                This is one of Ajaz's slides.  I think he


      showed a similar one earlier which is basically a


      slide just to show you that the total variability


      you are going to see in any test method is going to


      be the variability that is inherent to your product


      and your manufacturing process and the variability


      that is inherent to your test method.  For


      dissolution the variability inherent to the test



      method can be quite large, especially if you don't


      understand how all the different parameters can


      affect your product.


                I am going to just show some examples of


      some of the variability.  You can see I have a lot


      of information up on this slide, and every single


      one of these bullets can be a source of variability


      when running a dissolution test method.  You have


      to make sure your operators are well trained.  You


      have to make sure you have set things up properly.


      You have to make sure that you understand how all


      the different media and equipment parameters,


      sinkers etc., can affect the variability of your


      specific product.  So, there are a lot of places in


      here where, you know, if you add a tenth or so, or


      a percent or two of variability by the end you have


      quite a wide range of potential dissolution


      parameters you could get even with the same lot of




                When it comes to mechanical calibration, I


      think I showed some of the USP parameters earlier


      and what I want to show you here is actually that



      in our lab, DPA, we use more stringent mechanical


      calibration than what is listed in the USP.  A lot


      of the criteria we use come directly out of the


      PhARMA recommendation.  I think that paper is in


      your packet.  It came out in the '90s, where they


      did a collaborative study to take a look at


      mechanical calibration a little more closely to see


      if tighter mechanical calibration might reduce


      variability when running the calibrator tablet.


                Because we run so many products in our lab


      and we don't necessarily have the time to stop and


      see if this product is really sensitive to


      centering or not, etc., we just try to be very


      careful about how we set up our equipment.  Some


      tools are now available to very easily set these


      parameters much tighter than what is currently in


      the USP.  So, you can see that for quite a few of


      these we are tighter, and for others we have added


      criteria that are not actually in the USP as


      specific numbers.  For instance for shaft wobble


      and vibration, we actually measure those and set


      criteria for those.


                Degassing is one of the things I think


      that really got us into trouble--I don't want to


      use that word, but with the malaria drug the



      different labs were degassing in different ways and


      this drug happened to be very sensitive to


      degassing.  So, typically in the past the way you


      decided whether your media was well degassed or not


      is that you ran the calibrator tablet.  The 10 mg


      prednisone is very sensitive to dissolved gasses in


      the media so if you weren't sure if you were


      degassed or not you could just run that calibrator


      tablet to see if you were in range and then decide


      if you were degassed properly.


                Well, it turns out that there is some


      equipment on the market that you can use to


      actually measure dissolved gasses so this is


      something we have done recently in our lab.  We


      have taken this meter, which is actually used in


      other industries and not in the pharmaceutical


      industry, and used it to try to determine how much


      dissolved gases are left after using different


      degassing techniques.


                There are many different ways in which


      people degas their media.  The reason you need to


      degas your media is because there are some products


      that if you take a vessel and you drop in a tablet


      or capsule, what will happen is you have gases in


      the media.  The bubbles will form around this



      tablet or capsule and oftentimes will prevent it


      from dissoluting.  So, you actually need to get the


      gases out of there before you start.


                Here is a little graph of the different


      ways people degas and the results we got with the


      total gas meter, measuring both total gas and


      oxygen.  You can see that for the first bar over


      there that is obviously atmospheric pressure and


      atmospheric oxygen in the media.  These are all


      done in just plain water.  The next bar is the way


      we degas at DPA, which is point of vacuum at less


      than 150 ml of mercury with agitation, and you can


      see we get rid of about a little more than half of


      the total dissolved gases and quite a bit of the




                The USP method is also very good.  There



      you are heating up to about 41 degrees and


      aspirating to remove the dissolved gases.  They


      also get half the total gone and about half the




                Some people actually helium sparge and you


      can see helium sparging and although you do reduce


      the oxygen significantly you do not reduce the


      total dissolved gases.


                So, does this matter or not matter?  You


      know, this all depends on the product you are


      testing.  So, I just want to show you some examples


      here.  These are 3 different products, called


      product 1, 2 and 3 so I don't give any product


      names.  You can see that for product 1 and product


      2 there is a huge difference between non-degassing


      and degassing.  For both of those graphs I have


      shown two different ways of degassing.  One is the


      USP and DPA method, both of which give similar


      results.  The other is helium sparging.  You can


      see in both cases that the helium sparging does


      give slightly higher results than either the USP or


      the DPA method.  Certainly, for product 2 helium



      sparging gives much more variable results than the


      DPA degassing.  You can see that on the helium


      sparging line which is kind of the green-yellow




                You can see that product 3 doesn't really


      care whether you degas or not.  One of those lines


      is non-degassed and one is the DPA method which had


      the lowest percent of dissolved gases.  You can see


      that you get essentially very similar dissolution


      whether you degas or not.


                Larger than just degassing is the actual


      composition of the media.  I think as Ajaz


      mentioned, Japan is looking at what type of media


      you actually want to be using.  We see a lot of


      acid here and some buffers.  Here is a product and


      the dissolution method is pH 7.2.  So, 7.2, as you


      can see on your left I guess, is the media that is


      used in this product.  It also turns out that with


      these 6 different tablets there is some variability


      between the 6 but they all passed the dissolution


      specification for this particular product.


                This is a product where we wanted to take



      a look at some lower pHs just because there are


      some patients who happen to use this drug who may


      have lower intestinal pH than 7.2 and so we went


      down to 6.8 and, lo and behold, every single tablet


      looked different to us and no two tablets were the


      same.  We repeated this over and over again, trying


      to figure out what is going on.  You can actually


      do a lot with dissolution by just watching your


      product.  There is nothing like the human eye




                If you watch this product in the vessel


      what you will see is that it sits there and does


      not dissolve and you get no dissolution until you


      see the coating split open.  Once the coating


      splits open, then it dissolves fairly quickly.  So,


      taking a look at that we were trying to figure out


      what could be the sources of variability of this


      product.  Is it the way we are handling it when we


      put it into the dissolution vessel?  Are we


      damaging the coating in some way?  Are these tablet


      differences real or is this the manufacturing


      process itself?  Do we have instrument variation? 



      These 6 tablets are in 6 different vessels so is


      there some difference in these vessels where maybe


      we have improper calibration or something?


                Well, after much investigation, what we


      found is that this is actually a product problem.


      If you cut open these tablets and take a look at


      the coating, not all of them have uniform coating.


      You can see there, on the left, one of the tablets


      that has a very uniform coating thickness.  Then


      every once in a while you ran across a tablet that


      had a void between the drug and the coating.  The


      drug is actually on the left side here; it is kind


      of the yellow sparkly stuff and the red is the


      coating.  So, some of the tablets had very uniform


      coating; some of the tablets had defects.  These


      defects were dissolving much faster or were


      breaking open, splitting open much faster than the


      ones that didn't have defects.  This is a situation


      where perhaps dissolution could help this


      manufacturer make a more consistent product if they


      were doing their dissolution at a slightly


      different pH or doing a dissolution test method at



      several different pHs to try to make sure they were


      making a consistent product.


                I was just going to mention sinkers


      because I talked about them and also because they


      do make a big difference.  The graph up there


      actually has nothing to do with the sinkers but it


      shows you what happens if you don't get your tablet


      at the center of your vessel.  The bottom blue line


      is product 1, right down at the bottom of the


      vessel, centered completely.  The green-yellow line


      is if it is off center by 1 cm.  So, if it is just


      off center by a centimeter you can see that it


      dissolves much faster.  There are different


      hydrodynamics in that area than at the bottom of


      the vessel.  So, if you have a tablet that is


      fairly light and is not going to stay put, then


      often you will put it inside a sinker.


                Traditionally, in our lab we have used the


      sinker at the top to the right.  That is the one


      that we have used in our lab.  It is very easy to


      use.  It has a spring load and you just pull back


      the spring and drop the capsule or tablet in and it



      is, you know, very convenient I guess.  The USP


      method is to use a wire and wrap the wire three


      times around the tablet or capsule.


                Well, we did run across a product--this is


      what I talked about, that you have to understand


      your product and how it reacts to different


      variables--that was sensitive to this actual


      commercial sinker.  This is the product we tested


      and with the commercial sinker that I just showed


      you it failed dissolution.  The specification here


      was 80 percent at 30 minutes and you can see that


      all 6 tablets failed.  Of course, we thought the


      product was perhaps a failure but it actually


      turned out that if you visually looked at what was


      going on, the product was being trapped.  It was


      swelling up and getting trapped inside that


      commercial sinker and so it could not essentially




                We went back to the USP method with three


      wire turns around the tablet, and you can see that


      the product passes wonderfully with no problems


      whatsoever.  So, we no longer use commercial



      sinkers in our lab but a lot of people use them so


      I just wanted to make you aware of the fact that


      something as simple as a sinker can affect the


      individual product that you are looking at.


                So, what I have tried to show you is just


      some data that illustrates the fact that different


      products are sensitive to different parameters when


      you are doing dissolution, and there are obviously


      a lot of places where you can introduce variability


      in your test method.  What we would like to propose


      is an alternate approach to calibration and


      validation which includes complete understanding of


      how dissolution and the measurement system in your


      product specific variables affect variability, and


      try and understand the relationship between your


      product properties and your dissolution results.


      This includes understanding the dissolution


      apparatus that you are using, why you are choosing


      it and why you are choosing the media you are


      choosing, and determine, hopefully, the best method


      to give you opportunities for improvement and to


      ensure that the quality of your product is good.


                You can see that because of the way


      dissolution is currently set up there are a lot of


      things you have to control, and perhaps there are



      new approaches we can also use to get the same type


      of information that might have inherently less


      variability.  Then, obviously, a part of this whole


      process needs to be communication and training.  If


      people are out there saying that FDA wants us to


      use apparatus 1 and 2, then that is what people are


      going to do.  So, the FDA is trained in a more


      open-minded look at other things.  If people feel


      that way at least, then they might be willing to


      look at other approaches.


                When it comes to alternative approaches to


      dissolution calibration validation, I think as I


      told you in our lab we do more stringent mechanical


      calibration because some products are very


      sensitive to how the apparatus is set up and,


      certainly, if you set it up properly your


      variability will be less than the variability of


      the calibrator tablet.  Certainly, when you are


      using your specific product itself, you need to ID



      and control all the source of variability that you


      are going to see.  You need to determine how your


      product is sensitive to things like the apparatus


      type, the setup parameters and the media, both type


      of media and whether it is degassed or not.  There


      is an interaction between the instrument you use


      and your product, and understanding that is going


      to also help you reduce the variability in the


      dissolution test method.  People like to use


      calibrator tablets.  I think it gives them a


      measure of confidence that they set everything up


      and their system is suitable.


                So, what we are proposing is that


      certainly the USP calibrator can be used if


      somebody wants to take a look and see that they


      have set up properly.  Perhaps it also might be


      useful to set up an internal calibrator maybe based


      on a bio. batch or clinical batch to make sure of


      system suitability.  The calibrators dissolve in a


      certain way or are sensitive to certain things and


      not sensitive to certain things, the USP ones, and


      those parameters may not be the parameters that



      your particular product is or is not sensitive to.


      So, creating your own internal calibrator and


      understanding how your product is sensitive to all


      the parameters is going to be perhaps better than


      an outside product that may not have the same


      sensitivities that yours does.  Obviously, you need


      to confirm the suitability of your internal


      calibrator using some kind of a gauge R&R study so


      you can really understand what the variability is


      in your product.


