FOOD AND DRUG ADMINISTRATION













                         JOINT SESSION WITH THE




                           ADVISORY COMMITTEE



                                VOLUME I













                       Wednesday, March 23, 2004


                               8:00 a.m.









                       Hilton Washington DC North

                           620 Perry Parkway

                         Gaithersburg, Maryland



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


      Alastair Wood, M.D., Chair


      Shalini Jain, PA-C, Executive Secretary


      Committee Members:


      Michael C. Alfano, DMD, Ph.D., Industry


      Terrence F. Blaschke, M.D.

      Ernest B. Clyburn, M.D.

      Frank F. Davidoff, M.D.

      Jack E. Fincham, Ph.D.

      Sonia Patten, Ph.D.., Consumer Representative

      Wayne R. Snodgrass, M.D., Ph.D.

      Robert E. Taylor, M.D., Ph.D., F.A.C.P., F.C.P

      Mary E. Tinetti, M.D.


      Special Government Employee (Voting):


      Michele L. Pearson, M.D.


      Government Employee Consultants (Voting):


      John S. Bradley, M.D.

      John M. Boyce, M.D.

      Ralph B. D'Agostino, Ph.D.

      Thomas R. Fleming, Ph.D.

      Elaine L. Larson, R.N., Ph.D.

      James E. Leggett, Jr., M.D.

      Jan E. Patterson, M.D.


      FDA Participants:


      Tia Frazier, R.N., M.S.

      Charles Ganley, M.D.

      Michelle Jackson, Ph.D.

      Susan Johnson, Pharm.D., Ph.D.

      John Powers, M.D.

      Curtis Rosebraugh, M.D.

      Debbie Lumpkins, Team Leader



                            C O N T E N T S


      Call to Order and Introductions

         Alastair Wood, M.D., Chair                              4


      Conflict of Interest Statement, Shalini Jain, PA-C

         Acting Executive Secretary                              8


      Issue Overview, Susan Johnson, Pharm.D., Ph.D.            10


      Regulatory History of Healthcare Antiseptic Drug

         Products, Tia Frazier, R.N., M.S.                      21


      Testing of Healthcare Antiseptic Drug Products,

         Michelle Jackson, Ph.D.                                31


      Microbiological Surrogate Endpoints in Clinical

         Trials of Infectious Diseases, John Powers, M.D.       54


      Antiseptic and Infection Control Practice,

         John Boyce, M.D., Yale School of Medicine             106


      Prevention of Surgical Site Infections,

         Michelle Pearson, M.D., CDC                           127


      Question and Answer Period                               163


      Open Public Hearing:

                Steven C. Felton, Ph.D.                        204

                J. Khalid Ijaz, DVM, Ph.D.                     211

          The Quset for Clinicaql Benefit

                Steven Osborne, M.D.                           214

      OTC-TFM Monograph Statistical Issues of Study

         Design and Analysis, Thamban Valappil, Ph.D.          224


      Industry Presentation:

         The Value of Surrogate Endpoint Testing for

            Topical Antimicrobial Products,

            George Fischler                                    250


      Statistical Issues in Study Design,

         James P. Bowman                                       276


      Committee Discussion                                     299




                         P R O C E E D I N G S


                    Call to Order and Introductions


                DR. WOOD:  Let's get started.  Welcome to


      the Over-the-Counter Advisory Committee.  Let's


      begin by going around the table and everybody


      introducing themselves, and we will start on this


      side, Charlie.


                DR. GANLEY:  Charley Ganley, Director of




                DR. POWERS:  John Powers, Lead Medical


      Officer for Antimicrobial Drug Development and


      Resistance Initiatives in the Office of Drug


      Evaluation IV.


                DR. ROSEBRAUGH:  Curt Rosebraugh, Deputy


      Director, OTC.


                DR. JOHNSON:  Sue Johnson, Associate


      Director, OTC.


                DR. LUMPKINS:  Debbie Lumpkins.  I am a


      Team Leader in OTC.


                DR. DAVIDOFF:  I am Frank Davidoff.  I am


      an internist and editor emeritus of Annals of


      Internal Medicine and a member of the OTC






                DR. FLEMING:  Thomas Fleming, Chair,


      Department of Biostatistics, University of




                DR. FINCHAM:  Jack Fincham, professor at


      the University of Georgia, College of Pharmacy, and


      I am a member of the committee.


                DR. CLYBURN:  I am Ben Clyburn.  I am an


      internist at Medical University of South Carolina


      and a member of the committee.


                DR. BRADLEY:  I am John Bradley, a


      pediatric infectious disease doctor from Children's


      Hospital, San Diego, and I am a member of the


      Anti-Infective Drugs Advisory Committee.


                DR. PATTERSON:  Jan Patterson, Infectious


      Diseases and Infection Control, University of Texas


      Health Science Center, San Antonio and South Texas


      Veterans Healthcare System.


                MS. JAIN:  Shalini Jain, Acting Executive


      Secretary for today's meeting.


                DR. PATTEN:  Sonia Patten.  I am the


      consumer representative on the panel, and I am an




      anthropologist on faculty at Macalester College in


      St. Paul, Minnesota.


                DR. SNODGRASS:  Wayne Snodgrass,


      pediatrician and clinical pharmacologist at the


      University of Texas Medical Branch.


                DR. LARSON:  Elaine Larson, from the


      School of Nursing and School of Public Health at


      Columbia University, in New York.


                DR. TAYLOR:  Robert Taylor, Chairman,


      Department of Pharmacology, Howard University, in


      Washington, internist and clinical pharmacologist.


                DR. BLASCHKE:  Terry Blaschke, internist,


      clinical pharmacologist, Stanford, member of the




                DR. TINETTI:  Mary Tinetti, internist,


      Yale University and member of the committee.


                DR. D'AGOSTINO:  Ralph, D'Agostino,


      biostatistician from Boston University, consultant


      to the committee.


                DR. LEGGETT:  Jim Leggett, infectious


      diseases at Portland Medical Center and Oregon


      Health Sciences University, and I am a member of




      the Anti-Infective Drugs Advisory Committee.


                DR. ALFANO:  I am Mike Alfano, New York


      University College of Dentistry, industry liaison


      to NDAC.


                DR. WOOD:  And I am Alastair Wood and I am


      the Chairman of the NDAC and Associate Dean at




                So, let's get started.  Shalini, do you


      want to read the conflict of interest statement?


      While she is digging that up, the weather has


      caught us and the first speaker from CDC is stuck


      in Atlanta--the story of people's life in the


      Southeast.  So, what she is going to do, she is on


      her way back to her office and she is going to


      e-mail us slides and then we will try and project


      the slides later in the morning, with her talking


      to us over the telephone.  So, that will be a


      nightmare I suspect.




                That means we will time shift everything


      up and then probably, depending on how she gets on,


      we may have the question and answer period  for the




      first ones a little bit earlier and take an earlier


      break and then come back to hear her, depending on


      how the technology is behaving.  Shalini, go ahead.


                     Conflict of Interest Statement


                MS. JAIN:  The Food and Drug


      Administration has prepared general matters waivers


      for the following special government employees who


      are attending today's meeting of the


      Nonprescription Drugs Advisory Committee on the


      microbiologic surrogate endpoints used to


      demonstrate the effectiveness of antiseptic


      products used in healthcare settings.  The


      committee will also discuss related public health


      issues, trial design and statistical issues.


                This meeting is held by the Center for


      Drugs Evaluation and Research.  The following


      meeting participants have waivers:  Dr. Jan


      Patterson, Dr. Sonia Patten, Dr. Thomas Fleming,


      Dr. John Boyce, Dr. Ralph D'Agostino and Dr. John




                Unlike issues before a committee in which


      a particular product is discussed, issues of




      broader applicability such as the topic of today's


      meeting will involve many industrial sponsors and


      academic institutions.  The committee members have


      been screened for their financial interests as they


      may apply to the general topic at hand.  Because


      general topics impact so many institutions, it is


      not practical to recite all potential conflicts of


      interest as they apply to each member.  FDA


      acknowledges that there may be potential conflicts


      of interest but, because of the general nature of


      the discussions before the committee, these


      potential conflicts are mitigated.


                With respect to FDA's invited industry


      representative, we would like to disclose that Dr.


      Michael Alfano is participating in this meeting as


      a non-voting industry representative, acting on


      behalf of regulated industry.  Dr. Alfano's role on


      this committee is to represent industry's interests


      in general and not any one particular company.  Dr.


      Alfano is Dean, College of Dentistry, New York




                In the event that discussions involve any




      other products or firms not already on the agenda


      for which FDA participants have a financial


      interest, the participants' involvement and their


      exclusion will be noted for the record.


                With respect to all other participants, we


      ask in the interest of fairness that they address


      any current or previous financial involvement with


      any firm whose product they may wish to comment


      upon.  Thank you.


                DR. WOOD:  Thanks a lot.  Let's go


      straight on to the first presentation from Susan


      Johnson.  Susan?


                             Issue Overview


                DR. JOHNSON:  Good morning.




                My name is Susan Johnson and I am the


      Associate Director of the Division of OTC Drug


      Products.  On behalf of the division, I would like


      to welcome the members of the Nonprescription


      Advisory Committee and the Anti-Infective Advisory


      Committee and our other guests.  As I am sure the


      committee members would agree, the bulk of the




      background package as a metric of the challenge


      that we face today is certainly significant, and we


      certainly appreciate everyone making as much


      headway as they could with that background package.


                We very much appreciate all of your


      assistance today.  There is a wide variety of


      issues to discuss and so you will see the


      representation of the committee being broadened


      from NDAC to include the Anti-Infective committee


      members, and we appreciate everyone's attendance,


      as well as our consultants.


                I will just be providing a brief


      introduction to the regulatory issues associated


      with the efficacy of OTC healthcare antiseptics.




                The OTC healthcare antiseptics include


      three categories of drug products, the healthcare


      personnel handwashes; surgical hand scrubs; and


      patient preoperative skin preparations that are


      used to scrub the skin prior to surgery.




                FDA's current approach to the evaluation




      of healthcare antiseptic efficacy assumes that


      healthcare antiseptics play a critical role in


      infection control, and Dr. Michelle Pearson and Dr.


      John Boyce will discuss this role in additional


      detail.  However, the efficacy of individual


      products must be demonstrated to meet regulatory


      requirements.  FDA's current regulatory standards


      are based on actual product performance and have


      been supported in previous public discussions such


      as this one.  Ms. Tia Frazier will explain more


      about the regulatory history of these products.


                FDA currently determines the efficacy of


      healthcare antiseptics using a surrogate endpoint,


      and that is used as the reduction in a log                              


                                                                     10 count


      of bacteria from the site of the test product


      application.  Dr. Michelle Jackson, from the


      Division of OTC, will discuss how the standard is


      used in the test methodology.




                This meeting has been convened because we


      have received citizen petition requests to change


      the threshold criteria for bacterial reduction.  We




      wish to present our review for your consideration


      of the efficacy data in the literature for these


      products.  We are asking that the advisory


      committee provide input about the standards that


      FDA needs to have in place to make regulatory






                What are some of the factors that can


      influence efficacy of the healthcare antiseptics?


      This is by no means an exhaustive list but is


      intended to give you an idea of why product testing


      is required to demonstrate efficacy.


                The first group of factors I am going to


      discuss are associated with the actual product.


      The active ingredient obviously affects efficacy.


      The spectrum of activity for each individual active


      ingredient is tested in associated testing criteria


      in vitro.  The potency or dose response of the


      active ingredient shall also be taken into


      consideration, although in some cases it is not


      well known.


                The formulation of the product can impact




      its efficacy and influence that to increase or


      decrease efficacy so the concentration and dose


      delivered to the site and vehicle and other


      inactives in the products can affect efficacy.  One


      thing that influences efficacy quite a bit is how


      the product is actually used, and that is led in


      large part by the way the product is labeled.




                Other factors that influence efficacy of


      healthcare antiseptics include actual use


      parameters, adherence to the labeling and other


      practice standards and actual implementation of


      both labeling and practice standards.


                There are many patient parameters that can


      affect the efficacy of these products, including


      things like health status which influences the risk


      for infection, as well as the type of procedure


      that is being conducted.


                Resident and transient bacteria, resident


      bacteria being normal flora and transient bacteria


      being those sorts that are introduced during


      healthcare processes, can affect efficacy as well. 




      The amount of bacteria that is delivered and that


      resides on the skin, either prior to or that is


      left residually after product use, is an important


      determinant of overall efficacy.  Virulence of the


      bacteria that exists on the skin affects efficacy


      as well.  A small amount of bacteria can be present


      and provide a great risk of infection.




                FDA in general assesses efficacy using


      randomized, controlled trials for the most part.


      These provide analytical strength and can be


      designed to control for multiple confounders.


      Critical to the design of controlled trials is the


      selection of active and vehicle control, and we


      will be discussing that later today.




                The endpoints that are normally used in


      randomized, controlled trials are clinical or


      surrogate endpoints.  Randomized, controlled trials


      typically use clinical endpoints because the


      relevance is more evident.  In some situations the


      difficulty and expense of conducting clinical




      trials is very important to industry.  An


      alternative to clinical endpoints is surrogate


      endpoints, and Dr. John Powers will later discuss


      the scientific and regulatory precedent for using


      surrogates.  Just as a reminder, and I am sure you


      have gleaned this from your reading already, but


      the current standards for OTC healthcare antiseptic


      efficacy are surrogate endpoints.




                The factors that should be considered when


      using a surrogate to assess healthcare antiseptic


      efficacy include validity.  We acknowledge from the


      outset of this discussion that there is limited


      information about the links between clinical


      outcomes and efficacy and use of the surrogates to


      determine efficacy.  Dr. Steve Osborne will discuss


      the literature surrounding this question a little


      bit later.


                The existing trials in the literature are


      not designed to validate our practice standards.


      Instead, our practice standards and use of


      surrogate are based on the use of antiseptics in




      practice and our experience with marketed drug




                Test methodology is also an important


      factor to consider when using surrogates.  Test


      methodology should evaluate the conditions of use,


      largely directed by the labeling or the intended


      labeling.  The test methodology to evaluate


      healthcare antiseptics with surrogates needs to


      characterize the tolerability of drug products.


      While we are talking primarily about efficacy


      today, the tolerability of these drug products is a


      major safety concern and does come up as part of


      the testing methodology.  Test methods do need to


      be standardized with regard to all inherent






                Other factors that should be considered


      when using surrogate endpoints are the decision


      thresholds and, as I have said, the current


      criteria are based on the NDA performance of


      existing approved products.  We suggest that any


      changes to these criteria on decision thresholds




      should be data driven.