                Ajaz mentioned gauge R&R a little bit.  If


      you pick a lot of product or a piece of a lot to


      maybe set up as an internal calibrator you need to


      carefully characterize that and determine what its


      variability is.  You want to make sure it is


      representative of your manufacturing process.  You


      want to make sure that it was manufactured while


      your process was under control.  Obviously, when


      you are doing a gauge R&R you need to take a look


      at what variability is introduced instrument to


      instrument, vessel to vessel.  As you can see, each


      instrument is like 6 individual little instruments.



      And variability from personnel to personnel and,


      obviously, media and whether it is degassed or not.


                We need to understand the benefits and


      limitations of the different dissolution apparatus.


      I showed you that there are 7 different ones in the


      USP.  We also sometimes get ones that are non-USP


      apparatus when people submit test methods.  So,


      there are a lot of different things out there to


      choose from and, better than just choosing one that


      someone thinks maybe the FDA wants to see, maybe


      try to understand how the hydrodynamics work; try


      to model your system.  Actually, I have been told


      by people who do modeling that apparatus 1 and 2


      are difficult to model so there may be some better


      systems out there where we can do some better


      predicting of what is going to happen as we change


      physical parameters of our product, and take a look


      at some other things we might be able to do.


                Of course, what would even be better is


      just quit doing dissolution as it is known today


      and maybe find some other ways to assess product


      quality.  People have done some work in our sister



      lab here, in White Oak, to try to correlate


      dissolution with NIR.  There is a lot of


      spectroscopy out there that can be used online as


      part of a PAT feedback loop, and perhaps good


      correlations and good models could be developed


      between those and quality and in vivo availability


      and we can dispense perhaps with the current


      dissolution test method, which has all of its


      parameters--things that can go wrong and need to be


      set very carefully.  Obviously, key to this is


      going back to the first principles and modeling and


      understanding your formulation, and how each


      component of your formulation contributes to the


      quality of your product.


                So, that is all I had to say and I just


      wanted to acknowledge Terry Moore, who is actually


      here today, who probably knows more about


      dissolution than anybody in the world.  He is


      sitting over there, if you want to know more about


      dissolution.  Then, Zongming Gao is also in our lab


      doing dissolution; and Lawrence who also knows a


      lot about dissolution; and Ajaz all helped with



      this.  So, thank you.


                DR. COONEY:  Thank you very much, Cindy.


      There certainly is time for questions.  Gerry?


                     Questions by Committee Members


                MR. MIGLIACCIO:  Cindy, first I applaud


      your last comment about using alternate methods.  I


      just want to point out that you made several


      comments about the use of apparatus 1 and 2 and,


      speaking I think for most companies, we don't use 1


      and 2 because we think FDA wants us to use them.


      You did a great job of pointing out the variability


      of the different parameters that can impact


      variability.  It is very important when you are


      testing thousands of batches a year that you have a


      really well trained work force that knows how to


      use this apparatus, and that you have consistency


      in the way you test because if you are switching


      from one apparatus to another it presents another


      level of complexity.  So, it is really the


      consistency.  Because of the variability that is


      inherent here, it is the consistency that drives us


      to apparatus 1 and 2 and not a lack of desire--


                DR. BUHSE:  To try something else?


                MR. MIGLIACCIO:  --but, you know, it is


      complicated enough so it is really consistency that



      drives us there.


                DR. COONEY:  Marvin?


                DR. MEYER:  The data you showed from your


      lab versus the specs on prednisone, and you said in


      one case you tend to be high and some cases fail,


      when you do fail the calibration what do you do


      about it?  Is it the calibration that is no good?


      Is it the USP specs that is no good?  Is it the lab


      that is no good?  Or, do you just keep going until


      you have 36 samples?


                DR. BUHSE:  Well, historically what we


      have done is double check your mechanical


      calibration and then you really run the calibrator


      tablet.  So, was the original failure the tablet?


      Rarely do we find something to adjust when we check


      the mechanical calibration.  We do the mechanism


      calibration much tighter than the USP anyway so


      essentially you rerun.  We actually don't run them


      anymore in the lab, the USP calibrator tablets.


                DR. MEYER:  That solves that problem!


                DR. BUHSE:  That solves that problem!  We


      have an internal calibration tablet that we use now


      that we have characterized ourselves in our lab


      that has lower variability.  We stopped using this


      one probably at the end of last year.  The data I



      showed was the data from 2004, 2003.


                DR. MEYER:  The other question I have or


      comment is that on one of the slides you suggest


      using perhaps an internal calibrator, a bio. batch


      or some known that you have produced.


                DR. BUHSE:  Right.


                DR. MEYER:  How do you know that that


      product, over the lifetime of the product being


      manufactured, hasn't changed?  Dissolution doesn't


      change, you are satisfied your equipment is in good


      order when, in fact, it isn't because you couldn't


      pick up the change--


                DR. BUHSE:  Stability is a big issue.


      Stability is an issue with the current USP


      calibrator.  It is known to drift down I believe


      with the paddle method over time, or whatever.  Do,



      you want to talk about that, Ajaz?


                DR. HUSSAIN:  Yes.  Marvin, I am going to


      go over that in detail.  The gauge R&R is actually


      for three purposes.  It is to establish and


      benchmark the variability.  I think the proposal


      actually is that mechanism calibration actually is


      sufficient.  The gauge R&R is an opportunity to


      establish your target.  You benchmark your


      variability and then use that variability for


      setting specifications, and so forth.  But then you


      have that and then you can keep the system stable.


      I think stability of the system has to be based on


      mechanism calibration.  That is what other


      countries do anyway.  So, I will go over that in a


      bit more detail.  So, the opportunity is more than


      just the internal calibrator.  So.


                DR. MEYER:  One follow-up, I kind of joked


      that you made the problem go away because you are


      not using it anymore.  What if you are a company


      and had in your NDA or ANDA that you would


      calibrate your dissolution using the prednisone and


      USP and you started to fail, your dissolution



      couldn't meet the calibration?  They don't have the


      luxury of just saying, well, we are going to use


      our own now because they are stuck with using what


      they said in the NDA, right?  What should a company


      do about that?


                DR. BUHSE:  You want to talk about that,




                DR. HUSSAIN:  Well, I think this meeting


      is step one to start addressing that in a sense.


      Here is an alternate procedure.  So, I think if the


      advisory committee will sort of endorse this and we


      move that way, we will put that in policy and there


      are many different ways to implement that.  So.


      But from the compendia perspective, I think you


      have to comply with the compendia so that is a


      different challenge that the industry and companies


      have to deal with.  So, all we are doing right now


      is creating an alternate regulatory decision


      pathway and our enforcement strategy based on that.


                DR. COONEY:  Nozer?


                DR. SINGPURWALLA:  Slide number 13, I


      thought you said it was Ajaz's slide.  Therefore,



      it is wrong!




                DR. BUHSE:  Yes, it was Ajaz's slide.


                DR. SINGPURWALLA:  Well, how do you


      distinguish between repeatability and




                DR. BUHSE:  Well, I was going to say with


      a destructive test it is very difficult.


                DR. HUSSAIN:  See, this is gauge R&R for a


      destructive test.  You really have to have design


      experiment and I was going to cover that in my


      talk.  What this does is, it actually ensures that


      the lot you choose is stable and in a state of


      control.  That is the only way you can actually


      move in this direction.  So, that achieves that


      target.  The destructive gauge R&R is a very formal


      experiment and it is a nested design which does get


      an estimate of whether a practice or an operator


      can repeat it.  That is repeatability.


      Reproducibility is the variability associated with




                DR. SINGPURWALLA:  So, the repeatability



      refers to a physical thing.  The other thing is I


      don't know how important it is for you to manage


      variability but if it is important to you to manage


      variability, then my sense is that as the product


      variability increases the measurement variability


      will also increase.  Therefore, there will be


      correlation and, therefore, the sigma squared total


      that you have will be underestimated the way you


      have put it down.  If it is of any importance, you


      may want--


                DR. HUSSAIN:  I think it is.  That is the


      reason the leverage--the quality by design having


      the pharmaceutical development information starts


      to allow us to dilute some of this.  But the


      variability that you are observing you are


      observing to the eyes of the measurement system so


      the measurement system and variability in the


      product are together.  I will try to come back and


      sort of explain some of that.


                DR. DELUCA:  I apologize for my voice.


      You very nicely pointed out the multitude of


      variables that are involved.  There is instrument



      variability as well as product variability so you


      have interaction.  You mentioned degassing.  But


      you are using a set agitation in your system.  When


      you start degassing, are you not sparging?  Now,


      you can create agitation or sparging during the




                DR. BUHSE:  No, it is done beforehand.


      You do it before you start and you put the media in


      the different vessels and there is no degassing


      during dissolution itself.  Questions come up,


      especially for extended-release products, where


      actually the dissolution test method lasts for 24


      hours per product, and the question then becomes


      what happens to the gas level over that time.  We


      hope to test that with this meter.  The one I


      showed you here is actually one that has a probe


      that is, like, 3 inches around so you have to put


      it in a giant vessel.  They are making a new probe


      that is small and will fit inside the dissolution


      vessel so we can see what happens actually in the


      dissolution vessel over time.  Like you say, with


      some of these test methods at high rpm, 100 rpm, we



      are getting a lot of agitation.  So, that is a good




                DR. DELUCA:  And I was worried about the


      product and how product variation can affect--so,


      you have an interaction between the instrument and


      the product where particle size might influence,


      you might have a set agitation rate.  If the


      particle size changes then it is going to change


      the result.


                DR. BUHSE:  Right, unless you have a


      method that can discriminate that if it is


      important to the acting of the drug.


                DR. DELUCA:  You have talked about


      modeling, I mean you mentioned it.  Maybe it is


      going to be covered later on, but I wondered if you


      include anything here to look at profiles, release




                DR. BUHSE:  We haven't done a lot of


      modeling yet in our lab.  I don't know if we are


      going to talk about that specifically later on or


      not today.


                DR. COONEY:  Ken?


                DR. MORRIS:  Just a couple of things.  One


      is that given the sort of lag--I guess I just have


      a philosophical problem with calibrator tablets in



      that if you are looking at a process and want to


      independently establish that it is in control or


      that it is doing what you think it is doing--we are


      producing these the same way we produce the tablets


      for testing--


                DR. BUHSE:  That are no better.


                DR. MORRIS:  What is that?


                DR. BUHSE:  That are no better.


                DR. MORRIS:  In fact, there are some data


      that I think we will see to day that there are some


      liabilities.  I think maybe this is something we


      will talk a lot more about, I am sure, but I think


      one of the things that may come out of this is that


      calibrator tablets just don't have a prominent


      role.  What I would say is that if you look at an


      immediate-release system--and we will also get into


      BCS exemptions--then the issues become sort of


      treatable in other ways.  If you are looking at


      sustained-release or modified-release, such as



      enteric or extended, then my argument is that you


      ought to be controlling the coating process and


      that sort of activity is really much more advanced


      than it was.  I mean, you have your example of the


      tablet that has the air pocket but probably what


      was more important was the difference between the


      80 and 50 micron coating thickness.  This is


      clearly a failure of reproducibility of coating and


      the dissolution may catch it or may not.  I mean,


      the statisticians--I don't know, there is the


      Bayesian argument but I have talked to Sandy


      Bolton, for one so, you know, if you have high


      variability dissolution maybe 6 tablets is enough


      to pick it up but, depending on what constitutes


      high variability, you know, it is in the laps of


      the gods whether you get it or not.  So, to the


      extent that things are surface-based alternate


      methods--I mean, in the first place, you want to be


      controlling the coating processes and then, to the


      extent they are surface-based, have you considered


      things within the group like the combination of


      that and, like, IGC to look at surface free



      energies or something that is at least a little


      less subjective?  I don't know if you have talked


      about it because everything else is a correlated




                DR. BUHSE:  Right.