                Analysis of test data is critical, and


      later today Dr. Thamban Valappil will be discussing


      the analysis of these data.  His talk is predicated


      on the previous discussions that we will be having


      about validity methods and thresholds, and he will


      talk about the need to evaluate the response of


      test products in the context of variability in both


      test methods and in patient response.




                Epidemiologic studies do provide


      information for healthcare antiseptics.  They


      provide actual use information on large populations


      and can often be used to suggest practice


      standards.  They are often used to generate


      hypotheses to be later studied in randomized,


      controlled trials.  But they are relatively


      insensitive to treatment differences and changes in


      things like threshold criteria.  So, using them to


      extrapolate for regulatory decision-making is of


      limited value.




                What specifically are we asking the


      advisory committee to address?  First, can we


      continue to rely on surrogate markers to assess




      healthcare antiseptic efficacy?  I would like to


      remind the committee, as we will several times


      today I am certain, that we have the need for


      ongoing assessment and decision-making of these


      products so we do need to have standards in place


      now and in the near future, as well as into the


      distant future.


                If surrogates can be applied, at least in


      the short term, is there compelling evidence to


      change our surrogate efficacy criteria now?  What


      is the best way to analyze the efficacy data?  And,


      what labeling information would be helpful for


      clinicians to understand product efficacy and


      potentially to compare among different products?


                With that, I will turn it over to Tia


      Frazier, who is a regulatory project manager in the


      Division of OTC Drug Products, and she will be


      discussing regulatory history.


                DR. WOOD:  Just before you take that slide




      off, there is sort of an underlying assumption


      there, which I think is right but I just wanted to


      articulate that there is a sort of regulatory


      inertia which is that in the absence of evidence we


      shouldn't change criteria.  Is that fair?  I am not


      disagreeing with that, I am just trying to put


      number two in that context.


                DR. JOHNSON:  Yes, I think that is very


      essential to this discussion.  What we have tried


      to make clear, and will make clear in other


      presentations, is that the surrogates are based on


      as much information as we have had prior to the


      mid-'70's, when this regulatory mechanism was


      invoked, until now.  There still is not a body of


      evidence, while we are asking you to assess that


      body of evidence and whether you think that compels


      us to change.  So, there are standards in place and


      we think that those standards are based on the


      information that has been available to this point.


      At this point we are reconsidering the standards


      and we do think, and we are suggesting to the


      committee that any change in the standards should




      be data driven.


                DR. WOOD:  Just to summarize, so what you


      are saying is that you don't want the committee


      particularly to consider the quality of the data


      supporting the standards; you want the committee to


      consider the quality of the data supporting a


      change in the standards.


                DR. JOHNSON:  Well, I think it is both but


      our concentration is really on the latter part of




                DR. WOOD:  All right, thanks.  The next


      speaker will be Tia Frazier.


                    Regulation History of Healthcare


                        Antiseptic Drug Products


                MS. FRAZIER:  Good morning.




                I am Tia Frazier, and I am a project


      manager in the OTC Division, and I will briefly


      review the regulatory history of the monograph for


      OTC healthcare antiseptic drug products.




                The monograph includes both consumer and




      professional use products.  Today we are addressing


      issues related to the professional use products


      included in the monograph, which we call the


      healthcare antiseptics.  I will start first by


      defining the healthcare antiseptics.  There are


      three recognized uses, that Susan has already told


      you about, included in the tentative final


      monograph.  These are patient preoperative skin


      preparations used to cleanse patient skin prior to


      surgery; surgical scrubs which are used by


      operating room personnel prior to performing


      surgery; and healthcare personnel handwashes which


      are the soaps and leave-on products that are used


      by all personnel in healthcare settings prior to


      contact with patients.




                We have two different mechanisms for


      regulating OTC healthcare antiseptics.  Companies


      can submit new drug applications, which we call


      NDAs, for specific drug products to the FDA.  Data


      provided in NDAs remains confidential.  The second


      mechanism that we have for regulating these




      products is the OTC drug monograph review process.


      Products submitted to the monograph review are


      judged on the safety and efficacy of their


      individual active ingredients.  The data review for


      monograph drug products is public.




                Just to add to this brief description, I


      will also tell you that the OTC drug monograph


      review began in 1972.  At that time, and for some


      years later, the agency made determinations about


      the safety and efficacy of over 200,000 OTC


      products that were on the market at that time.  We


      have reviewed 700 active ingredients in 26


      therapeutic categories with the help of expert






                The advisory review panel reviewed and


      made recommendations on ingredients and products to


      further the development of a drug monograph.  FDA


      then categorizes ingredients considered in the


      monograph review according to their safety and


      effectiveness for a particular use described in the




      review.  I won't say much more about how we


      categorize and evaluate ingredients since the focus


      of today's meeting is on the effectiveness criteria


      that we use to evaluate this particular group of


      professional use products.  The OTC review panel's


      recommendations are then published in an advance


      notice of proposed rule-making, or ANPR.




                After the ANPR is published we consider


      public comments as we develop a tentative final


      monograph, or TFM.  A TFM is FDA's proposed






                FDA usually receives more data and public


      comments on any TFM that we publish.  Typically, we


      publish a final monograph after a tentative final


      monograph.  In this case, we published a second


      tentative final monograph in 1994 after the first,


      which was published in 1978.




                We, at FDA, have the current view that


      antiseptics do play a pivotal role in the practice




      of infection control today.  We operate from the


      presumption that antiseptics can decrease the


      number of organisms on the surface of the skin and


      this probably reduces the spread and development of


      nosocomial infections.


                Based on this presumption, we adopted


      surrogate endpoints, measurements of log reductions


      on the skin surface that are intended to indirectly


      measure the effectiveness of antiseptics that we


      regulate.  This is the reason that FDA and the


      European regulatory bodies selected this particular


      surrogate endpoint, the reduction of the organisms


      on the skin surface, to evaluate the effectiveness


      of these products.




                The advisory review panel recommended in


      1974 that we use surrogate endpoints to measure


      antiseptic effectiveness.  To date, unfortunately,


      we still have not figured out how to design a


      clinical study that can measure the contribution of


      an antiseptic in reducing the likelihood of


      contracting or spreading nosocomial infection. 




      With any luck, today Dr. Pearson will explain later


      why designing studies like this is so difficult.




                So, now I am going to go into the history


      of the monograph as it relates to the surrogate


      endpoints.  The first defined surrogate endpoint


      for patient preoperative skin preparations appears


      in our 1974 ANPR.  It was also incorporated in the


      first tentative final monograph which, I said, was


      published in 1978.  Then the panel recommended a


      3-log reduction in organisms on the surface of the


      skin as the requirement for patient preoperative


      skin preparation.  At that time, NDA products were


      often approved for patient preoperative skin


      preparation indications based on their ability to


      meet a 3-log reduction and the monograph simply


      adopted this commonly used NDA standard.


                It is important to realize that the


      effectiveness criteria used today to evaluate


      products marketed under the monograph are really


      based on the effectiveness criteria often applied


      to NDA products.  NDAs, of course, can be approved




      with alternate clinical endpoints and are not


      necessarily bound by the monograph standards.




                Moving on to the surgical hand scrub


      criteria, the history on this is that Hibiclens is


      an NDA product that was approved in 1975 based on a


      new surrogate model developed to evaluate surgical


      scrubs.  FDA incorporated the effectiveness


      criteria applied to Hibiclens surgical scrub into


      the developing antiseptic monograph.  These


      criteria were published in our second tentative


      final monograph, on June 17, 1994.


                Hibiclens is often included as a positive


      or active control in testing designs for antiseptic


      products.  Because these are laboratory tests,


      companies are required to include a positive


      control arm using an approved product like


      Hibiclens to ensure that the tests are conducted






                The current 3-log reduction criteria


      proposed for healthcare personnel handwashes in the




      second tentative final monograph was based on FDA's


      evolving understanding of what the NDA products


      under review at that time could achieve.




                As I have said before, this monograph is


      unusual because there are two tentative final


      monographs associated with it.  In 1994 we elected


      to publish a second tentative final monograph


      rather than a final monograph to allow for public


      comment on the new testing requirements.  The


      current proposed testing requires in vitro studies


      of the product spectrum and kinetics of


      antimicrobial activity and of the potential for the


      development of resistance.  We also require in vivo


      studies of effectiveness under conditions that we


      think simulate how the product is actually used in


      that healthcare setting.


                Another unusual aspect of this monograph


      is that it requires in vitro and in vivo testing


      not only for the approval of new products but also


      for the approval of new formulations.  We require


      this testing to be done because changes in the




      inactive ingredients or dosage forms can affect the


      product's effectiveness.




                Products are required to meet key


      attributes important to their performance in


      healthcare settings.  We state that a healthcare


      personnel handwash should be persistent if


      possible.  We would like it to be non-irritating,


      fast acting and be able to kill a broad spectrum of


      organisms as well.


                Persistence, or the ability to have a


      residual effect for some time after the product is


      used, is also an attribute that we would want a


      surgical scrub or a patient preoperative skin


      preparation to have as well.




                We have had two prior public discussions


      about these effectiveness criteria.  We discussed


      performance testing at an advisory committee


      meeting in 1998.  This was a general discussion


      only and we did not present questions for the


      committee to vote on.  Then in 1999 we held a




      public feedback meeting to hear the industry


      coalition present an alternative model or framework


      for evaluating antiseptics.  Dr. Jackson will cover


      the effectiveness criteria proposed by this


      industry coalition in her presentation that follows






                I think everyone here today would agree


      that it is critical that FDA ensures it uses the


      right criteria to evaluate antiseptic products.


      There are many dangers we can imagine might occur


      if we allow ineffective products to be sold and


      used in hospitals.  We need these products to work.


      The OTC and anti-infective divisions admit that the


      effectiveness criteria we currently use are not


      based on data from clinical studies.  We recognize


      this as a limitation of our current standards.


                The divisions recently reviewed available


      scientific data on topical antiseptic products used


      in healthcare settings.  We searched for data that


      could be used to support effectiveness standards


      for this class of products.  Our review of more




      than 1,000 studies submitted by industry and picked


      up through our own literature search is included in


      the committee background packages.  Dr. Steven


      Osborne will present the results of his review and


      evaluation of a section of those references that


      address clinical benefit later on this morning.




                The monograph for OTC healthcare


      antiseptic drug products is in the tentative final


      monograph or proposed rule stage.  We are in the


      process of writing a final rule, and we need your


      recommendations on what the effectiveness criteria


      should be in order to finalize this monograph.


                Now I would like to introduce my


      colleague, Dr. Michelle Jackson, who is a


      microbiology reviewer in the Division of


      Over-the-Counter Drug Products.  She will review


      the testing methodologies used to evaluate these




             Testing of Healthcare Antiseptic Drug Products




                DR. JACKSON:  My talk will focus on the




      testing criteria for healthcare antimicrobial drug


      products, and currently the development and


      standardization of protocols regarding the testing


      criteria for healthcare antiseptic drug products


      are based on earlier NDA review process.




                My presentation will discuss where we are


      with the proposed monograph requirements in regards


      to clinical simulation testing procedures for


      healthcare personnel handwash, surgical hand scrub


      and patient preoperative skin preparation, and the


      use of surrogate endpoints, also referred to as log


      reductions, with the three healthcare professional


      products.  Then I will go over the industry


      coalition's position of wanting to use alternative


      criteria.                [Slide]


                During the early stages of the antiseptic


      NDA review process standardized protocols did not


      exist.  However, the agency requires standardized


      and reproducible methods, therefore, as the NDA


      review process evolved clinical protocols used


      throughout the NDA review process also evolved into




      protocols now recommended in the tentative final




                So, what makes a good clinical simulation


      test method?  It should simulate as close as


      possible the actual use conditions.  Ideally,


      clinical simulations should include design


      characteristics such as test product, also referred


      to as final formulation; the test product contains


      the active antimicrobial agent; a vehicle control


      arm is the test product without the active


      antimicrobial agent and vehicle, and negative


      control that shows how much contribution of


      reduction is due to just the mechanical action of


      washing the hands.


                A current trial design in TFM does not


      recommend inclusion of a vehicle for healthcare


      personnel handwash and patient preoperative


      testing.  The active control arm is also referred


      to as the positive or internal control.  The active


      control is used to assess the reproducibility of


      the clinical simulation studies and also used to


      validate the study.  This standard is usually a




      chlorhexidine gluconate containing product.


      Clinical simulations should also measure the


      desired product performance.  This simulation


      testing generates the surrogate endpoints and it


      should also be reproducible.


                I will briefly go over the three testing


      criteria for healthcare personnel handwash,


      surgical hand scrub and patient preoperative skin






                For healthcare personnel handwash, the


      label indicated use is handwash to help reduce


      bacteria that potentially can cause disease.  The


      products are used by healthcare professionals on a


      daily basis up for to 50 handwashes per day.  The


      testing process predicts the reduction of organisms


      that may be achieved by washing the hands after


      handling contaminated objects or caring for


      patients.  Here we are focused on the removal of


      transient organisms.  The testing process is


      designed for frequent use and it measures the


      reduction of transient organisms after a single use




      or multiple uses to initial baseline level.


                The studies are designed to demonstrate a


      cumulative effect of an antiseptic, meaning that


      the product gets better and better in reducing the


      bacterial load on the hands.  Thus, the products


      are considered broad spectrum, fast acting and, if


      possible, persistent.  The TFM surrogate endpoints


      propose a 2-log reduction for the first wash and a


      3-log reduction for the 10th wash.




                For the inclusion criteria subjects


      participating in the studies must be between the


      ages of 18-69, generally in good health, and have


      no clinical evidence of dermatosis, open wounds,


      hangnails or other skin disorders.


                The subjects are excluded if they have


      been diagnosed with having medical conditions such


      as diabetes, hepatitis, or having an immune


      compromised system, subjects having any sensitivity


      to antimicrobial products, pregnant or nursing


      women also would be excluded from participating in


      a study.


                For the healthcare personnel handwash


      there is a one-week washout period where subjects


      are instructed to use a non-antimicrobial product,




      such as soaps, deodorant and shampoos, and avoid


      bathing in chlorinated pools and hot tubs.




                The outline of the test procedure includes


      a test practice wash using bland soap.  This


      basically removes any oils and dirt from the hands,


      and the bacteria counts are compared to the


      baseline counts.  The hands are contaminated with


      Serratia marcescens and immediately sampled, and


      the baseline is determining the number of organisms


      on the surface of the skin prior to using an


      aseptic product.


                The handwashing schedule involves ten


      washes performed on one day.  At the first wash the


      hands are contaminated and washed with the test


      product.  The hands are then sampled for microbial


      counts.  Eight additional washes are performed, and


      at the tenth wash the hands are sampled for


      microbial counts and the product must achieve a




      specific log reduction after the first and tenth


      washes.  The repetitive hand washing aspect of the


      study design is intended to mimic the repeated use


      of a product in hospitals.  The repetitive washing


      is also used to measure the cumulative effect, and


      cumulative effect is a progressive decrease in the


      number of microorganisms recovered following the


      repeated application of the test product.




                Once the hand washing procedure is


      completed, the subject's hands are decontaminated


      by sanitizing the hands with 70 percent alcohol.


      The purpose of this is to destroy any residual


      Serratia marcescens left on the skin.  Typical


      handwashing procedures involve contaminating the


      hands with a microorganism, Serratia marcescens.


      The hands are rubbed together for 45 seconds, and


      the hands are held away from the body and allowed


      to dry for a few minutes.




                Once the hands are dry, a specific amount


      of test product is dispensed into the cupped hands




      and the next step is to lather and wash all over


      the surface of the hands and above the wrists.


      After the completion of the wash, the hands and


      forearms are rinsed under regulated tap water with


      a temperature of 40 degrees Celsius for 30 seconds.




                The hands are then placed in plastic bags


      and sampling fluid is added to the bag containing


      neutralizers.  Neutralizers are reagents that stop


      the antimicrobial reaction.  Sampling should occur


      within five minutes after each wash.  The bags are


      tightly secured above the wrist with a strap.  The


      hands are massaged for one minute, paying


      particular attention to the fingers and underneath


      the nails.




                An aliquot of the sampling fluid is


      aseptically withdrawn from the bag and transferred


      immediately to dilution tubes.  The microbial count


      determination is performed by surface plating and


      this is done within 30 minutes of sampling.  The


      plates are incubated for two days at 30 degrees








                This diagram depicts the colony forming


      units, CFUs, from two dilution plates.  CFUs are


      then converted into log counts.  Serratia


      marcescens produces a red pigment color for easy


      identification, and it distinguishes itself from


      the normal flora of the hands that appear white or


      yellowish on agar plates.  Here, I want to


      emphasize that we are just counting bacteria.




                Here the industry coalition suggest a 1.5


      log reduction for the first wash, and suggest


      eliminating the tenth wash.  We require the test


      product to show a cumulative effect, that is an


      evaluable attribute, that shows a progressive


      decrease in the number of organisms recovered


      following repeated application of a test product.




                For surgical hand scrub the indication use


      is to significantly reduce the number of organisms


      on the skin prior to surgery.  These products are




      used to reduce the resident and eliminate the


      transient flora of the hands of surgeons and


      surgical personnel, thus reducing the incidence of


      post-surgical site infection.


                The testing process is designed to measure


      the immediate and persistent reduction of resident


      organisms after a single or repetitive treatment.


      Here there is no artificial contamination of the


      hands, and the testing of the surgical hand scrub


      involves multiple test product use and repeated


      measurements of the bacterial reduction.  These


      antiseptics are considered broad spectrum, fast


      acting and persistent.  The TFM surrogate endpoints


      propose a 1-log  on day 1 for the first wash; 2-log


      on day 2 at the second wash; and 3-log on day 5 at


      the 11th wash.




                The subjects are selected through the


      inclusion/exclusion criteria for surgical hand


      scrub testing.  A 14-day or 2-week washout period


      is required.  Soon after the washout period the


      baseline counts are determined, and they are




      sampled two times, first on day one and the second


      estimate includes one of the three options.  On day


      3 and 5, 5 and 7, or 3 and 7.


                Subjects with a baseline greater than or


      equal to 5 logs after the first and second baseline


      estimates will qualify for the study testing


      period.  So, no sooner than 12 hours and no longer


      than 4 days after completion of the baseline


      determination subjects perform the initial scrub


      with the test product.  The surgical hand scrub


      testing requires a total of 11 scrub washes over a


      5-day period.  The sampling occurs on day 1, day 2


      and day 5.


                The reason we test 5 days is that the


      procedure mimics typical usage and permits the


      determination of both immediate and long-term


      bacterial reduction.  Each day the antimicrobial


      soap is used it produces a greater effect due to


      the persistence of minute residues left from the


      previous scrub.  This effect is called cumulative


      effect, and that is the reason why we test for 5






                An amount of the test product is dispensed


      according to the manufacturer's labeling




      instructions.  The soap is distributed all over the


      hands and two-thirds of the forearms.




                The hands are then scrubbed according to


      the manufacturer's directions, and if no directions


      are provided the TFM requires two five-minute scrub


      procedures.  A scrub brush is used to scrub the


      hands including the nails, the fingers, and


      interdigital spaces of the hands.




                A lab technician will don sampling gloves


      on the subjects.  One-third of the hands in a


      treatment group is sampled immediately.  The gloves


      remain on the test subjects' hands for either three


      hours or six hours prior to sampling.  Enumeration


      of bacterial flora three hours after the scrub is


      conducted in order to demonstrate continued


      effectiveness of the product during the time


      required for a surgical setting.  The enumeration




      of bacterial flora six hours after the scrub is


      conducted to demonstrate the suppression of


      bacterial counts over a period of time chosen as


      representing the maximum duration of most surgical


      procedures, that is, on average most surgeries will


      not last greater than six hours and, if so,


      surgeons usually rescrub.




                A specified amount of sampling fluid then


      is added to the glove pan, and the gloves are


      fastened securely above the wrist and strapped, and


      the hands are then massaged for one minute, paying


      particular attention underneath the nails.




                An aliquot of the sampling fluid is


      aseptically withdrawn from the glove and


      transferred immediately to dilution tubes


      containing neutralizers.  A microbial count


      determination is performed by surface plating, and


      this is done within 30 minutes of sampling. The


      plates are incubated for two days at 30 degrees






                Here the industry coalition agrees with


      the 1-log reduction for the first wash.  They




      suggest eliminating the second and 11th wash.  They


      suggest that persistence of antimicrobial activity


      should not be a requirement for surgical hand


      scrub.  We require an assessment of persistent


      activity in case there is a tear in the surgeon's


      glove, and it is assumed that the persistent effect


      will prevent the multiplication of resident flora


      on the gloved hand, thus preventing contamination


      of the surgical field.




                For the patient preoperative skin


      preparation or surgical prep labeled for the


      indicated use helps reduce bacteria that


      potentially can cause skin infection.  These


      antiseptic products must be fast acting, broad


      spectrum and persistent and, statistically reduce


      the number of organisms on intact skin.  They are


      designed for use by healthcare professionals to


      prep the patient's skin prior to invasive surgery




      or prior to injection.  These indications, however,


      do not cover more specific indications such as


      catheter insertions and open wounds.


                The testing process measures the immediate


      and persistent reduction of resident bacteria after


      a single treatment.  The TFM surrogate endpoint


      proposed a 1-log reduction for pre-injection; 2-log


      for the abdomen or dry site; and 3-log for the


      groin or moist site area.




                The subjects are selected through the


      inclusion/exclusion criteria for patient preop


      testing.  A 14-day washout period is required, and


      no bathing 24 hours prior to the baseline


      screening.  We want to try to obtain a high


      bacterial count for the baseline.  The TFM


      recommends the baseline screening counts for


      pre-injection to be greater than or equal to 3


      logs.  The TFM recommends that baseline screening


      counts for the common surgical sites for both dry


      and moist site areas, and the sites are to present


      bacterial populations large enough to allow the




      demonstration of bacterial reduction for up to 2


      logs centimeters squared for the abdomen sites and


      up to 3 logs centimeters squared on the groin






                For the abdominal site testing a 5 X 5


      treatment site area is marked on the skin using a


      permanent marker.  The template is divided into


      four quadrants for baseline, 10 minutes, 30 minutes


      and 6 hours sampling.




                The baseline sampling is performed using


      the cylinder sampling technique.  A sterile


      scrubbing cup is held firmly against the skin over


      the site to be sampled.  The scrub solution


      containing neutralizers is placed into the cup and


      scrubbed with moderate pressure for one minute


      using a sterile rubber-tipped spatula.  This


      procedure is also used for sampling for the


      treatment site.




                The application of the prep formulation is




      applied to the testing area.  For 30-minute and


      6-hour sampling sites a sterile gauze is placed


      over the prep area to help prevent microbial


      contamination.  The gauze pad is held in place by


      the sterile teeth dressing.




                The treatment samples are taken from the


      site areas using the cylinder sampling technique.


      A similar procedure is also used for testing the


      groin site area.




                Here the industry coalition agrees with


      the 1-log reduction at the pre-injection site, and


      they suggested that only a 1-log reduction should


      be required for the abdomen site and a 6-hour


      persistent is not needed.  For the groin site a


      2-log reduction should be required and a 6-hour


      persistent is not needed.




                FDA has received objections to the TFM


      proposed effectiveness criteria through comments in


      a citizen's petition.  Industry contended that the




      current performance criteria for healthcare


      antiseptics are overly stringent.  They claim that


      two category ingredients, alcohol and iodine, and


      one NDA approved ingredient, CHD, cannot pass the


      current testing requirements.  They claim that all


      antiseptic products only need to be effective after


      a single use, and they also do not want to meet the


      persistence requirement.




                This table summarizes the bacterial log


      reduction in industry's proposal for the healthcare


      antiseptic compared to FDA current standards for


      final formulation for healthcare personal handwash,


      surgical hand scrub and patient preoperative skin


      preparation I just reviewed.  Over the years the


      industry coalition has made several proposals for


      the revised effectiveness criteria.


                For the healthcare personal handwash, it


      should be effective following a single use.  A


      cumulative effect should not be a requirement.  For


      surgical hand scrub, it should be effective


      following a single use and also a cumulative effect




      should not be a requirement.  And for patient


      preop, the pre-injection and abdomen dry site a


      1-log reduction is suggested, and for a worst-case


      scenario such as the groin site area, it should


      need a 2-log reduction.




                We are aware the surrogate endpoints lack


      the clinical validation of a test method and


      performance criteria.  They do not measure the


      level of residual bacteria on the skin and


      virulence of the residual bacterial is not factored


      into the log reduction determination.  We realize


      that we are just measuring the mean log reduction.


                The criteria is based largely on earlier


      NDA performance and we have approved over 20 NDAs


      based on using surrogate endpoints.  These criteria


      are consistently applied to monograph products and


      many NDAs.  Industry has deviated from following


      the TFM in regards to variability in testing


      procedures such as scrub techniques and lab


      analysis, and it is not compared to vehicle or


      active control.  We will later hear from Dr.




      Valappil regarding improving statistical analysis


      that could be applied to the existing criteria.




                Overall, it is impossible to compare the


      data across studies due to the vast differences and


      methodologies that were used, and other limitations


      such as the following:  The majority of the studies


      were designed as product comparisons; studies were


      not designed to assess the product's ability to


      meet the TFM effectiveness criteria.  There were


      significant variations in how the studies were


      conducted; different testing procedures were used;


      and neutralizer validation data were not generally


      provided.  More than half the data submitted did


      not include neutralizers in the testing procedures,


      which can result in artificially high log


      reductions.  Generally, sample sizes were small in


      the studies and there was a limited number of


      subjects included in the testing procedure.  And,


      alcohol alone did not meet the 10th wash 3-log


      reduction.  However, most were able to meet the


      3-log reduction of the first wash.  We are




      currently evaluating the alcohol leave-ons and


      alcohol gel products.




                This slide was included to show that other


      countries also use surrogate endpoints.  The


      European performance criteria for handwash require


      that the test product mean log reduction factor


      should be greater than soap that has an average


      reduction log of 2.8.  The performance criteria for


      hand rub require that the test product mean log


      reduction factor should be equal to or greater than


      60 percent isopropyl alcohol that has an average


      reduction log of 4.6.




                In summary, we measure bacterial log


      reduction and testing methodology for healthcare


      personnel handwash, surgical hand scrub and patient


      preop.  These log reductions are used as surrogate


      endpoints to evaluate effectiveness.  How should we


      analyze this data?


                Later this morning we will hear from Dr.


      Valappil a presentation on statistical analysis for




      healthcare and aseptic drug products.  You will


      also hear from Dr. Steve Osborne who will discuss


      the relationship of these outcomes and


      corresponding reduction in the incidence of


      nosocomial infections in healthcare settings where


      the product use remains undefined.




                We are aware of the limitations of these


      test methods, and we assume that the incidence of


      infections as related to current use of existing


      products and lowering these standards may increase


      the infection rates.  We need research to validate


      these surrogates, and we need to have products on


      the market now and in the use of actionable


      criteria in the meantime.  That concludes my




                DR. WOOD:  Mike, you approached me earlier


      about some confusion about the data.  Do you want


      to comment on that at this stage?


                DR. ALFANO:  Yes, I have been advised that


      industry is not recommending removal of the 6-hour


      persistence requirement but, rather, the cumulative




      effect requirements.  Apparently, that came about


      because of some confusion over a table that the


      industry submitted.


                DR. WOOD:  Can you put slide 12 back up?


      Is that the one that we are talking about here, on


      page 6?  Is that where the confusion is?


                DR. ALFANO:  Actually, it was brought to


      my attention versus the questions that we are to


      answer today, which is on the last page of the




                DR. WOOD:  I was just trying to clarify


      these slides.  So, there is no confusion about what


      industry's position is on the slides?  Is that




                DR. ALFANO:  That is correct.


                DR. WOOD:  Well, I think there is


      actually.  Somebody seems to want to comment.


                DR. FISCHLER:  George Fischler, manager of


      microbiology for the Dowell Corporation,


      representing the STA-CTFA coalition.  Yes, there is


      some confusion.  On this slide, yes, where it says


      surgical hand scrub, there is an asterisk and




      patient preoperative skin preparation, an asterisk.


      Industry has not recommended the removal of the


      6-hour persistence criteria.  The only criteria


      that we recommended approval for is the cumulative




                DR. WOOD:  Okay.  Well, let's come back to


      discussing that later.  I am even more confused now


      but let's go on to the next speaker.


                DR. JACKSON:  The next speaker is John


      Powers.  He is the lead medical officer in the


      Antimicrobial Drug Development and Resistance


      Division, and he will discuss the biological


      surrogate endpoints in the clinical trials of


      infectious disease.