                DR. DELUCA:  Whereas, something that


      actually measures surface free energy, even though


      there is no practical instrument right now, is a


      direct measure.


                DR. BUHSE:  We haven't done that with that


      particular product.  We have tried to do some


      spectroscopy correlations.


                DR. HUSSAIN:  If I may?


                DR. COONEY:  Yes.


                DR. HUSSAIN:  I think you make a good


      point, and I think the goal that we have, number


      two, desired state, specification based on


      mechanistic understanding--so, if the mechanism is


      controlling the dissolution based on a coating


      thickness, if you are able to measure the coating


      thickness reliably, and so forth, that should be


      sufficient.  So, I think that is the direction we



      wish to move in, and some of the new technologies


      and science sort of helps that.


                There is another point I think which I do


      want to make and this is my graduate school


      training; this is biopharmaceutics 101.  When we


      approach trying to develop a product we first think


      about the patient, and so forth.  Prof. Richard


      always insisted you don't even think about an in


      vitro test.  You first try to get initial


      information in humans and then say, all right, what


      sort of testing will we need.  So, you establish


      your formulation, human connection or patient


      connection first before spending time in an


      artificial way.  In my consulting role before I


      came to FDA, one company I worked for carried out


      53 experiments, screening and so forth; they had no


      idea whether dissolution was useful or not.  They


      spent all this development effort trying to


      optimize a hypothetical, what they thought was the


      dissolution rate and the first experiment they did


      was completely off.  So, all the experiments were


      actually off target.


                So, there is a tendency within industry to


      assume that in vitro dissolution is going to guide


      them to a formulation without even understanding



      its relevance.  I think Marvin knows that company


      very well.  We actually had a paper on that issue


      together.  So, there are challenges I think.  So,


      quality by design actually forces us to think what


      is the patient and then think about the tests so


      that is what we are trying to achieve here.


                DR. COONEY:  Ajaz, perhaps we can capture


      that as a point, that the purpose of formulation


      development is to optimize patient care, not


      dissolution assay.  We hear you.


                DR. FACKLER:  Could I just make a point?


      Dissolution can function for a number of different


      purposes and on one of your slides you suggested


      that finding a discriminating method might be


      useful, and I would agree under certain




                On the other hand, if you look at that


      enteric-coated product really the purpose of the


      enteric coating is to protect the tablet for the



      first hour, or whatever time it might exist at the


      very acidic condition of the stomach.  Whether or


      not coating breaks open at one hour, two hours or


      three hours might have no relevance to the in vivo


      performance of the product.


                So, I think it is important, as we talk


      about the future of dissolution testing, to


      recognize what it is intended for.   If it is


      intended to predict in vivo performance, that is


      one thing and a predictive or correlative method


      then I think would be the ideal.  If it is to look


      for product quality and to reduce the inherent


      variability in products, well, then a more


      discriminating method that might have no relevance


      to in vivo performance might be our goal.  I think


      we just need to keep that in perspective as we


      think about the future of dissolution testing.


                DR. HUSSAIN:  If I may since we have time,


      I think this is one of the first steps in our


      tactical plan.  Since we have time, if we could


      engage the advisory committee to make sure is this


      an acceptable step further discussion is needed. 





                DR. COONEY:  Tom?


                DR. LAYLOFF:  Yes, I was going to say


      because of my concern with the problem with


      degassing--I never degas my stomach before I take


      my medication--




                DR. FACKLER:  You probably don't swallow


      900 ml of water either.




                DR. MORRIS:  Just a couple of comments.


      First, when we validate equipment we have to


      understand what tests we are doing and what we are


      trying to validate, and the standard tablet just


      doesn't--intuitively, it doesn't get there for me


      because we are looking at validating a piece of


      equipment and all of a sudden the variables that we


      are throwing into the pot include what is the


      dissolution medium and how we handle that; what is


      the size of tablet and how we handle that when we


      are trying to validate a piece of equipment.  So,


      probably the first step is saying what validates



      the equipment, and anything else we do is a waste


      of time.


                Then, the next step, to get right to what


      Paul said, is what is my dissolution test telling


      me because I am a manufacturer and I want to keep


      my process under control, or am I predicting what


      is happening in people?  We have seen for 35 years,


      as far as I know, that dissolution doesn't predict


      the human results in terms of bioavailability or


      bioequivalency.  You can't do it that way.  You


      have to get that data and then try to correlate.


      So, if we are using dissolution for quality


      control, for process, fine, then there is a set of


      variables and we do it that way.  But if we are


      trying to say that I can do a dissolution study


      and, therefore, I will know that my formulation is


      going to work in a person I think we are really


      biting way more off than we can chew.


                DR. HUSSAIN:  I think I agree with you,


      but in may aspects you do establish correlation.


      Actually, Lawrence, in his talk, will actually make


      that same proposal as you did.  So.


                DR. COONEY:  Tom?


                DR. LAYLOFF:  The early work done on


      digoxin was designed to go for in vivo/in vitro



      correlation for about 35 manufacturers, and that is


      how that standard was set.  Prednisone subsequently


      was done the same way.  In reviewing that, it would


      determine that if the FDA continued down that path


      it would eventually take all the resources of the


      FDA to do it because of the cost of performing


      those in vivo/in vitro correlations.  Then the


      dissolution standard was just arbitrarily applied


      across the board.


                DR. COONEY:  Marvin?


                DR. MEYER:  Ajaz, I think you ask if you


      are on the right track and I think you definitely


      are.  You know, when you first said we are going to


      revisit dissolution I said, oh, my God--




                --so, I think you are on the right track.


      I mean, for me, when I used to do some dissolution


      just in a university laboratory, I loved the wide


      range for the calibrators because then my equipment



      always passed and I didn't have to worry about it.


      But now, sitting around this table, I have a


      different hat on and it is shocking, 51-81 percent.


      How can you have a calibrator--if somebody comes in


      with analytical data like that you would say go


      away; this is a very poorly controlled analytical


      procedure.  So, I think that revisiting the issue


      is very important.


                DR. HUSSAIN:  Marv, in may ways, you know,


      I was blind to this.  I actually was not fully


      aware of the scenario, and Cindy will attest to


      this.  When I started writing this paper I put


      Lawrence through hell.  I said how could this


      happen?  Because our standard criteria for


      specification is plus/minus 10 percent and the


      instrument is this way so there was a disconnect


      that I was not aware of and I have to apologize for




                DR. COONEY:  Are there any other comments


      or questions at this point?


                [No response]


                What I would like to suggest is that we



      take a break for 15 minutes and reconvene at 10:25,


      and we are in good shape for continued discussion


      and I have no doubt there will continue to be more.


                [Brief recess]


                DR. COONEY:  I would like to now welcome


      Dr. Mehta to speak to us about an overview of the


      current guidance on the documents and decision


      process in biopharmaceutics.


              Overview of Guidance Documents and Decision


                   Process:  Biopharmaceutics Section


                DR. MEHTA:  Good morning.  As you can see


      on my slide here, I am asked to give an overview of


      guidances documents and decision processes from a


      biopharmaceutics perspective.


                Before I start, I want to acknowledge a


      couple of people in my division, Dr. Ramana Uppoor,


      she is sitting in the back in the audience, and Dr.


      Patrick Marroum, team leaders in neuro. and


      cardiorenal in my division and some of the experts


      in biopharmaceutics in my division.


                This is the outline of my presentation.  I


      am going to give you an overview of



      biopharmaceutical aspects of dissolution-related


      guidances.  That is a formidable task.  My first


      draft that I sent to Ajaz had 100 slides and Ajaz


      replied by saying an excellent overview but cut it


      down.  So, I am now down to 60.




                But I still intend to finish in time.


      Then with a quick overview I will take you through


      some examples from our NDA reviews of


      immediate-release and modified-release products,


      and share with you my perspective on opportunities


      for improvement.


                These are the guidances I am going to


      quickly take you through.  Chronologically they are


      different but in terms of science, the way the


      ideas are represented I have shifted them around.


      I am going to first start with the BCS guidance.


      In parentheses are the references.  I will follow


      that by the immediate-release dissolution guidance


      that came out in 1997.  The BCS guidance was


      finalized in 2000.  The IR dissolution guidance


      invokes BCS principles and that is why I have



      arranged it that way.  That will be followed by a


      quick overview of the IVIVC guidance and that is


      for modified-release products, in vitro/in vivo


      correlation.  Then a couple of slides on general


      bioavailability and bioequivalence guidance, which


      was finalized in 2003.


                I will quickly switch to something known


      as scale-up and post-approval changes for


      immediate-release products and modified-release


      products, and the topics covered there.


                So, let me start with the BCS guidance


      summary.  Maybe it is known to everybody, but just


      for the sake of completeness let me point out the


      highlights of the BCS guidance.  This guidance


      takes into account three major factors that govern


      the rate and extent of drug absorption from the


      immediate-release solid oral dosage form.


                These are the solubility and intestinal


      permeability of the drug substance, and dissolution


      of the drug product.  So, based on the solubility


      and permeability characteristics of the drug


      substance the drugs are classified into four



      categories: high solubility, high permeability; low


      solubility, high permeability; high solubility, low


      permeability; and then the fourth category, low


      solubility, low permeability.


                The third bullet is the central idea, the


      central concept, a very sound scientific concept of


      BCS which is, you know, if a drug product is BCS


      class 1, and for different formulations of this


      class 1 product if they are rapid and similarly


      dissolving you can give a biowaiver for the test


      formulation without requiring an in vivo


      bioequivalency assessment, provided you show


      similar dissolution profiles over the physiological


      pH range.


                The last important point about this


      guidance is that in this guidance we have defined


      what determines rapid dissolution, and we say if


      your drug product dissolves 85 percent in 30


      minutes over the pH range absorption should not be


      dissolution limited.  So, that is all for BCS.


                Moving on quickly to the immediate-release


      dissolution guidance summary, and again I will do



      my little bit of acknowledgement here, Dr. Shah and


      some members on the panel here have contributed to


      this guidance and, from my personal perspective,


      this is scientifically a very well written document


      although it was almost ten years ago.


                These are the topics covered in this


      guidance.  The guidance lays out approaches for


      setting dissolution specifications for a new


      chemical entity.  As I said, it takes into


      consideration BCS nature of the drug product and,


      depending upon that, you can have minimal


      dissolution requirements in setting specifications


      or more stringent.


                Another very important point from my


      perspective is that this guidance has outlined


      something known as mapping or response surface


      methodology.  Again, this is supposed to be for


      immediate-release products.  The guidance says that


      undefined clinical manufacturing


      variables--manufacture your products at the


      extremes of CMVs and in vivo performance and, if


      you have that information, you will have a very



      sound rationale for coming in with appropriate


      dissolution specifications.


                Finally, in this guidance there is a


      discussion of how do you compare dissolution


      profiles of two products.  One of the approaches


      that I recommend is known as the f2 or the


      similarity factor which essentially looks at the


      differences in dissolution at each time point, with


      a range of 0-100.  An f2 of 50 or greater than 50


      indicates similarity of the dissolution profiles.