            Microbiological Surrogate Endpoints in Clinical


                     Trials of Infectious Diseases


                DR. POWERS:  Thanks, Michelle.




                Today I am going to discuss issues related


      to microbiological surrogate endpoints in clinical


      trials of infectious diseases.  Some of the members


      of the Anti-Infective Drugs Advisory Committee




      won't be surprised by any of this since this is an


      issue that has come up in infectious disease trials


      over and over again.  So, I am going to try to


      discuss just some of the general points that have


      to do with selecting surrogate endpoints in these


      types of trials.




                The first thing I am going to talk about


      is differentiating what we do in clinical practice


      and how one develops clinical practice guidelines


      with what one actually does in a clinical trial,


      and how those are very different situations.  Then


      what I would like to do is define our terms and


      talk about what is an endpoint; define what a


      clinical endpoint and surrogate endpoints are and


      differentiate those from biomarkers.  One of the


      things you will hear often, and probably we will


      make the mistake today, is using the term surrogate


      markers rather than surrogate endpoints, which is


      rather non-specific and causes some confusion.


                Then we will talk about the utility of


      surrogates in clinical trials and differentiating




      surrogate endpoints from surrogates as risk


      factors, which is an entirely different


      consideration.  I will talk about some of the


      strengths and limitations of surrogate endpoints


      and then, finally, relate all of that information


      to the use of surrogates in the setting of topical






                What we do in clinical practice is we are


      using drug products that are already proven to be


      safe and effective and, hopefully, we are not


      experimenting on our patients; we are using the


      products in a way where they are already shown to




                In clinical practice we impose several


      interventions on patients and hope they get better.


      We are not really concerned with why they get


      better when we do all that stuff to them, only the


      fact that they get out of the bed and they leave


      the hospital cured.  We develop treatment


      guidelines to help us describe the use of the


      products based on whatever the best available




      evidence is, and a lot of current treatment


      guidelines actually put grades on the evidence


      where you will see A-1 all the way down to D that


      talk about whether it is from randomized,


      controlled trials versus observational evidence as


      well, but optimally these treatment guidelines are


      based on randomized, controlled trials.  When that


      data is not available we oftentimes have to put


      things into these guidelines based on the best


      available evidence that we have.


                The unfortunate thing is that sometimes


      these guidelines then become the reason for not


      getting the data from randomized, controlled trials


      because people will come to us and say the


      guidelines say this, therefore, you can't do a


      trial to evaluate it.  And, that is probably not


      what the people who alter these guidelines actually


      are intending.


                This differs from clinical trials which


      are experiments in human beings to determine if


      drug products are safe and effective.  Clinical


      trials differ from clinical practice in that we are




      using the scientific method.  We are trying to hold


      as much as possible constant, except for the


      interventions, so that we can apply the outcomes to


      causality related to the interventions themselves,


      which is very, very different from clinical


      practice.  So, how we do this is often to use


      concurrent controls which is something that we do


      not do in clinical practice.  In clinical practice


      we look at what the patient is at baseline and


      compare what happens at the end.  That is not what


      we do in clinical trials where we are comparing


      what happens at the end in patients who receive the


      test product versus a control.


                These clinical trials are, hopefully, to


      provide the evidence for formulation of practice


      guidelines and, as I said, hopefully, it is not


      vice versa where the guidelines determine that we


      can or cannot do a clinical trial.  But the big


      issue in clinical trials is that we need to


      determine some yardstick to determine if products


      are safe and effective.  How are we going to


      measure those products to make that kind of




      assessment?  That is really what we are asking




                And, the reason for this slide is to sort


      of outline the real question today.  We are not


      questioning whether handwashing is important or


      whether handwashing should be done in clinical


      practice.  What we are asking today is how do we


      develop a yardstick to determine which products are


      safe and effective to use in handwashing.




                So, let's define some of the terms that we


      are going to use today.  An endpoint is a measure


      of the effect of an intervention on an outcome,


      outcome being defined, for instance, as success or


      failure in a clinical trial in the treatment or


      prevention of a disease.  Again, it is important to


      realize that what we are talking about here is a


      disease.  We are not preventing someone getting an


      organism on their skin.  What we are really trying


      to look at is does that prevention of getting an


      organism on the skin, in turn, result in prevention


      of disease.


                But whenever we are picking an endpoint we


      have several questions that we have to address.


      The first one is what are we going to measure?




      Obviously, this should be clinically relevant to


      the disease in question.  We are not going to ask


      if your left earlobe hurts when we are trying to


      evaluate something that has to do with foot pain.


                The next question is how to measure it?


      And, we should be able to measure differences


      between therapies, should they exist, and that gets


      to this issue of the yardstick and that we need to


      be able to differentiate effective from ineffective




                The next issue is when do we actually


      measure it?  If we apply a product and come back in


      two years and then try to determine if there are


      differences between the patients we are probably


      not going to see a whole lot in a non-lethal




                The next question is how much to measure,


      what magnitude of difference actually makes a


      difference to patients?  A lot of this has to do




      with sample size.  We could take a product that is


      99 percent effective and show that it is


      statistically different than a product that is 90


      percent effective if we studied thousands and


      thousands of patients.  So, it gets to the issue of


      clinical significance versus statistical




                Then, one of the big issues I am going to


      ask you to talk about today is when we get some


      results, how do we analyze those results so that we


      can logically draw conclusions from them?




                This is a cartoon from the New Yorker,


      which sort of outlines the issue in choosing


      endpoints that are relevant to patients.  Here


      there is a doctor who has just done an endoscopy on


      a miserable patient, and the doctor says


      congratulations, the endoscopy was negative;


      everything is perfectly all right.  So, according


      to the surrogate endpoint of what the doctor saw on


      the endoscopy, the patient feels great but the


      patient is saying my symptoms bother me.  I am




      worried and concerned.  I can't exercise; I can't


      eat.  My whole life is affected.  So, that gets to


      the difference between measuring a surrogate and


      measuring what the patient actually feels.




                This seems sort of redundant but it is


      probably important to define what a disease


      actually is.  In these terms we are talking about a


      constellation of signs and symptoms experienced by


      the patient.  Although infectious diseases are


      caused by pathogenic organisms, those result in a


      host response and it is actually the host response


      that causes a lot of the symptoms that we see.


                When we are talking about surrogates we


      often hear about Koch's postulates.  Well, these


      fulfill Koch's postulates so the surrogate must


      work in the setting of an endpoint of a clinical


      trial.  But Koch's postulates relate to proving the


      cause of a disease, that a pathogen actually causes


      that particular illness, and Koch's postulates were


      never designed to measure the effect of an


      intervention.  It is very important in our




      discussion today to separate out cause from effect


      which are two different considerations.


                One of the issues we always talk about is


      that patients seek the care of clinicians because


      they have symptoms when they have a disease, not


      because of the presence of an organism.  So, a


      patient may come and say, doctor, I have this


      terrible cough I can't get rid of it.  They don't


      come in and say, doctor, I have mycoplasma in my


      respiratory tract.  Although that may be the cause


      of it, the reason patients come to see us is for


      relief of symptoms.


                In prevention trials, on the other hand,


      we are actually seeking to prevent those symptoms


      from ever occurring, but still here we are talking


      about the relevant endpoints being those actual


      symptoms that patients may encounter.




                So, what is the difference between


      clinical endpoints and surrogate endpoints?  We are


      so used to using surrogates that sometimes we call


      things clinical endpoints that are, in fact,




      surrogates.   The definition of a clinical endpoint


      is actually fairly simple.  It is measures of how


      the patient feels, functions or survives, and a


      simple way to think of it is anything that measures


      something other than that is a surrogate endpoint.


      For instance, clinical endpoints would be measures


      of mortality or resolution or prevention of


      symptoms of a disease.


                On the other hand, surrogate endpoints are


      laboratory measurements or physical signs used as a


      substitute for a clinical endpoint.  Fever is a


      surrogate endpoint.  Fever does not necessarily


      measure how the patient feels.  Although fever may


      make the person feel terrible, what we really want


      to measure is the person feeling terrible not what


      the level of the temperature is but we are so used


      to using this in infectious disease trials.  But


      other things like culture results, which we are


      going to talk a lot about today, chest x-rays,


      histology or even data like pharmacokinetic


      information are all surrogate endpoints and need to


      be correlated with what is actually clinically




      happening to the patient.


                The important part here, as discussed at


      NIH Biomarkers Definition Working Group, published


      in 2001, is that surrogate endpoints by themselves


      do not confer direct clinical benefit to the


      patient and we need to make that link.  This is


      also reiterated in the International Conference on


      Harmonization, ICH E9 document.  The International


      Conference on Harmonization is a group consisting


      of U.S., Japanese, European regulators and members


      of the pharmaceutical industry.




                So, how do we differentiate biomarkers


      from surrogate endpoints?  Biomarkers are any set


      of analytical tools that are used to assess


      biological parameters so it is a big, broad


      category.  Biomarkers are useful for many other


      purposes other than surrogate endpoints in trials.


      This is why the term surrogate marker isn't really


      very helpful to us because we can use these


      biomarkers for any number of things.  One may be as


      a diagnostic tool.  We can use the test as




      inclusion criteria to define the disease based on


      the presence of organisms.  Differentiating


      diagnosis from endpoint is a very, very important


      process.  As members of our Anti-Infective Drugs


      Advisory Committee that are here will tell you, we


      have had several advisory committees for instance


      addressing acute otitis media in children and acute


      bacterial sinusitis in children and adults where we


      have tried to make the distinction between needing


      microbiologic data to diagnose that the person


      actually has the disease, but how useful it is as


      an endpoint is an entirely different consideration.


                We can also use biomarkers to describe the


      mechanism of action of the drug and the effect on


      the organisms of an antibacterial or antiviral


      product is really the mechanism by which it


      achieves its effect, not necessarily the goal of


      therapy alone.  We have certainly been told by a


      number of sponsors--the direct quote, all


      antibiotics do is affect organisms.  Well, that is


      true but that is the mechanism by which they do


      what they do, not the goal of why we give them to




      patients in the first place.


                The third thing is that biomarkers can be


      a risk factor for acquiring the disease.  For


      instance, we know that colonization with a


      particular organism is a risk factor for getting an


      infection.  That doesn't mean that risk factors end


      up being the same thing as an endpoint.  Also, some


      of these things can be risk factors for outcome.


      They can indicate disease prognosis and how poorly


      or well the patient is going to do.  For instance,


      HIV viral load and CD4 counts in HIV--we can look


      at those to actually predict how a patient is going


      to do down the line. Then, finally, biomarkers can


      be used as surrogate endpoints, which are different


      from the previous four things we talked about.




                The word surrogate comes from the Latin


      root surrogatus, which means to choose in place of


      another, or to substitute or put in place of


      another.  So, what we are doing with a surrogate


      endpoint is actually substituting microbiologic


      outcomes in patients for clinical outcomes.  One of




      the problems in looking at this is that


      investigators have looked at people only who have


      failed and then tried to relate clinical and


      microbiological outcomes in only the failures.  But


      we need to look at these correlations both in


      people who succeed and people who fail, which is


      pivotal in these clinical trials to prove drug






                Surrogate endpoints are very useful.  They


      can be used in early drug development as proof of


      principle that the drug has some biological


      activity, and they can be used in selecting


      candidates to go on and study in future phase 3


      trials.  They are also useful in phase 3 trials


      when the surrogate endpoint can be measured sooner


      in time than the clinical endpoint.  The obvious


      example of this is HIV trials, which I will go into


      in a little more detail.


                When the clinical endpoint events are more


      rare it allows us to complete a trial with a


      smaller sample size.  In other words, if the effect




      on the surrogate endpoint is quite large and the


      effect on the clinical endpoint is small, we can do


      a trial with a smaller amount of patients in a


      shorter amount of time.  Of course, this is all


      predicated on knowing that the surrogate actually


      predicts clinical outcomes.


                Some examples of where the agency has


      allowed surrogates and they have been used


      successfully are things like lowering cholesterol


      which, in turn, has been shown to prevent


      cardiovascular disease; lowering blood pressure to


      prevent cardiovascular disease; and perhaps the


      best example is suppression of HIV viral load as a


      surrogate endpoint in the prevention of either


      AIDS-defining events or death in the treatment of


      HIV and AIDS.




                In this example what we see is a


      three-dimensional graph.  On the right-hand side


      there are CD4 counts which actually are predictors


      of the host's immune response.  On the other axis


      is the viral load, or HIV RNA concentration.  On




      the upward axis there is the three-year probability


      of patients progressing to AIDS.  You can see from


      this that as the person's CD4 count declines and as


      the HIV viral load goes up, the risk of developing


      AIDS-defining events and death also goes up.  So,


      both HIV viral load and CD4 counts are predictors


      of what is going to happen to the patient




                The interesting thing about this is that


      this is measuring the organism but CD4 count is


      also measuring the host's immune response.  HIV is


      very unique in that the virus itself blunts the


      host's immune response so one of the things that


      complicates the measurement of surrogates is that


      measuring the surrogate itself often doesn't


      measure what is happening to the person.  So, viral


      load is very unique in that the virus itself knocks


      out the immune response and takes that piece out of


      the equation.




                So, HIV viral load and CD4 counts are also


      a good example of the difference between risk




      factors and endpoints.  Both HIV viral load and CD4


      counts are risk factors for disease progression to


      HIV and AIDS, as I showed you on the previous


      slide, however, only HIV viral load functions well


      as a surrogate endpoint, much better than CD4 count


      does in clinical trials.


                Seven of eight trials with a positive


      effect on CD4 count also showed a positive effect


      on progression to AIDS or death.  But the effect in


      6/8 trials that had a positive effect on CD4 count


      also showed a negative effect on AIDS progression


      or death.  This again gets back to the issue that


      you cannot cherry-pick which studies you like.  You


      need to look at both success and failure of the


      surrogate to be able to get an overall assessment


      of what is going on here.  If we only looked at


      these studies we would think that CD4 count was


      great as a surrogate endpoint.


                This also gets to the issue that how you


      use the surrogate is very important.  It may be


      that CD4 count would function as a decent surrogate


      endpoint if we followed patients for longer periods




      of time than we follow the viral load because it


      just may be that the CD4 count may not change fast


      enough over the time that we measure it in a


      clinical trial to be very useful.  But if we


      measured it for longer, that may be a different






                What are some of the strengths and


      limitations then of evaluating surrogates?  Part of


      this is the logic string we go through as related


      here to topical antiseptic products.  We know


      colonization with organisms precedes infection and,


      therefore, the surrogate may be useful as a risk


      factor for disease.  We know that these organisms


      can cause infection and result in a host response.