      As we have said in that guidance, dissolution


      specifications are established in consultation with


      Biopharmaceutics and the CMC review staff. The


      general bioavailability/bioequivalence guidance


      summary, again limited only to dissolution


      considerations, we have a section in there that


      talks about what should be submitted in an NDA or


      an ANDA in terms of a dissolution method.  There


      should be a dissolution method development report


      for an NDA, new drug application.  It should


      contain a pH solubility profile of the drug


      substance; dissolution profiles generated at



      different agitation speeds; and dissolution


      profiles generated on all strength in at least


      three dissolution media.  Essentially you want to


      see the in vitro performance of your product over a


      variety of conditions, including different media


      and different agitation; and select the agitation


      speed and medium that provides adequate


      discriminating ability, taking into account all the


      available in vitro and in vivo data.


                For ANDAs, abbreviated new drug


      applications, the guidance states that one should


      start with an appropriate USP method if it is


      there, in the USP.  For some reason, if it is not


      there for this product, then if the FDA method is


      publicly available, utilize that.  If that is not


      available, also publicly available, then submit the


      dissolution method development report, as described


      above for a new drug application.


                Again, for modified-release products for


      ANDAs the dissolution profiles use the appropriate


      USP method, if available, otherwise use the FDA


      method for the reference listed drug if available. 



      In addition, and I think this is probably because


      you could have for a generic similar or different


      release mechanisms, so additional dissolution data


      in three different media.


                Now switching to the IVIVC guidance which


      is, you know, in vivo/in vitro correlation for


      modified-release products, again from my


      perspective, this is a very useful guidance also.


      The main purpose of this guidance was to provide an


      outline for waiver of bioequivalency studies for


      modified-release products if one was able to


      establish an in vivo/in vitro correlation, a


      quantitative correlation.


                The guidance defines correlation in


      different categories, A, B, C and D.  Level A


      correlation is most quantitative, and I have listed


      in my presentation just the level A discussion.


      Level A correlation is supposed to be a


      point-to-point relationship between the in vitro


      dissolution and the in vivo input rate of the drug


      from the dosage form.  Usually this is a two-stage


      process, meaning that you take your dissolution



      data, convert that into dissolution rate, and you


      take your in vivo data and convert that into


      absorption rate and correlate the two.  Generally,


      this relationship is linear but non-linear


      relationship is also acceptable provided it is


      adequately characterized.


                So, this is an example of how level A


      IVIVC would look.  On the Y axis you have percent


      of drug absorbed and on the X axis is the percent


      of drug dissolved; your linear relationship over


      the range and this establishes your correlation.


      For the purpose of obtaining biowaivers, you need


      validation of this level A correlation.  From the


      point of view of setting dissolution


      specifications, that level of validation is not


      necessary, and I will get into that subsequently in


      my examples.


                In the IVIVC guidance for modified-release


      products we have some general concepts laid out for


      what the dissolution specification should mean.


      Ideally, as we say in the guidance, all lots within


      the lower and upper limit of the specifications



      should be bioequivalent.  At the minimum, those


      lots should be bioequivalent to the clinical trials


      lots or an appropriate reference standard chosen by


      the agency.  In other words, you have your


      reference performance and the upper limit should be


      similar to the reference and the lower limit should


      be similar to the reference.  Ideally, the extremes


      should be bioequivalent.


                Some further considerations are that


      variability alone should no longer be a primary


      consideration in setting specifications for


      modified-release products.  Specifications wider


      than 20 percent are acceptable only when evidence


      is submitted that lots with mean dissolution


      profiles that are allowed by the upper and lower


      limits are bioequivalent.  In other words, you can


      have specifications wider than 20 percent if you


      have a correlation, a quantitative correlation.


                If you don't have an IVIVC and you want to


      set dissolution specifications for modified-release


      products, these are some of the characteristics of


      what the data should be.  The profile should have



      at least three time points.  The last time point


      should be the time where 80 percent of the claimed


      labeled amount is dissolved.  Specifications are


      set to pass at stage 2, meaning that there are 12


      dosage units.


                As I mentioned a while ago, for setting


      dissolution specifications with the IVIVC, external


      validation is not required and, as I already


      mentioned, wider specifications based on what the


      correlation predicts can be done.


                This is graphically presenting that.  On


      the left panel you see that in the middle is the


      performance of your product, the variability around


      the mean dissolution profiles.  The blue line is


      the upper limit of the specification.  The red line


      or orange line is the lower limit of the


      specification.  You take that data using your in


      vitro/in vivo correlation model.  You predict the


      plasma concentration based on the two limits.


                On the right panel, the diamonds are the


      actual blood levels, the predicted blood levels at


      upper and lower limit, and the predicted level for



      Cmax and AUC should not be greater than 20 percent.


      Back in '97, what we could come up with was setting


      the consideration based on the mean difference.


      So, the upper and lower limit would not differ on


      the mean AUC and Cmax by 20 percent.  We could not


      build into this consideration the variability


      aspects and, as we have already heard in an earlier


      presentation today, that is an opportunity for


      improvement for future consideration.


                Switching gears, I am going to quickly


      tell you about what the SUPAC guidances mean as far


      as immediate-release and modified-release products.


      There are also a few guidances that came out


      subsequent to the issuance of the SUPAC in 1997,


      which is called equipment addendum, FDAMA and the


      changes approved to an NDA or ANDA guidance in


      2000.  Again, I am going to try to capture this


      very quickly.


                Conceptually speaking, these guidances


      identify what are the changes or what are the


      variables that are covered in terms of


      manufacturing considerations.  The level of changes



      for these variables, what are they?  They are


      defined; and then how do you deal with that?


                So, the variables covered in this


      guidance, manufacturing related, are composition


      and components.  For excipients it is


      non-release-controlling as well as


      release-controlling.  The non-release-controlling


      aspect is what is the part of the SUPAC-IR


      guidance.  That is taken as it is into the SUPAC-MR


      guidance and then what is added is the


      considerations for release-controlling excipients.


      Other variables covered are site, batch size,


      meaning scale-up and scale-down, manufacturing


      equipment and manufacturing process.


                I am going to take you through only one


      set of variables here and show you how the levels


      are defined and what are the related tests


      recommended and what are the related filing




                Essentially, the idea is this, the


      guidance has defined the level of change into three


      categories, level 1 is the minor change; level 2 is



      the moderate change; and level 3 is the major


      change.  So, moderate could have an in vivo impact


      on level 3 or major changes likely to have an in


      vivo effect.


                Related to those changes, the tests go


      along with them in terms of document evidence.  The


      lowest level, level 1, would usually require only


      application of compendia tests and stability data.


      Level 2 change would require extensive in vitro


      dissolution and release data.  That typically means


      that for immediate-release products you require


      profile comparison in five different media.  Then,


      for modified-release you need profile comparison in


      three different media.  Level 3 is the most


      significant change and that will be allowed only if


      you have an in vivo bioequivalency study or you had


      established in vitro/in vivo correlation.


                The filing requirements, again going from


      minimal to most which is annual report, changes


      being effected supplement, or prior approval


      supplement.  In the subsequent discussion I will


      just focus on the first two bullets, which is level



      of change and the tests.  I am not going to touch


      filing documentation at all.


                Here is an example of how the guidances


      break down changes into different levels.  For


      SUPAC-IR excipient levels excipients are listed for


      level 1 change, level 2 and level 3.  If you look


      at glidant, for example, for talc, plus/minus one


      percent change is allowed.  If you look at the top


      of the right-hand column, it is percent change


      weight of the change of the excipient over the


      weight of the total unit.  For talc it is


      plus/minus one percent.  Other glidants would be


      plus/minus 0.1 percent.  So, that is the lower


      limit of change, plus/minus 0.1 for talc.  If you


      look at filler, for example, it is also plus/minus


      five percent change.  So, this defines level 1


      change, minimal change.


                If you go to level 2 the ranges double.


      So, you go from plus/minus 0.2 to plus/minus 10


      percent.  Anything beyond 10 percent is considered


      a level 3 change.  Again, this is for


      non-release-controlling excipients.


                If you go down to release-controlling


      excipients for modified-release products, the


      criteria are more stringent.  Now, the change is



      measured as a percentage of the total


      release-controlling excipients and not the total


      dosage form unit so your denominator is a smaller


      number.  The percentage allowed is smaller for


      release-controlling excipients.


                For level 1 change, that means that the


      total additive effect of all release-controlling


      excipients should not be more than plus/minus 5


      percent.  Level 2 should not be plus/minus 10


      percent.  Changes beyond plus/minus 10 percent are


      considered level 3.


                So, this is a summary of what we have


      recommended in the SUPAC-IR and MR guidances.


      These guidances define the tests; filing document


      recommendations; level of changes in composition


      and components, release-controlling and non-release


      controlling excipients; site changes; batch size


      changes; equipment and process changes.


                The following changes either need a bio.



      study or an established IVIVC:  Level 3


      release-controlling and level 3 non-release


      controlling change; level 2 release-controlling


      change for NTR drugs; and level 3 site change and


      level 3 process change.  All of those changes,


      meaning level 2 changes, would require comparable


      dissolution documentation, meaning, as I said,


      profile comparison in several media.


                As I mentioned in the title slide for


      these guidances, the equipment addendum came out a


      little later and there we identified equipment by


      class and subclass for all major unit operations,


      and a change to a different class is generally


      considered a change in design and principle.  So,


      if you have equipment changes within the same


      design and operating principle it is considered a


      minor change.  If you go to a different design and


      principle it is a major change.  Finally, the


      changes guidance allows for multiple different


      level changes.  As we all know, these changes do


      not occur only one at a time; it is a composite of


      changes for any change.  So, if you have, say,



      several level 1 changes and one level 2 change for


      your new product you would be held to the most


      restrictive individual change of level 2, and


      whatever requirements go with that level of change.


                So, that was a quick overview of the


      guidances.  These documents are available on the


      web, and if you have any questions please look them


      up.  Let me switch gears here and take you through


      some examples of the way the specifications are




                But before that, let me share with you


      generally what we see in an application in terms of


      information available for setting specifications.


      The data that are available for a typical


      immediate-release product in an NDA are as follows:


      Dissolution results under a variety of agitation


      and media conditions.  Then typically what we see


      are several methods.  One method is selected by the


      sponsor which generally provides you with a rapid


      dissolution profile.  Using that method, we have


      data of 6-12 units and that is the limit of data we


      have for any given lot.  So, that is the range of



      variability that you would typically see for a


      particular lot.  Using that method, you have


      dissolution data from the bio. batch, the batch on


      which bioavailability has been characterized, plus


      few to several production lots under this


      condition.  Again, as I said, these batches are


      usually in very large quantities, hundreds of


      thousands to million units.  We see the data on


      6-12 units.


                Then we do have a lot of bioavailability


      data on this product.  Actually, bioavailability,


      relative bioavailability, bioequivalency trials and


      dissolution data of lots used in efficacy trials


      and stability data.  So, we look at all this


      information and try and come up with a meaningful




                What do we do when we consider setting


      specifications?  These are the factors that are


      taken into consideration when setting specs. for an


      immediate-release product.  The in vivo behavior of


      a drug product, particularly how rapidly the drug


      is absorbed and an indicator for that is Tlag time



      or what is the Tmax of your product.  Since the


      issuance of the BCS guidance we look at the


      permeability data very closely.  In vivo


      permeability would be based on mass balance studies


      as well as absolute bioavailability studies and


      that, in my mind, is the gold standard by which you


      define whether a drug is highly permeable.  If it


      is quantitatively absorbed, then you say this high


      permeability, along with your high solubility data,


      puts the product into BCS class 1.  Then that


      carries its own benefits.  I have an example of


      that to show you a little later.


                That is what one pays attention to, in


      vivo behavior of the drug product from a


      bioavailability point of view.  We look at


      dissolution behavior across all conditions in vitro


      and then we try to come up with an adequately


      discriminating method, taking all this data into


      consideration based on any quantitative or


      qualitative in vitro inference.


                What is very helpful for evaluation of an


      NDA is if you have data like this where a solid



      dosage form in vivo is compared to something that


      is even more rapidly dissolving, meaning your solid


      dosage form's performance in vivo with respect to,


      like, a solution.  If we have this data, this tells


      us a lot about what is the in vivo dissolution of


      your solid dosage form and that can help us


      evaluate the in vitro considerations for setting


      specifications for that product.  So, this can


      guide how discriminating the in vitro method needs


      to be.


                As I said, we look at all the available


      dissolution data and pay particular attention to


      the lots that have in vivo data, and then discuss


      with our chemist colleagues about what is available


      in the stability domain, the data there and the


      specifications we are considering.  If we see a


      significant change or time with stability


      performance, that will have to be resolved by a


      bioequivalency study.


                Possible outcomes in terms of setting


      specifications, one is everybody is happy.


      Sufficient data are submitted and specs are



      finalized.  It is possible that insufficient data


      are submitted.  Based on the product's indication,


      the product needs to be approved with reset interim


      specs.  We agree with the sponsor what additional


      data needs to be generated.  We agree upon a


      time-line.  We evaluate the specs and we finalize


      the specs.  In the rare instance where there is


      insufficient data submitted--I have not seen this


      happen in my lifetime where we have withheld


      approval for a drug product because of insufficient


      dissolution data.  At the least, we will set specs


      on the clinical trial product.  So, if insufficient


      data are submitted and specs can't be finalized


      even including interim specs, then we have to


      resolve that prior to approval.


                Now let me take you through some specific


      examples, starting with simple to a little bit more


      complex.  This is an immediate-release drug product


      A.  The drug is highly soluble over the pH range of


      1.2-6.8, or 6.9 in this case.  Based on the


      bioavailability and the in vitro permeability, we


      established that the drug is highly permeable.  So,



      we have high solubility, high permeability criteria


      met.  The drug product is rapidly dissolving over


      the pH range of 1.2-6.8.  So, we have seen this.


      We are sure of these characteristics and we say


      okay, this is BCS class 1.


                We have dissolution results of the


      bioavailability lot and the clinical lot so all


      that data is utilized in setting the


      specifications.  There was stability data also


      available that was taken into consideration.  It


      turned out to be a straightforward case.  The


      sponsor's proposal was that they use a USP 1


      apparatus at 100 rpm in 900 ml 0.1 normal


      hydrochloric acid; specs of 80 percent in 30


      minutes.  We agreed with the sponsor.


                Just as a note, Ajaz and I didn't exchange


      notes beforehand but in this case the sponsor chose


      apparatus 1 to avoid coning effect.  Ajaz had an


      example from the Canadian database where that was


      the reason why you saw a big investigator


      difference compared to the reference, but the in


      vivo data turned out to be fine.


                Another example for an immediate-release


      drug product, product B, the drug is a free base


      with 2 pKs of 5.4 and 7.2.  It is highly soluble at



      pH 1 but it is practically insoluble at pH 7, and


      the solubility drops sharply between pH 4-5.  I


      have a graph that shows that clearly.  The drug is


      absorbed slowly, at Tmax ranging from 3-5 hours.


      The half-life is long, 45 hours.  It is not highly


      permeable.  The fraction absorbed is around 0.75.


                So, what do we do with this?  This is the


      dissolution behavior across the pH ranges.  As you


      see, below pH 5, which is the third curve from the


      top, dissolution starts dropping rapidly as the pH


      increases.  The sponsor chose the dissolution


      method at pH 5, and showed that the clinical and


      to-be-marketed formulations had similar profiles.


                This is what that comparison is at pH 5.


      We had bioequivalency data on these two


      formulations and that turned out to be clearly


      bioinequivalent in vivo for the test, meaning that


      to-be-marketed product showed a clear difference in


      Cmax.  The Cmax was 17 percent lower.  We



      interacted with the sponsor and they optimized


      their method to come up with an adequate


      discrimination condition to evaluate this


      formulation further.


                This is what they came up with, 5 percent


      volume Tween 80 and the same two formulations that


      were clearly bioinequivalent in vivo, they were


      able to identify their in vitro performance and


      show that, indeed, they were different.  This was


      verified further by taking the two formulations


      that were bioinequivalent in vivo and the method


      showed that they were similar in vitro.


                This was the availability of dissolution


      data across several batches.  All I want to point


      out to you is that, as I said, dissolution data for


      different batches, from 6 units, mean and range is


      available and if we look at the right-hand column,


      the lowest range is 86 percent.


                Taking all that data into consideration,


      the sponsor proposed the specification with


      apparatus 2 at 50 rpm and 1000 ml to Tween 80 in


      water; Q of 75 percent in 45 minutes.  We



      recommended no changes in the condition but a Q of


      80 percent in 45 minutes.  Here is an example of


      availability of in vivo data optimizing the




                The final example I have is for a drug


      product, a modified-release drug product with in


      vivo/in vivo correlation.  For this drug product a


      level A correlation was established.  Correlation


      was obtained from in vivo data from 6 different


      studies, and the media consisted of pH 1.5 for the


      first 1.5 hours and then pH 6.8 for the remainder


      of the 24 hours.  This is a once a day product.


                These are the results.  I think this was


      excellent work on the sponsor's part.  We worked


      with them and we were very happy to figure out the


      specs with them.  Look at the hatched region.  That


      is the observed range of dissolution data.  That is


      the extent of variability across the entire


      manufacturing experience for this sponsor.  So, the


      hatched area is the dissolution variability,


      dissolution range the product showed in vitro.  The


      specs we agreed upon are the two dotted lines above



      that hatched region.  So, those were the


      specifications proposed and we agreed with them.


                The best part is that if you look at the


      third level of curves, which are the topmost dotted


      lines, the topmost and the bottom, those are the


      predicted in vitro dissolution behaviors of two


      formulations that would be comparable in vivo.  So,


      the specifications were set within the limits of


      what products would be bioequivalent, so a good


      IVIVC that could lead to meaningful specifications.


                Now let me conclude with some personal


      comments on opportunities for improvement.  Before


      I get into my own suggestions, I want to cite this


      article that Ajaz already mentioned from Dr. Janet


      Woodcock, a clinician who has written beautifully


      on pharmaceutical quality.  I am just going to cite


      two quotations out of this article.  I mean, I can


      stand here and tell you a great deal about all the


      complexities involved in clinical trials but I


      think Dr. Woodcock has summarized this very well in


      this first bullet, which is, as she says, for the


      purposes of clinical use, the established drug



      quality attributes are generally adequate because


      they achieve much tighter control of the level of


      variability than could be detected in patients


      without extensive study.


                These are part of all the variabilities,


      specially manufacturing variability.  It can be


      done but it is a difficult task and it would be


      very extensive, and that is not the paradigm


      currently used.


                But maybe even more important, as she


      points out here in the very second line of the


      previous quotation, in contrast, for regulatory and


      manufacturing processes, the lack of detailed


      understanding of the real-world importance of


      quality attributes is a serious problem, leading to


      many disputes that might be resolved easily were


      relevant information available on the relationships


      between various quality parameters and clinical


      performance.  I personally couldn't agree more with


      that concluding comment.


                So, clinical performance, if I were to


      dissect that further--everybody talks about



      variability and this is my share of what are the


      different types of variability in therapy.  You


      start with manufacturing variability, then you have


      variability associated with the drug exposure and


      then you have variability associated with the drug


      response.  You have compliance issues.  You know, a


      lot of people can actually add more bullets to this


      and provide a complete picture of how complex the


      system is when a patient is being treated in vivo.


                But I have taken a shot at just making a


      point on exposure-related variability and


      manufacturing aspects associated with that.  The


      next table is a snapshot.  We have an internal BCS


      database of almost 200 NDAs.  That is in the


      process of being audited and we hope to publish


      that soon.  So, what I requested Dr. Uppoor to do


      is to randomly select a few drugs and prepare a


      table that would show variability in AUC and Cmax


      and the exposure parameters of different BCS




                Again, this is tentative because this is


      not fully audited so that is why I have starts in



      this table.  This is BCS class 1, 2, 3 and 4 across


      the top horizontal line.  You have the permeability


      associated with the AUC parameter and the Cmax


      parameter for these products.  As you can see,


      staring with class 1, we have variability in the


      range of 17 to about 24 percent.  Class 3 shows


      maximum in vivo variability.


                So, if I want to take this tentative class


      information further, the point I want to make--the


      numbers might be off when we have the actual


      publication coming out, but this is the point I


      want to make, that if I assume that the clinical


      trial formulation for this product was


      optimized--if it is not optimized, I think it is in


      the interest of the sponsor to optimize that so


      that even a little bit of manufacturing variability


      does not reflect in the in vivo performance at


      least from a drug exposure point of view.  But


      assuming that this formulation is optimized, even


      for BCS class 1 products you do see a decent amount


      of variability in vivo.  Again, this is reflecting


      how the drug is handled by an individual and the



      variability of handling that across individuals.


      This information can be utilized by a sponsor to


      come up with rational specs.


                These are some of my thoughts in terms of


      opportunities for improvement.  The first point is


      nothing earth-shattering but I still think it is a


      point that has to be made, to select an appropriate


      dissolution method based on physicochemical in


      vitro and in vivo characteristics of the drug and


      the drug product.


                It would be useful to have an estimate of


      in vitro variability for low solubility and low


      permeability.  Estimate of variability of lots used


      in pivotal efficacy trials would facilitate setting


      of rational specifications.  For modified-release


      products estimate the in vitro release


      variability--the example I showed where if you had


      a handle on the variability across your entire


      manufacturing process, then you can bring that into


      setting a meaningful specification.  As I already


      mentioned, right now the IVIVC current guideline is


      based on the limit mean estimates only and if you



      can build in the variability aspect and in vivo


      performance based on estimate of mean as well as


      variability, I think that would lead to more


      rational specs, maybe even wider specs compared to


      what we are doing now.


                The things that I see in the near future


      are new technologies like PAT.  Hopefully, it can


      provide in vitro and in vivo relationships based on


      the performance of an individual dosage form unit.


      I mean, this would be a non-destructive method.


      You would be able to assess the dissolution


      performance of a unit without breaking it up and


      then you would administer that to an individual and


      you would get that individual's exposure parameters


      so you would have correlation relationship on an


      individual dosage unit form in an individual


      patient taking it.  I think that would be a very


      powerful set of data to set meaningful




                We are getting more and more complex


      products like drug eluting stents and liposomes.


      For these complex dosage forms I think it would be



      essential to study drug elution, drug release using


      mechanistic models and new techniques in imaging


      and fluid dynamics.  Hopefully, future


      specifications will be based on in vitro mean and


      variability estimates.


                Moving from a science point to a process


      point--I didn't know our good friend Dr. Chuck


      Hoiber [ph.] would be here but this is in those


      days when Chuck and I were on the same floor and we


      started implementing this which is that from the


      process point of view there are also a lot of


      opportunities to optimize setting of specifications


      and that, from my perspective, is come and meet


      with us early.  A meeting would be useful if you


      have good quantity and quality of data.  As we have


      done on several occasions, we have had separate end


      of Phase II meetings with CMC Biopharmaceutics and


      colleagues on our side and the industry, going over


      the development plan and that has led to a quicker


      review and arriving at meaningful specifications at


      the time of NDA approval.


                Finally, I do personally believe that good



      homework will always bring dividends.  If you have


      good data, please share them with us and we will


      work with you to come out with rational


      specifications.  Thank you.