      So, the logic is that since the organisms cause


      infection, eliminating or decreasing the organisms


      should result in positive clinical outcomes for


      patients.  This seems very logical.  It seems very


      objective and reproducible.  But the question is,


      is it correct?


                This article by DiGruttola, and Dr.




      Fleming is a co-author on this, talks about are we


      being misled in terms of looking at these


      surrogates?  What we just did up here was an


      example of the old Arthur Conan Doyle Sherlock


      Holmes deductive reasoning.  We worked backwards


      from the end and said, well, it must be caused by


      this.  However, what we do in clinical trials is


      inductive reasoning.  We start off with a


      hypothesis and we test the hypothesis.  So, we need


      to test this logic to see if it is actually true.


      One of the seminal articles on surrogates was


      written by Prentice where he actually says that in


      a given clinical trial we need to test does the


      intervention have an effect on the clinical outcome


      and, in the same trial, does that intervention also


      have an effect on the surrogate so that we can link


      the two together?




                Well, why may it be that an intervention


      having an effect on a surrogate which, in turn, has


      an effect on the clinical does not predict what


      actually happens to the patient?  And there are




      five potential reasons why this may happen.


                The first is that there may be unmeasured


      harms caused by the intervention which actually are


      not picked up by just measuring the surrogate.


                The second is that there may be unmeasured


      benefits, that the intervention actually does


      something good that is not measured by the


      surrogate and actually has a better clinical


      outcome than predicted by the surrogate.


                The next issue is that there may be other


      pathways of disease that result in a clinical


      endpoint that have nothing to do with the


      intervention that you applied.


                Finally, there are issues with how we


      measure the surrogate and how we measure the


      clinical endpoint.  Let's go through each one of


      those one at a time.




                As I said, surrogates may not take into


      account unmeasured harm and benefits.  This gets to


      the issue of we cannot just look at whether a


      surrogate correlates with a clinical endpoint




      because, even if there are these unmeasured harms


      and unmeasured benefits, there will still be an


      association between the surrogate endpoint and the


      clinical endpoint.  It will be, however, that that


      association is not predicting the net clinical


      outcome in patients because it is not taking into


      account these other unmeasured benefits and harms.


                It is not too hard to understand why this


      occurs because the body actually has a finite


      number of processes to accomplish the things it


      wants to accomplish.  So, giving a drug product is


      still giving a foreign antigen to the body which


      may affect processes other than the ones that we


      actually intended to affect in the first place.  We


      know that, for instance, in antimicrobial products


      what we are really trying to affect is the organism


      which, in turn, has a positive effect on the host.


      The reason why we get adverse events is that all of


      these products have some effect on the host that is


      unintended in terms of adverse events.




                What are some examples of unmeasured




      benefits?  Well, there may be effects of the drug


      other than eradication of the organism.  Actually,


      this is a misnomer.  We constantly use this term


      "eradication" but what we really mean is that we


      have suppressed the organism to below a level of


      detection.  If we think that we are actually


      sterilizing somebody's body, we really are fooling


      ourselves.  There may be sub-inhibitory effects of


      antimicrobials on the organisms.  Even though those


      organisms are present, they can't do what they


      normally do in terms of invading.  It may be that


      we don't need to kill the organisms to actually


      have some effect on the ultimate outcome and,


      again, that may be because we are having other


      effects, other than killing, that do something to


      the organism.  Then, again, there may be direct


      effects of the antimicrobials on the host immune


      system.  These articles that I have shown up here


      are actually things that talk about the effect of


      antimicrobial products on white cell phagocytosis


      and other processes on the human immune system.


                There also may be unmeasured harms in




      terms of deleterious effects on the host that may


      promote infection.  For instance in topical


      products, if a product actually would cause


      micro-breaks in the skin that would not be visible


      to either the infection or the patient that may


      allow more invasion of organisms to cause wound


      infections.  We also may have replacement of one


      organism with another.  We get rid of the one


      organism we are worried about and, nature abhors a


      vacuum, and something else comes in its place that


      is actually worse than what we got rid of.  There


      may be other sources of infection, other than those


      affected by the drug.




                Are there some examples of where we have


      seen this happen in the past?  The answer is yes.


      This is why we have such pause when evaluating


      surrogates.  For instance, last year the FDA


      approved rifaximin as a treatment for travelers


      diarrhea.  If one evaluates the rate of negative


      cultures from the stool in rifaximin compared to


      placebo, there was no statistical difference




      between the number of organisms at the end of


      treatment in the stool in patients who received the


      drug versus those who did not.


                Regardless of that, there was still


      decreased time to resolution of diarrhea with


      rifaximin compared to placebo.  You could say,


      well, that means rifaximin isn't acting as an


      antibacterial agent; it is doing something else, it


      is decreasing GI motility.  Well, if that is the


      case, then why did rifaximin have an effect on some


      organisms like E. coli, but not on diarrhea caused


      by other organisms like Campylobacter?  If it was


      just acting as a motility agent it should have


      equal effects on everything.  So, perhaps this drug


      is doing something to the organisms other than


      killing them.


                Other examples of unmeasured harms--well,


      a classical example of this is the dose escalation


      trial of clarithromycin that was studied at 500,000


      and 2,000 mg for disease due to Mycobacterium


      avium-intracellulare in patients with AIDS.  When


      we looked at that dose response, the higher doses




      had higher rates of negative blood cultures for


      MAI.  However, those higher doses also had higher


      mortality in terms of the clinical outcomes.  So, a


      better microbiologic outcome actually resulted in a


      worse clinical outcome in this trial.




                Are there also other pathways of disease


      that may be unaffected by the intervention?  Do we


      have an example of that?




                Well, several trials showed decreased


      rates of colonization in the nose with Staph.


      aureus with intranasal mupirocin.  However, three


      trials now done in the last several years show that


      prevention of infections with mupirocin, the


      clinical outcome, was not lower in patients than


      placebo even though there was a dramatic effect in


      terms of negative cultures done from the nose with


      this particular product.  One hypothesis for why


      this may not be effective is that Staph. aureus is


      on numerous sites on the body other than just your


      nose and we may not be affecting that just by




      putting a product on one site in the body.




                The next issue is with accuracy of how the


      surrogate is measured.  One of the things that we


      constantly hear about surrogates is that they are


      reproducible.  Well, reproducibility talks about


      precision, but the example you can think about here


      is how to differentiate precision from accuracy.


      If I take a bow and arrow and I shoot it at a


      target I can hit the same spot on the target all


      the time, but it may be way far away from where the


      bulls eye actually is.  So, even though we are


      getting reproducibility, are we getting accuracy?


      Are we getting the correct inference?  This has to


      do with what, when, how and the magnitude of what


      is measured for that particular surrogate.




                The culture techniques that we use for


      bacteria are based on methodology actually from the


      late 1800's.  We know that there is inherent error.


      For instance, if we take the exact same colony of


      organisms and measure it two separate times we can




      get minimum inhibitory concentrations for a


      particular drug that are actually off by one or two


      tube dilutions jut by testing it a second time.


      So, we know that there is some inherent error here.


                There are a lot of issues with


      microbiological outcomes.  For instance, what is


      the patient population that we sample?  What is the


      sampling technique that was used?  What was the


      methodology used to get the culture?  Actually, I


      see Al Sheldon sitting in the back.  When he used


      to work for us he gave a great talk last year on


      diabetic foot infections where we talked about how


      superficial cultures from the foot may not tell us


      anything related to deeper cultures from the foot


      in diabetic infections, and that methodology is


      very important.


                When is the culture performed?  On therapy


      cultures may be very misleading because when we


      take a sample we are actually taking the antibiotic


      with it and putting it onto the culture plate as


      well, which may give false-negative cultures.


                How often do we sample, and what is a win?




      What is the criteria for classifying that this


      organism is there or not?  Do we have an all or


      nothing approach that says bug present/bug not


      present?  Or, do we so something like HIV viral


      load where we have a quantitative assessment of how


      much organism is present?




                The quantitative assessment may be very


      important, as I show on this graph.  On the bottom


      axis we have time where we can make a baseline


      measurement and on therapy measurement and what


      happens when a drug is gone after the study is


      over, compared to microbial load.  If one patient


      starts out at a higher level than the other


      patient, they both may decrease simultaneously at


      exactly the same rate, but if we make an on therapy


      assessment this patient may still have a positive


      culture and this one does not just because we have


      gone below some level of detection of how many


      organisms we can actually detect.  Does that mean


      that these two patients are really different?  We


      don't know.  It may just be a factor of how many




      organisms we were actually able to detect.  If we


      only looked at an on therapy assessment, that may


      not tell us what happens after the drug is removed


      from the body.  In one patient the bugs may come


      roaring back because all we did was suppress them.


      In the other patient it may continue to decline and


      we get rid of the organism altogether.




                One of the issues that I am sure we will


      talk about today is this issue of practicality, and


      practicality is in the eye of the beholder when it


      comes to clinical trials.  People have said because


      it is difficult to measure the clinical endpoint we


      should just rely on surrogates, which is very


      difficult logic in terms of perhaps needing to do a


      better job of actually measuring clinical


      endpoints.  An inaccurate measurement of clinical


      endpoints does not justify the use of unvalidated






                For example, there is a recent article,


      and there has been an ongoing debate in the




      Clinical Infectious Disease journal about the


      utility of catheter tip decolonization which, in


      this study, are claimed to be validated as a


      surrogate endpoint for clinical trials in


      prevention of catheter-related bloodstream


      infections based on the correlation of the two


      endpoints.  What they did, however, in these trials


      is they defined a bloodstream infection in some of


      these trials as a positive blood culture and a


      positive culture of a catheter tip.  So, this


      correlation is highly dependent upon the definition


      of the clinical endpoint.


                Dr. David Patterson, from the University


      of Pittsburgh, wrote in about one of these studies


      and said, residual antimicrobial activity in the


      removed catheter sufficient to prevent growth from


      the cultured catheter segments would substantially


      reduce the apparent rate of catheter-related


      bloodstream infections--and I put the emphasis on


      there--could it be that use of these coated


      catheters impregnated with antibiotics prevents


      growth from catheters in the microbiology




      laboratory but does not eliminate the clinical


      syndrome of catheter-related bloodstream infection?


                So, a more rational use of an endpoint


      here would be all people that have positive blood


      cultures and symptoms of a clinical infection, not


      just those who have to have a positive catheter tip


      because that is circular reasoning.


                One of the issues we always get into at


      the FDA is what gets published is all the


      successes, and people will look at those and say,


      look, there is this great correlation.  What is


      missing, and there has also been a lot in The New


      York Times recently, is about negative trials.


      What is missing is the data the FDA sits on showing


      where those surrogates did not work.  We have had


      several examples now, both in catheter tip


      decolonization and in products that are actually


      put on topically around the catheter site, where


      they had a dramatic effect on decolonizing the


      catheter and no effect at all relative to placebo


      in preventing bloodstream infections.  I cannot


      enlighten you anymore than that because this is




      proprietary information and we can't share it, but


      the interesting thing sitting at the FDA is you


      always wish that you could talk about the negative


      examples but, unfortunately, we can't share those.




                One of the other issues with correlating a


      surrogate is how well does it actually predict


      outcomes?  A perfect correlation would be a slope


      of 1 in terms of evaluating the surrogate related


      to clinical success so an 80 percent success rate


      with a surrogate would result in an 80 percent


      success rate in the clinical outcomes.  But we


      don't expect that to happen, especially in


      prevention trials where we know that a good number


      of people on these trials will achieve no benefit


      from the product.  So, what we want to look at is


      what is the actual correlation between the


      surrogate and the clinical outcome.




                The other thing that is very important is


      that the correlation may differ from drug class to


      drug class or from drug product to drug product,




      and this may actually be highly misleading in terms


      of what we actually measure.  For instance, let's


      take drug A and drug B that have two different


      correlations in terms of the clinical and the


      surrogate.  If we did then a measure of drug A and


      drug B in terms of the surrogate, it appears here


      that drug B is better than drug A in terms of the


      outcome with the surrogate.  But if these two


      slopes of the correlation are different what


      actually is misleading is that in reality drug A is


      actually better than drug B in terms of clinical


      success so the surrogate actually flip-flops these


      and misleads us in terms of telling us why would


      these slopes be different.


                That gets back to the five things we


      actually talked about.  Unmeasured harms,


      unmeasured benefits and those other things may be


      why these products have different correlations.  We


      actually did this with otitis media and showed that


      the spread of lines here actually goes from 0.4 all


      the way down to 0.1 for various different drug


      products.  So, saying that this won't occur--we




      have actually seen places where this correlation is


      actually all over the map for various drug






                Finally, there are regulatory issues with


      surrogate endpoints.  Traditional approval is based


      on surrogate endpoints only in cases where the


      endpoint is already validated to predict clinical


      benefit.  However, there is an accelerated approval


      clause in the Code of Federal Regulations based on


      surrogate endpoints for serious and


      life-threatening diseases, otherwise known as


      Subpart H.  This is where a surrogate endpoint is


      reasonably likely to predict clinical outcome.


      However, this part of the Code of Federal


      Regulations requires confirmatory post-approval


      trials based on the clinical endpoint to prove that


      what we saw with the surrogate is actually true.


                The important thing to note today is that


      this clause actually came out in the mid-1990's and


      what we are talking about today is a monograph that


      started out in the early 1970's.  So, if you ask




      the question, well, why doesn't the monograph jive


      with what we are saying up here, it is because we


      are talking about something that happened 20-30


      years before this regulation.




                Let's relate all of the stuff we just


      talked about with surrogates to the issues related


      to topical antiseptics.  Are there some potentials


      for unmeasured harms with topical antiseptics?


      Well, we may have unintended effects on microscopic


      breakage in the skin which may actually result in a


      greater clinical infection rate.  We know this can


      happen, for instance, in trials that examine


      peri-operative shaving.  This trial by Seropian,


      done in the American Journal of Surgery in 1971,


      actually showed a 5.6 percent rate of postop


      infection with shaving compared to a 0.6 percent


      rate without shaving.  So, we know that there can


      be unintended effects.