                     Questions by Committee Members


                DR. COONEY:  Thank you very much.  Some


      questions from the committee?  Ken?


                DR. MORRIS:  Two things.  I was a little


      surprised to see the high variability with BCS 3.


      In principle, you would expect BCS 3 to be a good


      candidate for waiver because, as long as your


      driving force doesn't change, you would expect that


      the absorption is rate limiting and falls into the


      same basic concept as 1.


                DR. MEHTA:  That is a very good


      observation.  We are looking at the data carefully


      ourselves, but I think it is maybe one product that




                DR. MORRIS:  Driving the variability?


                DR. MEHTA:  Yes.


                DR. MORRIS:  Or is it that the absorption


      itself is just variable?


                DR. MEHTA:  Again, we can think about it


      but it is a question if you have a class 3 high


      solubility, low permeability drug and if low



      permeability is not leading to the same conditions


      in vivo that is going to take away some of your


      high solubility benefit.


                DR. MORRIS:  Not the same conditions on


      which side?  Are you talking about in the gut?


                DR. MEHTA:  Yes.


                DR. HUSSAIN:  Sorry, if I may, I think one


      of the challenges is that this was always a


      question when we were deliberating the BCS


      guideline, high solubility.  But the in vivo


      dissolution actually is more sensitive for low


      permeability drugs and we actually have published


      on this with Lawrence--


                DR. MORRIS:  Right.


                DR. HUSSAIN:  So, people often say this is


      high solubility so dissolution is not rate limiting


      but in vivo dissolution behavior is quite complex.


      Plus, you add site-specific absorption of these


      compounds that adds to all the sources of





                DR. MORRIS:  Right.  I guess that is my


      point in a sense.  Shouldn't the compounds be


      segregated into site-specific and passive absorbed


      compounds to really do a valid experiment?


                DR. HUSSAIN:  I fully agree with you.  We


      came up with the classification system and those


      four classes are beautiful but there is nothing


      that black and white.  Greater than 90 percent


      permeability, highly permeable, but there is a


      gradation of that and, you know, we have to take


      that into account.  You know, there are, like,


      windows of absorption.  So, we need to subclassify


      those four classes and then come up, you know, with




                DR. MORRIS:  Yes, but it would be nice if


      you could identify some more waiver-worthy classes.


                DR. HUSSAIN:  Yes.


                DR. MORRIS:  Just another quick comment is


      that I am sure it won't surprise you but, you know,


      with the general BA/BE guidance people, because of


      what is in the guidance, are actually doing pH



      solubility profiles of non-ionizable compounds.




                DR. MEHTA:  That is taking us too




                DR. COONEY:  Paul?


                DR. FACKLER:  I have a comment and a


      question.  The comment had to do with the slide


      where you suggested there might be about 20 percent


      variability for even BCS class 1 compounds.  I


      would suggest that that is vastly understated, that


      the variability is much higher than that because I


      am guessing that your data comes from


      bioequivalence studies where all of the subjects


      take exactly the same amount of water, the same


      amount of food.  None of them are BMI greater than


      a particular number.  If they are old studies they


      were all men.  I would just say that in the general


      population with the way pharmaceuticals are really


      taken--some people run three miles, come home and


      then swallow their tablets; some people roll out of


      bed and swallow them without water--the variability


      even for class 1 is significantly higher than 20



      percent.  But it is just my opinion.


                DR. MEHTA:  That would just add to the


      thought I had which is, you know, use that


      information to evaluate your in vitro


      specifications.  That will help you.


                DR. FACKLER:  The question I had had to do


      with that same chart where you looked at 17 drugs


      that were randomly pulled out of the pool of 200.


      It was interesting to see that the class 3 is N


      equals 7.  I am just wondering if the distribution


      of these 17 in any way represents the distribution


      of the 200 drugs.


                DR. MEHTA:  I don't think so.  The whole


      idea was to see if we can get a handle on what is


      the exposure variability for these products.  A few


      years ago I presented this database at one of the


      APS workshops what was surprising is that we saw a


      lot of NDAs falling into class 4 category.  If it


      is a class 4, then you would see very few drugs,


      low solubility, low permeability.  You know, they


      would fall out of drug development.  But, as I


      mentioned a little while ago, the way classifying



      we have created these four classes, 90 percent of


      data goes in class 1 over this 85 percent


      absorbed--you know, it is still low permeability.


      So, I don't think when we come out with this


      information, all audited, that there is going to be


      a majority of them falling in class 3.  I don't


      think so.


                DR. COONEY:  When you presented the table


      of the 17 samples, your intent is to expand that?


      This is just a piece of work in progress?


                DR. MEHTA:  Yes, very much so.


                DR. COONEY:  So, the idea is to really


      address the question that was just asked, that is,


      to have an analysis that is representative of that


      whole set?


                DR. MEHTA:  Yes.  I mean, right now we are


      going through each drug and making sure, to our


      level best effort, that the data available


      classifies that drug product in the appropriate


      class.  We have the information put together and


      now it is like careful auditing going on.


                DR. COONEY:  Good.  Marvin?


                DR. MEYER:  I did come up with a couple of


      questions.  It always bothered me that the BCS


      system had this quadrant drawn and then the lines



      kind of floated depending on how you wanted to


      define high and low--


                DR. MEHTA:  No, it is rigid right now.


                DR. MEYER:  I know it is rigid but the


      rigidness was arbitrary.




                It is arbitrarily rigid.


                DR. HUSSAIN:  I will defend it tomorrow;


      don't worry!


                DR. MEYER:  Okay.


                DR. MEHTA:  We started out with a


      conservative position and now with the availability


      of more data we want to expand that rationally with


      proper evidence.


                DR. MEYER:  It also bothered me that this


      permeability goes all the way from a very rigorous


      intubation of humans to a K2 cell to looking at


      Tmax.  So, how it is defined or determined can be


      another source of variability in where it falls in



      this rigorously arbitrary quadrant.  So, I think


      that may be a reason in part why the class 3 seemed


      to be more variable than 2.  One drug in that would


      have expanded the range.


                DR. MEHTA:  That is just the way those


      drugs got pulled out.  That is why I have that


      range.  That may not be reflective of what it is.


      I don't want to take up too much time, but we look


      at permeability assessment now very carefully and,


      in my mind, hopefully, if we have data on the NDA


      side, which is mass balance data and


      bioavailability data, that is the maximum way in


      terms of assessing, you know, whether the drug is


      90 percent absorbed or not.  Sometimes we have an


      issue with that.  Then we utilize the in vitro


      methods for that decision.


                DR. MEYER:  One last question.  Do you


      feel that the f2 test has been rigorously




                DR. MEHTA:  A good question, Marv.




                There are people in the audience that--


                DR. MEYER:  Do you feel--do you feel it


      has been rigorously evaluated so it will detect


      differences when they should be detected and will



      allow passage when it should be allowed?


                DR. MEHTA:  Well, I mean we do state in


      our guidances under what conditions this approach


      should be employed.  You know, if your variability


      is very high in dissolution on each formulation


      this is not the right way of comparing those


      profiles so then you need to get into more complex


      assessment, and all that.  If it is done properly,


      yes, I do myself.


                DR. COONEY:  Pat?


                DR. DELUCA:  In the BA/BE guidance summary


      for modified-release products you are saying that


      they should profile using at least three other


      dissolution media and water.  Why do you need three


      others if you have a correlation?


                DR. MEHTA:  No, it doesn't say that there


      is a correlation.  This is just a question--well,


      usually correlation is release formulation




                DR. HUSSAIN:  It is just for that product.




                DR. MEHTA:  It is right now.


                DR. COONEY:  Nozer?


                DR. SINGPURWALLA:  When you don't


      understand something you start asking technical







                You showed a picture of linear correlation


      long ago, one of your early slides--


                DR. MEHTA:  Yes, level A correlation.


                DR. SINGPURWALLA:  Level A correlation.  I


      have two comments.  The first is that you are


      looking for relationships between the percent of


      drug dissolved and the percent of drug absorbed so


      correlation only measures linear relationships.


      You may have dependence which may be not linear but


      still of value to you, but correlation does not


      measure that.  So, I just want to say that as a




                The second more serious comment is that


      that particular correlation misses the time index. 



      What you really need is a third axis also showing


      the time at which all these happen.  For you to do


      that, you want to look at these two as what we


      would call stochastic processes or time series, and


      you want to cross-correlate the two time series.


      So, if you want to improvise on that particular


      theme, you may want to look not at correlation but


      what I would consider cross-correlation where you


      also introduce the time axis.  That is the only


      comment I want to make.


                DR. MEHTA:  Thank you.  That is helpful.


                DR. SINGPURWALLA:  Do you want to


      challenge me now?


                DR. MEHTA:  No, I didn't say that.


                DR. COONEY:  Are there any other questions


      at this point?


                [No response]


                Thank you.  The next presentation will be


      by Dr. Shah establishing dissolution




                Establishing Dissolution Specifications:


                            Current Practice


                DR. SHAH:  Good morning.  Mehul gave a


      nice overview on the BCS guidance and other


      guidances which are used in setting dissolution



      specification from a biopharmaceutics perspective.


      My job today is to cover the CMC aspects of setting


      the dissolution specifications.  In this


      presentation I am going to start with an overview


      of the current practice, and in that overview I am


      going to cover the CMC assessment and bring in some


      of the ICH Q6A principles, how we evaluate the ICH


      Q6A principles in our CMC assessment, and then I


      would like to talk about a case study example for


      extended-release oral suspension and in that


      example I am going to cover the drug development


      strategy by the applicant, the dissolution results


      obtained based on that development strategy, then


      what we identified as critical issues, followed by


      our recommendations and based on those


      recommendations, what were the improvements


      implemented by the applicant and what was the


      outcome out of those implementations.  I would like


      to end my talk with some concluding remarks based



      on this example as well as general remarks in that




                As Mehul suggested in his presentation, I


      want to reemphasize that establishing dissolution


      specification is a shared responsibility between


      the Office of New Drug Chemistry and the Office of


      Clinical Pharmacology and Biopharmaceutics.


                In the next three slides I have presented


      the considerations that should be given during that


      development, as well as the focus of CMC assessment


      during the NDA review, and what forms the basis of


      setting the dissolution specifications from CMC




                As I have pointed out here, it is a known


      fact that physicochemical properties of the


      formulation components, such as drug substance and


      other excipients, such as the solubility, pKa,


      particle size distribution, polymorphic forms and


      there may be some others, have a significant effect


      on the dissolution.  The physicochemical properties


      impact the processibility of the formulation


      components, as well they may affect also the



      safety, efficacy and stability of the drug product.


      In addition to that, the manufacturing processes,


      especially those having the potential to influence


      the release profile of the drug substance also


      should be studied during the development.  And, the


      control strategy of the critical process parameters


      and in-process testing also should be developed


      during the development, and those are the focuses


      of the CMC assessment.


                During the drug development one should


      expect that there should be a relationship of


      in-process testing to the critical quality


      attributes, such as dissolution of the drug


      product.  Some of the in-process testing that may


      be carried out might be particle size distribution;


      release rate; and the compression force, tablet


      hardness and friability in the case of solid oral


      dosage form.


                In addition, during the CMC assessment we


      focus also on the development and validation


      aspects of the proposed in vitro dissolution


      method.  Cindy already covered some of these



      aspects in terms of how the methodologies are being


      developed and what are the validation criteria that


      need to be covered, especially pertaining to


      specificity, linearity, accuracy, precision,


      ruggedness, etc.  In addition, we also focus on the


      release time point intervals and what should be the


      adequate tim point intervals.


                Once we have this information we need to


      see or need to provide during development, as well


      as the NDA submission data, what is the


      relationship between the in vitro dissolution data


      from development, clinical, bio. and primary


      stability batches, and also identify a discerning


      trend on storage.  We also evaluate the proposed


      shelf-life of the drug product on the basis of the


      stability data analysis of dissolution, as well as


      other drug product attributes.


                In the end, it is in coordination with


      Office of Clinical Pharmacology and Pharmaceutics


      that appropriate dissolution specifications are


      recommended and these specifications are reflective


      of the dissolution data from various batches



      including clinical, bio., stability and other




                In terms of the ICH Q6A document, ICH Q6A


      discusses the potential relevance of particle size,


      polymorphic content and polymorphic changes, and


      how it affects the dissolution.


                Here I have these three decision trees


      just for reference.  I just wanted to point out


      that CMC assessment very well integrates these


      principles in our assessment for the quality


      assessment of the drug product.  This is about the


      particle size distribution and the decision tree


      guides you on how to set acceptance criteria.


                This is in terms of polymorphic content.


      That also guides you on how to set acceptance


      criteria.  The next one is how to set the


      polymorphic change acceptance criteria in the drug




                Now I would like to focus on the case


      study example for extended-release oral suspension


      for the remainder of my talk.


                Let me give you just some background. 



      This was submitted as a 505(b)(2) application.  As


      a result, there was no clinical trial required


      because the safety and efficacy of the proposed


      active ingredients for the proposed indication was


      established through immediate-release products


      available under the tentative OTC monograph for the


      same indication.  The proposed dose was a single


      dose given every 12 hours to patients 6 years of


      age or older.  That was equivalent to the nominal


      OTC monograph which was given every 6 hours twice.


                In terms of the formula, the drug product


      contained two different active ingredients, and I


      will call them drug substance 1 and drug substance


      2.  For proprietary reasons, most of the data I am


      going to discuss here are well concealed and they


      are masked but the data are real.  Drug substance 1


      is anchored to a drug carrier support and coated


      separately with semipermeable polymer to prevent


      dose dumping and to impart the extended-release


      profile.  Drug substance 2 binds the drug carrier


      support in situ during the manufacturing process,


      but it is not coated.  Both active ingredients,



      along with other excipients, are suspended in


      aqueous solution.


                The concerns we had here arise from the


      safety implications due to the potential dose


      dumping, and efficacy implications due to


      insufficient rate and the extent of release of the


      actives.  These concerns were brought to the


      applicant's attention during the end of Phase II


      meeting as well as pre-NDA meetings, and they were


      very mindful of those two concerns.


                This was the strategy adopted by the


      applicant in the beginning.  They wanted to


      demonstrate bioavailability of the drug product


      formulas, and that was coated with 6 percent


      coating of drug substance 1, to a reference drug


      which as an immediate-release solution, and it was


      containing the same two active ingredients.  They


      had no other choice but to start with the


      immediate-release solution because there was no


      existing extended-release product containing these


      two ingredients.


                Their plan was to formulate three



      experimental drug formulations, each differing only


      by the coating level of semipermeable polymer on


      drug substance 1.  They were low coating, for


      example, 2 percent; medium coating, example, 5.5


      percent; and high with 9 percent coating on drug


      substance 1.  They labeled them as fast-release


      solution, intermediate-release formulation and


      slow-release formulation.  The approach was to


      establish IVIVC for each active among these three


      experimental formulations, and establish


      dissolution specifications for both actives based


      on generated dissolution profiles from the slow-


      and fast-release drug product formulations.


                In the NDA the data submitted include five


      formulations of the drug product containing drug


      substance 1 coated with varying levels of


      semipermeable polymer, 2 percent, 5.5 percent 9


      percent, as well as 6 percent and 10 percent.  They


      performed the following PK studies, multi dose


      bioavailability studies with immediate-release


      solution and single dose food effect study


      containing 6 percent polymer coating, and single



      dose IVIVC study containing three formulations, 2


      percent, 5.5 percent and 9 percent polymer coating.


      In support, there were PK results from four batches


      and stability results from four PK and five


      stability batches.


                Based on these PK studies, these were the


      applicant's claims, that level A IVIVC was


      established for both actives of the ER suspension.


      The mean individual level A IVIVC models for drug


      substance 2 met the FDA validation criteria and, in


      their opinion, it can be used for setting


      dissolution specifications and biowaivers.


                The mean and individual level A IVIVC


      models for drug substance 1, which is coated,


      failed the FDA validation criteria in that the


      predicted values had a larger error than


      recommended.  However, if the dissolution criteria


      remain within dissolution profiles tested in IVIVC,


      they proposed that the drug substance 1 results can


      serve as a mapping study for the formulations.


                Now let's see what was the agency's


      finding in terms of the PK results.  On the



      bioavailability and food effect studies, which was


      the 6 percent coating of drug substance 1, the


      agency found that systemic exposures of both


      actives were favorable between the extended-release


      suspension and multi dose of reference


      immediate-release solution, and there was no food


      effect on both actives.


                However, in terms of the IVIVC study,


      where the drug substance was coated with the 2


      percent formulation, 5.5 percent formulation and 9


      percent formulation, with respect to drug substance


      1, the agency found that it failed to establish the


      in vivo/in vitro correlation, and observed more


      than 20 percent of difference in Cmax for


      formulation of fast and slow dissolution profiles.


                With respect to drug substance 2 that was


      not coated, level A IVIVC was established, however


      it failed to validate the IVIVC.  The formulations


      used in the IVIVC study were found to be


      bioinequivalent, that is to say the Cmax of the


      formulations used in the IVIVC study were different


      by more than 20 percent.  The proposed dissolution



      specification and the approach to set a dissolution


      specification based on IVIVC by mapping was found




                Now let me share the stability results


      analysis.  This is what we review in our CMC


      assessment.  What we found was contradictory


      release profiles observed between drug product


      formulations containing 6 percent and 9 percent


      coated drug substance.  Drug substance 2 showed


      more decrease in dissolution than drug substance 1,


      and we observed substantial decrease in dissolution


      at 1-hour, 3-hour and 6-hour time points for both


      actives from the corresponding initial values among


      all batches, including bio. and primary stability


      batches, at all storage conditions.  The decrease


      in dissolution was most notable at 3-hour and


      6-hour time points.  The decrease in dissolution is


      minimum at the 12-hour time point and the decrease


      in dissolution for both actives levels off by 9


      months on storage.


                This is displayed on this slide.  This is


      the dissolution results of drug substance 1.  For



      clarity purpose, I have labeled the coating for the


      dissolution curves.  The yellow bar shows the 6


      percent coating that was used in the


      bioavailability study.  The purple is the 2


      percent.  The middle one is blue, which is 5.5


      percent coating of drug substance 1.  The red one


      is the 9 percent coating of drug substance 1.


                Now, what I explained in the previous


      slide is what you can see is a decrease in


      dissolution profiled for all the solutions.  You


      would expect the 9 percent would be showing a slow


      dissolution compared to the 6 percent but it is


      quite the other way.


                If you look at drug substance 2, the


      decrease is more compared to drug substance 1,


      which is shown basically from the least point and


      at the 18 months time point.  That is more than


      about 20 percent decrease in dissolution over time.


                So, based on this analysis these were the


      critical issues discussed with the applicant, and


      they concerned the raw material controls,


      manufacturing processing and in-process controls



      and controls related to particle size distribution


      and dissolution method.


                I just want to point out over here that


      these discrepancies in the results showed that the


      coating process was not in control and we discussed


      that issue with the applicant.  They decided to


      reformulate the drug product and decided to abandon


      the idea of the IVIVC approach to set dissolution


      acceptance criteria; conduct PK studies on


      commercial scale bio. batch containing drug


      substance 1 at the specified target coating level,


      rather than a range, and compare it to the


      reference IR solution; manufacture additional 3


      pilot scale primary stability batches of the drug


      product containing drug substance 1 at the same


      specified target coating level; and propose


      dissolution acceptance criteria based on in vitro


      dissolution profiles obtained for both actives from


      the bio. batch.


                These were the process improvements


      implemented.  They coated the drug substance with a


      specified target coating level of semipermeable



      polymer; revised the coating and subsequent


      manufacturing processes; instituted appropriate


      process controls to stabilize binding of both


      actives to the drug carrier support in the


      suspension; and manufactured one commercial scale


      bio. batch and three pilot scale stability batches.


                They instituted appropriate particle size


      measurement method, for example laser diffraction,


      for drug carrier support and coated drug-bound


      carrier particles.  They revised particle size


      distribution acceptance criteria for the drug


      carrier support, coated drug substance bound


      carrier support particles and suspension


      stabilizing excipients.


                Based on these results, they conducted


      three PK studies utilizing the drug product


      formulation with coating of drug substance 1.  They


      conducted BA/BE assessment; PK at steady state; and


      food effect studies.  The results showed that the


      PK profiles of drug substance 1 and drug substance


      2 from test extended-release suspension were found


      comparable to the reference IR solution following



      single and multiple dose administration, and food


      had no effect on bioavailability of both actives.


                Now let me share with you the stability


      results analysis.  After the implementation of the


      improvements in manufacturing process for coating


      and instituting adequate process controls in terms


      of particle size, we observed stable and consistent


      release profiles at 1-hour, 3-hour, 6-hour and


      12-hour time points for both drug substance 1 and


      drug substance 2 on storage within each of the bio.


      and three primary stability batches.  There was no


      discernible trend in release profiles of drug


      substance 1 and drug substance 2 and on bio. and


      primary stability batches at all storage


      conditions.  And, there were comparable release


      profiles for both drug substance 1 and drug


      substance 2 among bio. and three primary stability




                This is displayed in this graph for drug


      substance 1.  You can see, as opposed to the


      dissolution rates that we saw before and after


      implementation of manufacturing processes.  This is



      with respect to drug substance 1, which was coated.


      This is the bio. batch and these are the three


      primary stability batches.  Most of the


      dissolution, as you can see, ranges between 5-7




                This is with respect to drug substance 1


      dissolution profile.  This is the bio. batch and


      you can see these are the three primary stability


      batches and you do not see any discernible trend


      and most of the dissolution ranges between 5


      percent if you compare it to drug substance 2 prior


      to the implement.


                Then I would like to conclude my


      presentation with the following remarks.  We were


      able to identify probable causes of discrepant and


      inconsistent dissolution results for drug substance


      1 and drug substance 2, and recommend corrective


      measures to address the issues.  The outcome was


      consistent manufacturing process; acceptable BA/BE


      results; stable and consistent release profiles


      without any discernible trend on storage for both


      drug substance and drug substance 2.  Dissolution



      criteria which were set were better reflective of


      the data.  There was a substantial improvement in


      the quality of the drug product and there was a


      significant improvement in assurance of the safety


      and efficacy concerns.


                However, the case study example


      highlighted two significant points.  There was a


      lack of or poor understanding of the raw material


      properties and manufacturing processes that were


      critical to be controlled for consistent quality


      and thereby desired performance, for example,


      extended-release dissolution of the drug product.


      It also identified inadequate efforts invested by


      the applicant during the drug development to


      understand the causal links of dissolution




                The case study example stresses a dire


      need for improvement to the existing drug


      development efforts to understand the relationship


      between the raw material properties of formulation


      components and critical quality attributes of the


      drug product; the effect of raw material properties



      of formulation components on their processibility


      for selected manufacturing processes, and the


      effect of manufacturing processes and associated


      critical process parameters on the critical quality


      attributes of the drug product.