                If you go back and look at the hypothesis


      of that trial, it was exactly what we are trying to


      say today, clipping hair off may decrease the




      amount of bacteria near the wound and, therefore,


      should result in a decrease in infections.  It


      didn't; it did the exact opposite because of


      unintended harms that they didn't think about until


      after the trial was done.  It is always fascinating


      to see how someone's hypothesis changes after the


      actual results come out.


                Also, the effects on common pathogens may


      be less than that on the marker organisms on the


      skin.  Michelle Jackson showed you that what we are


      measuring here is resident microbial flora in two


      of the three indications and we are contaminating


      people with Serratia marcescens in another.


      Serratia marcescens is not a common cause of skin


      infection so the question is does predicting an


      effect on Serratia tell us anything about staph.,


      strep., E. coli, enterococci and the other common


      causes of infection?


                Also, there is this issue of are we


      selecting resistance to systemic antimicrobials by


      using these topical antibiotic products?  This


      really is something that deserves its own whole




      discussion, but there is some evidence at least in


      the test tube that there may be afflux pumps which


      confer resistance to both topical products and to


      the systemic antimicrobials simultaneously, at


      least in E. coli and Pseudomonas.  People have


      questioned what is the clinical relevance of that


      but that really is the question, isn't it?  Once


      again, it is how does that surrogate predict what


      is going to happen clinically?  I always think it


      is fascinating when you don't want to use a


      surrogate, all of a sudden it is not relevant.


      When you do want to use a surrogate, we will accept


      everything we want to believe about it.


                So, can there be unintended benefits?


      Well, it may be that some of these products have


      positive effects other than those on the organisms.


      It does something to the host immune system that


      actually results in a decreased infection rate,


      more than we would predict by what it does to the


      bug.  Also, could the effects on common pathogens,


      like staph. or strep. be greater than on something


      like Serratia?  So, it may be a better benefit than




      what we think.




                Are there other mechanisms not affected by


      the intervention?  Well, at least in terms of


      patient preop, for that indication we can look at a


      study that was done by Brown et al. in 1989 at the


      University of Virginia.  The data that we are


      obtaining from this surrogate is really from the


      most superficial layers of the stratum corneum of


      the epidermis.




                Here is an anatomical picture of the skin.


      What you see here is that the top 30 layers of the


      skin are this dead, keratinized layer called the


      stratum corneum of the epidermis.   What is down


      here is the stratum germinativum where these cells


      come from.  The cells die off.  They become highly


      keratinized at the stratum granulosum layer which


      forms a barrier between this and the stratum


      corneum.  What we are measuring in these trials is


      what is way up here.




                So, what is way up there is right here on


      this graph.  This is actually from the CDC


      guidelines on prevention of surgical infections.




      What we are worried about is infections here, here,


      here and here.  So, the real question is does doing


      something up here do something down here in terms


      of affecting the organisms?




                This group in Virginia actually did a very


      elegant experiment with a methodology that was


      developed by Pincus in 1952.  What they did was


      they took regular old cellophane tape and they


      showed that by putting cellophane tape and


      stripping it off the skin you can take one layer of


      that stratum corneum off at a time.  They evaluated


      this in 12 different sites on the body, and they


      showed that these 12 different sites in the body


      had highly variable colony counts of organisms


      depending upon whether you are looking at the arm,


      the back or other sites.


                They also showed that the number of


      colonies decreased over the top five layers of the




      stratum corneum but then stabilized in the


      remaining 20 layers of the stratum corneum.  So,


      there were more organisms up at the top than there


      were in the lower layers of the stratum corneum.


                But then they did something very


      interesting.  They took alcohol and decolonized the


      area that they had stripped, put a gauze pad over


      it and came back 18 hours later.  They then did


      plasmid profiles on the coagulase-negative


      staphylococci that were there at the beginning of


      the experiment and there 18 hours later and saw


      identical plasmid profiles for those staphylococci.


                So, they hypothesized that this indicates


      a reservoir for these organisms that may be below


      the stratum corneum, in the hair follicles and


      sebaceous glands of the dermis so where infection


      may come from is actually from the organisms that


      are lower down.  This is one of the reasons why we


      give systemic antimicrobials as perioperative


      prophylaxis, trying to affect those organisms that


      may be down deeper in the dermis.


                We also know that studies in perioperative




      systemic antimicrobials show that if the antibiotic


      isn't around at this layer at the time you get


      operated on they will not be effective.  For


      instance, you cannot give the antibiotic two


      seconds before you make the surgical cut because


      they will not affect the subsequent infection rate.




                Then there are all the issues with


      measurement of the surrogate, which we are going to


      talk about today.  Are we actually measuring the


      surrogate in a population that we are going to use


      it in?  No, we are not.  We are measuring healthy


      volunteers, not healthcare workers or patients.


                As we already discussed, the organisms


      measured are not necessarily those that cause


      infection.  Is the timing of these measurements


      relative to the disease process we are actually


      trying to prevent?  That gets at this issue of do


      we need to get persistent effect or not; how long


      do we have to look for that; and how long should we


      look for it?  For instance, we know that some


      patients may undergo prolonged surgery.  Surgeries




      may last hours and hours so an immediate effect is


      not the only thing we want to look at.


                Are the conditions of testing the same as


      those that would be encountered in real-life


      situations?  And, what happens with variations in


      the methodology?  One of the things that is


      interesting at the FDA is that you will see people


      submit things that say I am using the such-and-such


      method approved by the CDC or the NIH.  But it is a


      modified method.  I always joke I am a modified


      millionaire movie star; I am just not a movie star


      and I don't have a million dollars.  So, modifying


      the method--it is no longer the method.  So, we


      need to take into account that changing the method,


      even if we have a valid surrogate, may actually


      change the correlations between the surrogate and


      the clinical outcomes.


                The next question is what log reduction is


      clinically significant?  And, how do we analyze


      those numbers obtained on log reductions?  Dr.


      Thamban Valappil is going to go through a great


      talk that actually walks through some of these




      issues with how do we analyze the numbers.




                What is the data showing correlation of


      reduction of bacteria with a decrease in infection


      rates?  Steve Osborne is going to go through our,


      believe me, exhaustive, over 1,000-paper literature


      search.  You should have helped us out with this;


      that was a thrill!


                What does the dose-response curve look


      like for infection rates and numbers of bacteria?


      Is it a threshold effect, or is it a continuous


      variable, and is it the same for all types of






                What do I mean by dose response?  Down on


      the bottom it should read numbers of bacteria on


      the skin, not change in numbers of bacteria.  On


      the Y axis we have rates of infection.  What we


      want to know is does the dose-response curve look


      like this?  Sorry, this doesn't show up very well


      but it is a straight line.  Or, does the


      dose-response curve look like this?  The first




      straight line is a continuous variable.  The more


      organisms there are, the more infections patients


      get.  The curved line is really a threshold effect


      that we talk about.  At some certain level of


      bacteria people are more likely to get infected and


      below that level they are less likely to get




                Why is this important for us?  Well, if we


      look at a linear correlation between numbers of


      bacteria and rates of infection, what we will see


      is that the decrease of the numbers of bacteria by


      this much will actually result in a corresponding


      decrease in the number of infections by some






                On the other hand, if it is a sigmoidal


      threshold type effect, what we will see is that


      that same, exact change in the number of bacteria


      if it is on the flat part of the curve results in


      very little change in infection.  So, this gets to


      what does a 3-log reduction actually mean?  If this


      is 10                                       7 and this is 104 that is a

3-log reduction but




      we are on the flat part of the curve so there is


      very little effect on what happens to the patient.


      If we go from 10                                                        

4 to 101 that is a 3-log reduction


      too but if we are on the steep part of the curve


      that may be telling us something very, very


      different.  So, where you start may be as important


      as what the delta change is, and we don't have any


      information to tell us what this dose response


      actually looks like.




                What I would like to leave you with then


      is sort of the thought process we have had to go


      through for the last several months in terms of


      trying to look at this.  The first question you


      have to ask is what kind of endpoint are you going


      to pick to evaluate these products?  Are we going


      to pick a clinical endpoint or a surrogate


      endpoint?  Ideally, there would be the data right


      here that links the clinical and the surrogate


      endpoint together, and Steve Osborne is going to


      talk about our attempts to actually make that kind


      of a link.


                The second question is what are we


      actually going to measure?  Let me get back to this


      issue of practicality.  As I said earlier,




      practicality ends up being in the eye of the


      beholder.  One of the things you will hear about is


      that it takes more patients to do these clinical


      trials than it does to the surrogate endpoint




                Well, size is actually an issue but size


      really relates more to the time that it takes to do


      a trial which, let's be honest, relates to cost to


      do the trial.  One of the questions you have to ask


      when you are getting into this debate is how much


      does it cost to do it wrong?  How much does it cost


      the patients if we don't get this information and


      we don't actually know whether these products are


      effective?  That side of the equation needs to be


      factored in as well.


                The other issue that comes up is ethics.


      Ethics are only if you are denying somebody a


      proven effective treatment.  What we are trying to


      evaluate here is are these things proven effective




      or not, so we need to keep that in mind when we are


      discussing the ethics issue.  When we talk about


      clinical trials the endpoint is very simple, it is


      infection in patients.  On the other hand, with the


      surrogate we are looking at numbers of bacteria.


                Then we need to talk about how do we


      design these studies and how do we define success.


      Well, the definition of success, again, with the


      clinical endpoint is much simpler actually.  It is


      just the percent of patients that don't get an


      infection.  However, when we talk about selecting


      an endpoint for a surrogate we have several


      decisions to make that Thamban is going to go


      through.  Do we look at mean log reductions, median


      log reductions, the percent of subjects who meet


      some log reduction?  And, where do you get this


      information from?  Well, actually optimally it


      would be from a clinical trial that evaluated both


      of these things simultaneously.


                Finally, how do we analyze the results


      that we get?  Again, it is much simpler in a


      clinical trial.  We just compare it with a




      concurrent control.  This is one of the issues when


      people point to the studies, and Steve is going to


      go through this in some detail, they say we already


      know these things work.  There is no concurrent


      control.  What these things are is quasi


      experimental studies where they took what we were


      doing last year and they applied something new in


      the hospital and said, look, my infection rate went




                What that ignores is natural changes in


      baseline infection rates that may occur.  Even


      though the trials say, well, we didn't do any other


      interventions on these patients, you know in the


      real world and, hopefully our AIDAC members can


      enlighten us on this, when you have an outbreak of


      some particular organism you do not do one


      intervention.  You cohort patients together; you


      start using gowns and gloves on those people; you


      do a lot of other interventions that really call


      into question what was the cause of why the


      infection rate went down.  Was it just related to


      the product that you used?


                So, here we would make this comparison and


      either design these as superiority or


      non-inferiority trials, otherwise called




      equivalence trials, that show that the product is


      no worse than something that is already out there.


                On the other hand, there are a lot more


      complex decisions with a surrogate endpoint.  Do we


      say that these things meet some threshold that we


      set?  If so, where does that threshold come from?


      Where does the data come from to say?  And, do we


      still need some comparison with a control given the


      variability in the method?  Michelle Jackson showed


      you on one of her slides that at least that article


      in The Journal of Hospital Infection, based on the


      European methodology which is slightly different


      from that that is in the TFM, shows at least a 2 to


      2.5 log drop with soap and water all by itself.


      So, do we need to look at how these things compare


      to some vehicle or another product?  And, again, we


      have the choice of superiority or non-inferiority.




                To conclude then, surrogate endpoints must




      not only correlate with clinical outcomes but they


      must also take into account unmeasured harms and


      benefits; the methodology and uncertainties in


      measuring the surrogate; and the appropriate


      measurement of the clinical endpoint.


                The clinical endpoint for efficacy of


      topical antiseptic products would be prevention of


      infections but actually the clinical design of


      these trials would vary depending upon whether we


      are talking about patient preop surgical hand


      scrubs or healthcare personnel handwash.


                One of the things that I am sure we will


      hear about is what Semmelweis did in 1847 was he


      showed that medical students who went and examined


      corpses with their bare hands and then went and


      delivered babies--there was actually a higher rate


      of death in the mothers who had their babies


      delivered by these medical students than the


      midwives who were spared the odious task of doing


      the autopsies.


                That is not what we are doing today.  We


      are not digging our hands into gram-negatives of




      dead people and then going and operating on


      someone.  So, the conditions of Semmelweis were


      huge bacterial load, probably with gram-negative


      organisms.  So, what Semmelweis showed was that


      washing your hands is a good thing.  Semmelweis did


      not do a randomized trial of one product compared


      to handwashing alone or handwashing compared to


      nothing.  We are not debating that Semmelweis was


      correct and that you need handwashing.  What we are


      debating is handwashing with what, and how do we


      determine that that "what" is effective compared to


      just maybe plain soap and water?  So, we are going


      to discuss further today what is known about


      surrogates in the setting of topical antiseptics,


      and Steve Osborne is going to go over this clinical


      correlation and tell us some more about it.




                I would like to leave you with this quote


      by the statistician John Tukey which I think really


      relates to surrogates:  Far better an approximate


      answer to the right question, which is often vague,


      than an exact answer to the wrong question, which




      can always be made precise.  I will stop there.


      Thank you very much.


                DR. WOOD:  Thanks very much.  It appears


      that we still don't have the slides from Michelle


      Pearson.  Is John Boyce here?  Yes?  Good, so at


      least our next speaker is here.  I suggest that we


      take a quick break right now and be back at ten


      o'clock and we will start again.  We are hoping to


      get Michelle Pearson in before we do the questions.


      We will get back at ten o'clock.


                [Brief recess]


                DR. WOOD:  Let's go to Dr. Boyce and then


      we will come back to Dr. Pearson, whose talk we do


      now have somewhere in the building, as they say,


      but we have been unable to play it yet.  So, Dr.




               Antiseptic and Infection Control Practice


                DR. BOYCE:  Good morning.  I am having


      some Power Point problems today because of a switch


      in versions so I hope this is going to work.




                First I want to talk a little bit about




      the importance of hand hygiene in preventing


      transmission of healthcare-associated infections.


      Most of you know that transmission of


      healthcare-associated pathogens often occurs via


      transiently contaminated hands of healthcare


      workers.  For that reason, handwashing has been


      considered one of the most important infection


      control measures for preventing


      healthcare-associated infections.  Despite this,


      the availability of published handwashing


      guidelines has not helped, and compliance with


      healthcare workers with recommended handwashing


      practices has remained low for decades.




                This slide shows the percent compliance on


      the Y axis in 37 published observational studies of


      healthcare worker handwashing compliance.  The main


      point here is that compliance rates varied from


      about 5 percent to 80 percent.  The second point is


      that there is no trend towards improvement over


      this more than 20-year period.  So, getting people


      to wash their hands as frequently as possible has




      been a very difficult chore.