                I would like to end my talk with the last


      remark that there is no substitute to a systematic


      and scientific approach to drug development for a


      safe, efficacious and quality drug product.  Thank




                     Questions by Committee Members


                DR. COONEY:  Thank you.  There is an


      opportunity for questions.  Nozer?


                DR. SINGPURWALLA:  Just a point of


      information, you repeatedly used distribution,


      particle size distribution.  What particle size


      distributions do you use in your activities?


                DR. SHAH:  I am not following the




                DR. SINGPURWALLA:  Particle sizes are




                DR. SHAH:  Correct.


                DR. SINGPURWALLA:  They are not the same.


      So, they have a probability of distribution.


                DR. SHAH:  Yes.



                DR. SINGPURWALLA:  Now, there is a lot of


      literature, perhaps not in your business, on what


      should be the distribution of particle sizes.  This


      morning we heard the viable distribution attacked


      by my colleague here, but the log normal


      distribution is often used as a distribution of


      particle sizes.  My question is what distributions


      are used in the pharmaceutical industry for


      particle sizes, or is this a completely different




                DR. SHAH:  I am not sure how to answer


      that question, but I will tell you what we practice


      in CMC review.  We ask for the applicant to


      identify the particle size range in D10, D15 and


      D90.  That means 90 percent of the particles--


                DR. SINGPURWALLA:  Right.


                DR. SHAH:  And we ask for the span,


      basically the ratio of D10 to D90 divided by D15


      and that gives you where the distribution lies. 



      Basically, that kind of gives control of


      consistency of the particle size distribution.


                DR. SINGPURWALLA:  Actually, you answered


      my question.  What seems to be not there in your


      industry is you are just looking at the percentiles


      and if the distributions are skewed one way or the


      other it makes a big difference what they are when


      you simply work with the percentiles.  So, I am


      just encouraging you to look into that.


                DR. SHAH:  I agree.  Thank you.


                DR. COONEY:  Ken?


                DR. MORRIS:  I think one of the things


      that occurs is that people don't control the


      distributions.  They tend to be log normal sort of


      in a general sense but people don't intentionally


      control this.  They usually control to a mean,


      which is a real big problem--


                DR. SINGPURWALLA:  If you control the mean


      you start to control the distribution.


                DR. MORRIS:  Yes, you try to control the


      mean but there is no real--and I am not sure what


      historically the reason is for that but that is



      sort of the case.


                DR. SINGPURWALLA:  You need to know what


      it is.


                DR. MORRIS:  But you need to know what it




                DR. SINGPURWALLA:  You can't control it.


                DR. MORRIS:  That is right.  My question


      is do you think that there problem was control


      simply of film thickness or was it perhaps


      incorporation of one of the compounds into the film


      unintentionally during the coating process?


                DR. SHAH:  No, that was definitely the


      coating process, and this was like black art in


      that they were mixing and matching and they never


      had a handle on the coating process itself.


                DR. HUSSAIN:  One point that I think I


      wanted to illustrate with this presentation was


      that really to control, to achieve a state of


      control, and so forth, you have to get down to


      upstream activities, starting with raw materials,


      and so forth.


                The point I also wanted to sort of



      emphasize was that just focusing on a test, even


      when you have a correlation, which is just a


      correlation and may not be causal, I think is that


      gap that we are also trying to fill with focusing


      on the CMC part of the manufacturing controls.


      Without that the system really--the method is


      weakened.  So, the quality by design aspect is to


      emphasize that part of it.  So.


                DR. COONEY:  I think another dimension


      with this particular case is that there is a


      significant amount of complexity because you are


      dealing with multiple products, complexity both in


      the process as well as in the product itself.  This


      is I think a particularly good example where


      quality by design can have a greater impact with


      these more complex processes and products, and the


      processes and the products need to be thought


      through together, which is your point.  It is very




                I think we are actually going to begin


      lunch ten minutes early.  However, beginning lunch


      ten minutes early does not mean that you get an



      extra ten minutes for lunch.  We will reconvene at


      12:50--guess what, you can get an extra ten minutes


      for lunch.  We will reconvene at one o'clock.


                [Whereupon, at 11:50 a.m., the proceedings


      were recessed for lunch, to resume 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


                DR. COONEY:   If I could have people's


      attention, welcome back from lunch.  I hope that


      everyone appreciated the extra 9.5 minutes that you


      had for lunch.  It is one o'clock.  It is the


      opening period for open public hearing.  We have


      one presentation for this afternoon by Will Brown


      from USP, and he will speak with us on USP and


      dissolution testing.  Thank you.  Welcome


                          Open Public Hearing


                DR. BROWN:  Thank you so much, and I would


      like to thank the various FDA staff members for


      giving a staff member at USP the opportunity to


      speak before this committee.  I am a member of the


      staff of the Department of Standards Development at


      USP, and I serve as one of the liaisons to the


      Biopharmaceutics Expert Committee.


                This is breaking news.  USP reorganizes


      itself once every five years, and part of that


      reorganization is the election of the chair of the


      Council of Experts.  We have a reelected chair,


      Thomas Foster, for the Biopharmaceutics Expert



      Committee.  You can see on this slide the


      membership, and you will see names you recognize




                USP and dissolution--well, we are terming


      dissolution one of the performance tests.


      Performance tests currently mean dissolution or


      disintegration test, and by test I mean part of the


      specification.  The ICH definition, and it is very


      easy to use terms loosely, says that a


      specification is a list of tests, associated


      procedures and acceptance criteria.  So, that is


      kind of the idea of the USP dissolution.  It is


      part of the specification.  You will find the


      public specification in the USP monograph.


                The general dissolution test is found in


      the general chapter, 7-11 on dissolution, and that


      gives a general description of the techniques that


      are available, with the understanding that those


      techniques can be modified.  We saw this morning


      what the modifications might represent.  They might


      represent the appropriate medium or agitation or


      apparatus as determined by the applicant and the





                Now, the study design that is embedded in


      the dissolution test and the analysis is in three


      stages.  We have a fixed number of samples tested


      at each stage and there are acceptance criteria


      again that are determined by the applicant and the


      agency, and then communicated to USP by what I am


      terming the sponsor, who is the same party as the




                The general approach is to test by


      attribute.  In other words, a product is either


      good or bad.  It either conforms or it doesn't and


      that is a fairly decent approximation and


      convenient for application by an independent


      analyst but it doesn't necessarily address


      underlying distributions of performance.


                In the USP test by attributes there is a


      control on the spread of the data.  By example, at


      the S3 level where you tested 24 units there is a


      limit that says that no individual unit value can


      be below Q-25 percent.  So, there is an


      acknowledgement that there may be an underlying



      distribution at least on stability.


                For the Biopharmaceutics Expert Committee,


      in this cycle the expert committee is working on


      revising general chapters to include performance


      tests by dosage form, by route of administration.


      The current approach to applied dissolution is


      typically two oral products and some transdermals.


      The routes of administration that USP has


      identified were discussed in a stimuli article in


      Pharmacopeia Forum, in September, 2003 and


      basically identified five basic routes of


      administration, topical dermal, gastrointestinal,


      mucosal, by injection and by inhalation.  It is


      just a way to cut the universe.


                The intention is to work with the FDA and


      industry as appropriate but to facilitate this work


      the Biopharmaceutics Expert Committee has asked for


      the formation of advisory panels, which have been


      formed.  They were formed in the last cycle and


      they are currently meeting.


                My general feeling is that meetings may be


      productive but oftentimes they are not.  I have two



      examples of meetings that are productive.  In 1993,


      I am told that the predecessor to this committee


      met and out of that ultimately, in '97, came the


      immediate-release and extended-release guidances


      that were talked about this morning.  Another set


      of meetings that happened in that same time frame


      are the meetings of the Pharmacopeia discussion


      group.  The Pharmacopeia discussion group includes


      the Japanese pharmacopeia, the  European


      pharmacopeia, the USP and the World Health


      Organization.  In the process of harmonization,


      there actually has been a common statement with


      respect to system suitability.  It doesn't talk


      about calibrators, however there is a provision to


      have national text and in the national text portion


      of system suitability the USP continues to describe


      calibrators as part of the system suitability


      determination.  The general chapters are currently


      at stage six and that information can be found in


      the current PF and the corresponding Japanese and


      European documents.


                I was told that I only had ten minutes so



      this presentation is briefer than I usually intend,


      but I would like to draw the committee's attention


      to possibly a useful document.  This document


      article by Walter Hauck and a group at USP talks


      about oral dosage form performance tests, new


      dissolution approaches.  It is in the recent


      Pharmaceutical Research, I think February 22.2.  It


      talks about an approach that has explicit


      hypothesis testing.  Parametric tolerance interval


      is involved.  It gives an improved way, or at least


      the authors believe that it is an improved way to


      set dissolution acceptance criteria, and allows


      more flexibility in the design of a protocol.  So,


      I will just point you at that resource.  It may


      have some value.


                It allows the industry representatives


      more control on study design; allows the


      opportunity for tiered testing.  It doesn't


      specifically talk about tiered testing but allows


      that there may be an opportunity for some kind of


      successive testing on failing to meet the criteria


      at the first level.  It allows some flexibility in



      the number of units that are tested within each


      tier, and it allows the possibility that the test


      protocol, the test design could be changed from


      manufacturer to manufacturer.


                The idea is to set a probability of


      passing units from a batch where the clinical


      properties are known.  So, you characterize the


      batch for in vitro dissolution; determine, in some


      kind of a discussion with the agency--again, I am


      speaking from industry perspective even though I


      don't represent any industry perspective--sets the


      fraction of the units in this idealized reference


      population or this actual reference population that


      must conform to the standard.


                This approach, and I won't be able to


      describe this more fully, the authors believe will


      allow the consumer and producer risks to be clearly


      assessed, managed and communicated.  Ideally, if we


      continue with the model of dissolution for


      performance assessment, this could be communicated


      publicly in the compendium.  The basic underlying


      approach conforms to the approach for uniformity of



      metered dose inhalers that I believe this committee


      will be talking about tomorrow.


                Finally on to calibrators, the system


      suitability determination is written into the


      general chapter and, as I interpret it, is part of


      the performance of any dissolution test.  So, if a


      dissolution test is performed for compendial


      purposes, currently USP requires that the apparatus


      is demonstrated to be suitable, and the


      demonstration of suitability includes successful


      performance of the calibrators.


                In actual point of fact, the use of the


      calibrators has a GMP function.  Test apparatuses


      need to be demonstrated to be suitable twice a


      year.  So, that is the actual application of what I


      believe to be more comprehensive suitability


      determination.  I don't currently work in the lab


      but when I was in the lab if there were critical


      dissolution experiments to be performed, they were


      performed on an apparatus that was calibrated


      before and after so that the integrity of the data


      was not suspect on the grounds of an unsuitable



      apparatus.  The idea of calibration is not to focus


      on the performance of the apparatus but to rule out


      unsuccessful or unacceptable apparatus, so rule out


      apparatus on the extremes.


                The extremes--there is a range of


      acceptable results that is determine from a


      collaborative study, and we try to cast the net as


      widely as possible so that we can capture the


      sources of variability in properly operating labs.


      Inter-laboratory variability is a major component


      of the ranges.  I would submit that any one


      dissolution apparatus or assembly, because the USP


      looks at the apparatus as a single vessel, single


      spindle combination but, in fact, we have


      assemblies, groups of apparatus.  So, that is part


      of the wideness of the range.  We can talk about


      that if you wish.


                Calibrators, what we do with calibrators,


      USP is aware of problems.  Salicylic acid has


      elegance problems.  And, we go into unit packaging


      in the latest batch.  Prednisone tablets, the


      prednisone tablets that we distribute are a