                In 2002 the CDC published the guideline


      for hand hygiene in healthcare settings.  I am


      going to briefly mention a few indications for hand


      hygiene that are listed.  One is that it is


      recommended that we wash our hands with a


      non-antimicrobial soap or an antimicrobial soap if


      our hands are visibly contaminated with blood, body


      fluids or other proteinaceous materials.  If the


      hands are not visibly soiled, then the guideline


      recommended the routine use of an alcohol-based


      hand rub for decontaminating hands in most other


      clinical situations.  Alternatively, hands can be


      washed with an antimicrobial soap and water in


      other clinical situations.


                The guideline recommends that healthcare


      workers decontaminate their hands before having


      direct contact with patients, donning sterile


      gloves to insert a central intravascular catheter,


      before inserting indwelling urinary catheters or


      peripheral IV catheters, and before eating.




                It is recommended that we decontaminate


      our hands after having direct contact with a




      patient's intact skin, like taking a blood


      pressure; contact with body fluids or wound


      dressings if our hands are not visibly soiled;


      after moving from a contaminated body site to a


      clean body site during an episode of patient care;


      after contact with inanimate objects in the


      immediate vicinity of the patient; and after


      removing gloves.  So, there are a lot of


      indications for cleaning your hands.




                In fact, the number of hand hygiene


      opportunities that healthcare workers have can vary


      considerably.  In a large study, done by Dr.


      Pittet, they found that the average number of hand


      hygiene opportunities per hour of care was 24 in


      pediatric units, and the average was 43 per hour in


      intensive care units.  In fact, the lack of


      sufficient time to actually perform this large


      number of handwashing episodes is a major factor




      influencing poor handwashing compliance.




                This slide shows the results of a number


      of observational studies where healthcare workers


      were observed to see how many times they actually


      cleaned their hands.  You can see on your right


      that the average number of times per 8-hour shift


      was anywhere from 13 times to 26 times in an 8-hour


      shift.  So, we are talking about frequent use of


      these products.


                That sounds pretty frequent but let me


      present it another way, in a recent prospective


      trial that we conducted that involved 57 volunteer


      nurses working in intensive care units, a


      hematology-oncology ward and general medical ward,


      each nurse carried a portable counting device and


      prospectively clicked the counter every time they


      cleaned their hands.  On the right you see a graph


      that, along the X axis, shows the number of hand


      hygiene episodes that these nurses recorded during


      a 3- to 3.5-week trial period.  You can see that


      most nurses cleaned their hands anywhere from 100




      to 450 times in a 3- to 3.5-week period.




                So, one thing that is very clear is that,


      because of the high frequency of use of these


      products, providing healthcare workers with


      products that are well tolerated is very important.


      Poorly tolerated products result in poor compliance


      often because of irritant contact dermatitis, as


      shown in the picture, where this physician has


      bleeding knuckles after using soap and water


      handwashing 57 times over a period of a couple of


      weeks.  Products that have a high degree of


      antimicrobial activity, that is, a high log


      reduction, but are poorly tolerated may actually be






                Now, another important issue for which we


      have very little information is what level of log


      reduction of bacterial counts on the hands is


      actually necessary to prevent transmission of


      pathogens.  As you know, the efficacy of these


      agents is often expressed as a number of log




      reductions of bacterial counts on the hands of


      volunteers, 1, 2 or 3 log reductions for example.


                Although the review of the literature that


      I did apparently is not as big as what FDA has


      actually done, I reviewed over about 700 articles


      and couldn't find any evidence regarding the number


      of log reductions that are necessary to prevent


      transmission of healthcare-associated pathogens.


      So, we just don't know how many log reductions we






                Another thing for which I think there is


      little, if any, data relates to whether or not we


      need products that have a cumulative effect.  As


      you know, the tentative final monograph requires


      that healthcare personnel handwash agents produce a


      2-log reduction after the first wash and a 3-log


      reduction after the 10th wash, therefore showing a


      cumulative effect.


                In the review of the literature that I did


      I failed to identify any data supporting the need


      for a cumulative effect.  As a clinician with 25




      years of experience working in hospitals, I am not


      aware of any evidence that patients who are cared


      for in the middle or at the end of a work shift are


      at higher risk of infection than those that are


      cared for at the beginning of a shift.  I am also


      not aware of any evidence that patient care


      activities that are performed in the middle or near


      the end of a work shift result in greater hand


      contamination than those that are performed at the


      beginning of a shift.  So, frankly, from the


      standpoint of a clinician or of infection control,


      I fail to see the logic in requiring a cumulative


      activity of this type of product given the way they


      are used and the types of patients that we take


      care of.




                Another thing that actually has changed


      since the TFM was originally developed is the


      frequency of glove use.  Since the late 1980's


      nurses, physicians and other healthcare workers use


      gloves far more frequently than they ever did in


      the past.  A recent observational survey done of




      nurses working on a general medical ward found that


      these nurses visited patients an average of about


      54 times during an 8-hour shift, and they found


      that the use of gloves varied depending on the type


      of patient care activity.  When the nurses were


      going to have contact with body fluids they wore


      gloves 86 percent of the time.  If they were going


      to have skin contact only, then it was more like a


      little over 30 percent of the time that they wore


      gloves; even less frequently for equipment contact.


      So, in fact, glove use does vary among healthcare


      workers but it is certainly far more common than in


      the past.




                A number of studies, shown here, have


      documented that gloves can and do reduce the level


      of hand contamination when they are worn.


      McFarland looked at hand contamination with C.


      difficile and found that 46 percent of healthcare


      workers who did not wear gloves contaminated their


      hands with C. dif..  No healthcare workers who wore


      gloves had C. dif. on their hands.  Olsen and




      colleagues found that gloves prevented hand


      contamination in 77 percent of instances.  Dr.


      Pittet found that when no gloves were used and they


      measured hand contamination rates, they found out


      that the hands were contaminated with 16


      CFUs/minute of patient care when no gloves were


      used, but only 3 CFUs/minute when gloves were used,


      showing the protective effect of gloves.  Finally,


      Tenorio et al. found that gloves reduced the risk


      of hand contamination by vancomycin-resistant


      enterococci by 71 percent.  So, in fact, to the


      extent that people do wear gloves during patient


      care nowadays, their hands are probably less


      heavily contaminated than they were back in the


      '60's, '70's and early '80's.




                One thing that I thought that I was


      supposed to try to address was whether or not there


      is any evidence that the products that are


      currently on the market have any kind of clinical


      benefit in a healthcare setting.  I wanted to


      mention this model by Ehrenkranz.  It was a field




      study that was supposed to reproduce clinical hand


      contamination.  Nurses touched the skin of patients


      who were heavily contaminated with gram-negative


      bacteria.  They then cleaned their hands.  They


      either used plain soap and water handwashing or


      they used the 63 percent isopropyl alcohol hand


      rinse.  After cleaning their hands, the nurses


      touched catheter material, like a Foley catheter


      type material, and then that catheter material was


      cultured on agar plats.


                What they found is that bacteria were


      transferred from the hands of the nurses onto this


      catheter material in 11/12 experiments when plain


      soap was used to clean their hands but only 2/12


      experiments when the alcohol hand rinse was used.




                Now, in terms of clinical trials, which I


      think is a major issue as was discussed in part by


      the last speaker, this slide shows one sequential


      trial of three hand hygiene regimens.  It was done


      in the surgical intensive care unit by a very


      experienced infection control physician.  They




      looked at non-medicated soap, 10 percent


      povidone-iodine or 4 percent chlorhexidine


      gluconate.  Each product was used exclusively in


      the ICU for 6 weeks.  Surveillance for nosocomial


      infections was performed.  What they found was that


      the incidence of healthcare-associated infections


      was 50 percent lower during times when the two


      antiseptic-containing handwash agents were used,


      suggesting that these hand hygiene products that


      were available at that time reduced infections


      better than plain soap and water handwashing in


      this short trial which was only done in one ICU.




                This slide discusses a prospective trial


      done to compare two hand hygiene regimens.  It was


      a prospective trial with a multiple crossover


      design.  It was done in three intensive care units


      in a university hospital that just happened to have


      one of the largest and most highly respected


      infection control programs in the country at that


      time.  So, they had lots of resources relatively


      speaking.  They followed over 1,800 adult patients




      for nearly 8,000 patient-days at risk.  The two


      regimens compared were 4 percent chlorhexidine


      gluconate versus a combination regimen of isopropyl


      alcohol and a non-medicated soap.  Healthcare


      workers were told that when the alcohol and


      non-medicated soap were available they were


      supposed to use the alcohol routinely for cleaning


      their hands.




                What they found was that the number of


      patients who developed a healthcare-associated


      infection was 96 in the chlorhexidine time period


      and 116 when the alcohol and plain soap were


      available.  So, the incidence density was lower


      with the 4 percent chlorhexidine.  The number of


      healthcare-associated infections was 152 during


      periods when the 4 percent chlorhexidine was used


      compared to 202 when the combination regimen was


      available--again, a lower rate with the 4 percent


      chlorhexidine.  Infection rates were significantly


      lower in 2/3 ICUs when the chlorhexidine was used.




                Despite this being planned by a very


      experienced and highly respected individual, with a


      large team working with him, this clinical trial




      ran into some problems.  First of all, the overall


      compliance of healthcare workers, as shown on the


      left, was not the same during the two trials.  It


      was about 42 percent compliance when the


      chlorhexidine was available versus 38 percent when


      the other regimen was available in the units.  The


      difference was actually statistically significant.


                Another important problem that emerged,


      despite this trial being well planned and designed,


      was that the volume of the products used varied


      significantly.  The amount of soap and isopropyl


      alcohol used when available was significantly lower


      than the volume of chlorhexidine used when that


      product was available.  Even though healthcare


      workers were told they should use the isopropyl


      alcohol routinely when available, for reasons that


      are not either understood or discussed by the


      authors, the healthcare workers hardly ever used


      the alcohol.  So, this trial was really more a




      comparison of 4 percent chlorhexidine against plain


      soap and water for the most part.


                So, one problem with this trial is that it


      is very difficult to control the activities of all


      these healthcare workers in all these ICUs over an


      8-month period, and to get them all to do exactly


      the same thing and to do it with exactly the same






                From the eyes of a beholder here who works


      in a hospital, that is one of the problems with


      clinical trials.  When you use a nosocomial


      infection rate as the outcome measure for efficacy


      of hand hygiene agents, there are many, many


      confounding variables including host factors; the


      rate of importation of organisms from nursing homes


      or other sites into the hospital and onto the


      wards; the level of compliance of healthcare


      workers with recommended hand hygiene, with


      recommended barrier precautions, how frequently


      they follow guidelines for central line placement


      and for ventilator-associated pneumonia prevention.




      If you are talking about surgical site infections


      you have to worry about the skill of the surgeon;


      whether or not prophylactic antibiotics were used


      and timed appropriately; and whether or not any


      active surveillance cultures are being done on the


      wards where the studies are being conducted.


                So, from my viewpoint, there are so many


      confounding variables that that, in and of itself,


      makes the clinical trials extremely difficult to do


      and extremely costly.  To me, it seems like the use


      of surrogate endpoints to assess efficacy of hand


      hygiene products still has merit.




                I want to mention a little bit more about


      clinical benefit.  None of the things I am going to


      mention are carefully controlled, prospective


      trials partly for all the reasons I have just


      mentioned.  This one publication involved a surgeon


      whose hands, but not other body parts, were


      colonized with a virulent strain of Staphylococcus


      epidermidis that caused an outbreak of surgical


      site infections related to cardiac surgery.  This




      surgeon was using a noon-antimicrobial soap for a


      preoperative scrub because of previous problems


      with hand dermatitis so he followed the


      recommendation of his dermatologist.


                An epidemiologic investigation that


      included case control studies and molecular typing


      clearly implicated the surgeon as the source of


      this outbreak, and we told him he had to stop doing


      cardiac surgery and to start using a 4 percent


      chlorhexidine gluconate surgical scrub.  After he


      did so the outbreak terminated and we did not see


      that strain any further in cardiac surgery


      infections, demonstrating that the antimicrobial


      soap that was available didn't appear to have






                An outbreak of vascular surgery-related


      surgical site infections occurred when an operating


      room was not provided standard povidone-iodine.


      The surgeons were used to using preoperative


      surgical hand scrubs.  The vascular surgeons in the


      hospital decided to use plain soap for hand




      scrubbing before surgery, while other surgeons used


      a 2 percent iodine with 70 percent alcohol for


      preoperative hand scrubbing.  Hand scrubbing with


      plain soap was significantly associated with the


      occurrence of this outbreak of surgical site


      infections and reinstitution of povidone-iodine


      hand scrubbing terminated the outbreak, again


      suggesting that this povidone-iodine product had


      value in reducing surgical site infections.




                Of course, the CDC guideline for hand


      hygiene was published in 2002 and the guideline


      recommends routine use of alcohol-based hand


      sanitizers for cleaning hands before and after


      patient contact as long as the hands are not


      visibly contaminated.




                Not long after the guideline was


      published, actually in January of 2003, the Joint


      Commission on Accreditation of Healthcare


      Organizations sent out a sentinel event alert to


      hospitals and recommended that hospitals comply




      with the CDC's new hand hygiene guideline.  So, I


      think both the Joint Commission and CDC are


      standing behind the guideline.




                This study was done where a 70 percent


      ethanol hand gel was introduced hospital-wide into


      the hospital.  A multidisciplinary program to


      improve hand hygiene was carried out.  During the


      following 12 months the alcohol hand product was


      used an estimated 440,000 times by healthcare


      workers and they found a consistent reduction in


      the proportion of all methicillin-resistant Staph.


      aureus that was hospital-acquired during the


      12-month period.




                This slide shows the impact of one of


      these alcohol hand sanitizers on the hand hygiene


      compliance in our hospital.  Compliance rate is


      shown on the Y axis.  Observational surveys


      conducted by the same infection control


      practitioners each time revealed that, by having


      this new alcohol hand gel available and promoting




      its use and educating people about it, the overall


      hygiene compliance improved from 38 percent to 63


      percent, and the proportion of all hand hygiene


      episodes which were performed using the alcohol


      hand gel, which is shown in the red part of the


      bars, increased significantly.


                Not shown on this slide is the fact that


      the proportion of all methicillin-resistant Staph.


      aureus--let me put that another way, the proportion


      of all Staph. aureus isolates that are due to


      methicillin resistance in our hospital levelled off


      about the time that survey 2 was done, and actually


      decreased by 5 percent over the following year and


      a half.  This decrease in MRSA in our hospital


      occurred during the same time frame when MRSA


      continued to increase in prevalence in the


      hospitals that participate in CDC's National


      Nosocomial Infection Surveillance program, or NNIS.


      Although it is rather crude data, we think that the


      hand hygiene program probably has helped reduced


      MRSA in our hospital as well.




                In conclusion, conducting clinical trials


      to assess the efficacy of healthcare personnel


      handwash products is, in fact, extremely difficult,




      expensive and, as far as I am concerned, largely


      not practical.  If they are to be done, they are


      going to be very expensive.


                Widespread experience with currently


      available products, combined with some of the


      epidemiologic studies that I mentioned, provide


      some evidence of their clinical benefit in


      healthcare settings.  Multiple studies have shown


      that promoting the routine use of alcohol-based


      hand santizers, when combined with educational and


      motivational material, can improve hand hygiene


      practices among healthcare workers.




                There are no published data that I am


      aware of demonstrating that cumulative activity of


      healthcare personnel handwash agents or surgical


      scrub products results in lower rates of


      healthcare-associated infections.  Removal from the


      market of hand hygiene products that are currently




      in widespread use in healthcare facilities would,


      in fact, disrupt national efforts to improve hand


      hygiene practices among healthcare workers.  So, I


      personally would hope that there is no regulatory


      action that ends up removing a lot of the current


      products from the market because I am convinced,


      again on a personal level, that they do have value.


      Thank you.


                DR. WOOD:  We have received Dr. Pearson's


      slides from the wilds of Atlanta and we think we


      can show them.  Is that right?


                MS. JAIN:  Yes.


                DR. WOOD:  Unfortunately, sort of like CNN


      breaking news, because the slides are just in we


      don't have a handout.  We are going to have her on


      the phone.  Dr. Pearson, can you hear us?


                DR. PEARSON:  I can.


                DR. WOOD:  As you go through the slides,


      Dr. Pearson, if you tell us when you want to change


      to the next slide, we will be able to do that.


      Let's go.


                 Prevention of Surgical Site Infections


                DR. PEARSON:  Good morning and thanks to


      the meeting organizers for tolerating my


      inconvenience and thank you for the opportunity to




      present on the topic.




                What I hope to do in the next few minutes


      is really to talk about some of the epidemiologic


      complexities of looking at the effectiveness of any


      preventive measure, whether it be cutaneous


      antiseptic or other preventive measures, using


      surgical site infections as the context for that


      discussion.  Next slide.


                What I am going to do is first provide an


      overview of what we know about the epidemiology of


      surgical site infections, including the incidence


      and risk factors for infection.  I will talk next


      about some of the preventive strategies that have


      been shown to decrease that risk; highlight some of


      the current surveillance systems for monitoring the


      incidence of surgical site infections; and conclude


      with talking about how we, here at the CDC, go


      about developing our policies and recommendations




      for prevention of healthcare-associated infections,


      such as SSIs.  Next slide.


                Just to give you a little bit of an idea


      of why this is an important topic and to frame it


      with some numbers, it is estimated that somewhere


      in the neighborhood of 20 million inpatient


      surgical procedures are done each year in the


      United States, and 2-5 percent of these procedures


      are complicated by a surgical site infection.


                Based on our surveillance system, surgical


      site infection is the second most common


      healthcare-associated infection, comprising about a


      quarter of all of the infections reported to CDC.


      These infections come not only at a cost to the


      patient but also a cost to the healthcare delivery


      system.  These infections result in anywhere from


      an additional week of hospital stay and they cost


      anywhere from $400 to $2,600 per infection, and


      these total well in excess, and approaching in some


      instances, close to a billion dollars a year in


      terms of healthcare dollars.  Next slide.


                In terms of the way we define or look at


      surgical site infections at CDC, we classify them


      either as incisional surgical site infections, and




      those include superficial infections which involve


      the skin and the underlying subcutaneous tissue, or


      deep incisional surgical site infections which


      involve the underlying soft tissue as well.


      Obviously, the most severe and costly infections


      are those that involve the underlying organ or


      organ space surgical site infections and those


      involve really any part of the anatomy other than


      the incision that might have been opened or


      manipulated during the procedure.  Next slide.


                This is a cross-sectional schematic to


      illustrate just a little bit more clearly an


      abdominal wall that shows the various


      classifications.  As you can see, a superficial


      incisional SSI would involve the skin and the


      subcutaneous tissue.  A deep incisional SSI would


      extend down into the fascia and the muscle.  The


      organ space surgical site infection, obviously,


      would include the organs in that surrounding




      tissue.  Next slide.


                Now, when we look at the organ or the


      potential sources for the pathogens that result in


      a surgical site infection, overwhelmingly these


      arise from the patient's own endogenous flora.


      There are also secondary sources for the pathogens


      that result in a surgical site infection.  Those


      can result from pathogens that are available in the


      operating room theater environment.  They may


      result from operating room personnel that are in


      and around the surgical field or, not uncommonly,


      at the head of the table of the anesthesiologist.


      Less commonly, these infections may result from


      seeding of the operative site from a distant site


      of infection.  Next slide.


                If we look at the microbiology of the


      surgical site infections--and this slide is


      somewhat dated but suffice it to say that the


      distribution of these pathogens is still


      predominantly--the primary organism are


      staphylococcal infections, not surprisingly because


      these arise primarily from the patient's own




      endogenous flora.  The predominance of these


      pathogens is Staph. aureus, and then with certain


      procedures like cardiac surgery, and more recently


      we have been looking at some data from prosthetic


      joint infections, and it appears that staphylococci


      now account for in the neighborhood of around 50


      percent of the infections causing surgical site


      infections.  We have also seen an increase in the


      proportion of those staph. infections that are due


      to resistant organisms, such as


      methicillin-resistant Staph. aureus.  Next slide.


                Less commonly, SSIs may be due to some


      unusual pathogens, such as the ones shown on this


      slide that are typically due to either contaminated


      products or solutions that are used in and around


      the surgical site, or to colonized healthcare


      workers, again, that might be part of the surgical


      team.  When you see clusters of infections that are


      due to these unusual pathogens you should think of


      a common source, such as the contaminated vehicle


      or potentially the colonized healthcare worker who


      is disseminating the organism.  Next slide.


                Regardless of where the organism arises,


      the pathogenesis of a surgical site infection can


      kind of be distilled into this numerical formula




      and relationship shown here.  That relationship


      really is a combination of the dose or the amount


      of bacterial contamination at the surgical site


      infection, the virulence of the colonizing or


      contaminating organism, and then the underlying


      sort of resistance of the host.  Those three


      factors are really give rise to the risk of


      surgical site infection.  Next slide.


                If we look at some of the epidemiologic


      factors that have been associated with influencing


      the risk of acquiring a surgical site infection,


      they can be broadly categorized into those that are


      host- or patient-related factors, such as age, body


      mass index, obesity, the presence of diabetes and,


      as we will see later it may not just be a patient


      who is labeled with diabetes but having


      hyperglycemia at the time of surgery, the


      nutritional status of the patient, whether the


      patient has a prolonged preoperative stay, again,




      whether there is infection at a remote site at the


      time of surgery, and whether the patient is on


      immunosuppressive medication such as steroids, or


      whether the patient is a smoker or uses nicotine.


                Some of the procedural factors that have


      been associated with influencing the risk of


      surgical site infection are things like hair


      removal or shaving, the duration of the procedure,


      surgical technique, the presence of foreign bodies


      such as drains, and things like the appropriateness


      or inappropriateness of antimicrobial prophylaxis.


      Next slide.


                What I am going to do now with the next


      series of slides is talk a little bit about some of


      these modifiable factors in terms of things that we


      recommend, or things that are recommended, to be


      done to minimize or moderate the risk of a patient


      acquiring a surgical site infection.  Next slide.


                There are a number of randomized,


      controlled trials showing the benefit of


      perioperative prophylaxis and I won't belabor you


      with those data.  The feeling is that this is




      probably one of the most important things that we


      can do in terms of modifying risk of infection.


      When we talk about antimicrobial prophylaxis we are


      really referring to a brief course, most commonly a


      single dose, of an antimicrobial agent that is


      given just before the operation begins.


      Antimicrobial prophylaxis is not intended as


      therapy.  It really is a preventive strategy ,and


      it really should be used as an adjunctive


      preventive measure and not really used to supplant


      basic things like aseptic technique and some of the


      other basic principles of preventing surgical




                Now, antimicrobial prophylaxis, as I said,


      has been studied in a number of procedures, a


      number of well done randomized, controlled trials


      and it is shown that its use, if done


      appropriately, can decrease the risk of surgical


      site infection at least 5-fold.  Next slide.


                But surgical prophylaxis--again, to show


      you how complex this whole issue is, is not a


      matter of just giving an agent and giving the right




      agent, but also giving it at an appropriate time.


      Now, this slide summarizes a study done by Classen,


      and I think it is one of the more classic studies


      looking at the importance of timing of


      antimicrobial prophylaxis in terms of its efficacy


      in preventing surgical site infection.


                What Classen did was actually study nearly


      3,000 elective clean and contaminated surgery.  He


      looked at the timing of the antibiotic and its


      influence or relationship to the risk of infection.


      If you look at what he called early antimicrobial


      prophylaxis, that is antibiotics given 2-24 hours


      before incision, the rate of infection in that


      cohort was 3.8 percent.  If he looked at


      antibiotics that were given postoperatively, that


      is 3-24 hours after incision, the rate of infection


      was 3.3 percent.  If he looked at antibiotics that


      were given within 3 hours after the incision, the


      rate of infection was 1.4 percent.  Lastly, the


      rate of infection was lower for antimicrobial


      prophylaxis that was given within 2 hours of the


      incision, 0.6 percent.  So, again, it is not just a




      matter of giving prophylaxis and giving the right


      agent, but this issue of timing is critically


      important.  Next slide.


                This next series of slides talks not only


      about this notion of giving antibiotics at a


      critical point before incision, but talks about the


      impact of prolonged surgical prophylaxis.  This is


      a study that was a prospective study that looked at


      a cohort of CABG patients.  They looked at those


      patients who received antibiotic prophylaxis within


      48 hours of the procedure and those for whom the


      prophylaxis was continued for greater than 48 hours


      after the procedure.  Next slide.


                They looked at two outcomes, not only the


      incidence of surgical site infection but also the


      likelihood of acquiring a resistant organism if a


      surgical site infection did occur.  Interestingly,


      what they found is that nearly half of the patients


      received antimicrobial prophylaxis greater than 48


      hours after the procedure.  Again, antimicrobial


      prophylaxis is intended to be given around the time


      of incision to get the maximal sterilization, if




      you will, of the surgical site.  But here we see


      that at least in half the cases patients are


      getting prophylaxis beyond two days after the




                What they found is that the incidence of


      infection in this cohort of patients really was no


      different if antibiotic prophylaxis was


      discontinued within 48 hours or if it was continued


      for greater than 48 hours.  But, interestingly, the


      rate of acquiring a resistant pathogen was 60


      percent higher in those patients who received


      prolonged antimicrobial prophylaxis.  So, again,


      antimicrobial prophylaxis and its influence on SSI


      is not only getting the right agent but getting it


      within the right interval and discontinuing it as


      soon as possible following the surgical procedure.


      Next slide.


                Another area that I think is particularly


      intriguing as to the complexity of things that


      would have to be considered or controlled for in


      looking at SSI risk is this whole issue of glucose


      control and perioperative management of




      hyperglycemia.  This slide actually summarizes a


      prospective study that was done in a group of


      diabetic patients who were undergoing cardiac


      surgery, over nearly a decade at one hospital.


                They had two groups of patients.  Again,


      this is a prospective intervention trial with a


      pre- and post-design.  The control patients were


      those who had received sort of the traditional


      therapy with their glucose being measured and


      monitored intermittently, and being given


      subcutaneous insulin.  What they called the treated


      group were patients who were placed on a continuous


      IV insulin drip for the immediate operative period


      and for up to 48 hours postoperatively.  Next




                The outcomes were that they looked at the


      levels of blood glucose that were below 200 mg/dL,


      and that was sort of the target level, within the


      first two days postoperatively.  The other outcome


      obviously was the incidence of surgical site


      infection, and they focused on deep SSIs.  What


      they found is that in the group who got traditional




      management using subcutaneous insulin on a PRN


      basis the rate of surgical site infection was 2


      percent as compared with the 0.8 percent in those


      patients who were managed with a continuing IV


      drip.  This difference was highly statistically




                Now, there have been some subsequent


      studies that have looked at sort of the prevalence


      of patients who are hyperglycemic who don't carry


      the diagnosis or label of diabetes.  Again, this


      notion of perioperative glucose management probably


      has broader implications beyond just the diabetic


      patient population.  Next slide.


                Another sort of titillating article that


      is summarized here and I think alludes to some of


      the complexity of this issue is this notion of


      perioperative oxygenation, the theory being that


      better oxygenated tissues are less likely to be at


      risk or be prone to developing an infection.


                This was a study that was published in the


      New England Journal in 2000.  It was a randomized,


      controlled, double-blind trial that looked at a




      relatively small group, 500 patients who were


      undergoing colorectal surgery.  Again, I want to


      emphasize that this was colorectal surgery.  The


      intervention was that patients were randomized to


      receive either 30 percent or 80 percent inspired


      oxygen during and for up to 2 hours following the


      surgical procedure.


                Now, what they found is that the incidence


      of surgical site infection was 5.2 percent in those


      who received higher 32 percent versus 11 percent in


      those who received 30 percent oxygen.  That


      difference was statistical significant.


                There has been a more recent study that


      came out in JAMA, and I did not summarize that


      here, looking at a more heterogeneous population of


      patients undergoing intra-abdominal procedures,


      again, randomizing them to receive 70 percent


      oxygen versus 30 percent inspired oxygen.  That


      study concluded that there was not only no


      beneficial effect to a higher level of inspired


      oxygen but, in fact, there might be some


      detrimental consequences.  In fact, they found a




      higher rate of surgical site infections in those


      people who got more oxygen.


                I say this to say again that this


      difference might be in part attributable to the


      population that was studied in terms of procedures.


      So, a lot of these things have to be factored in,


      in terms of trying to extrapolate findings from one


      cohort to another--not only what the intervention


      was but the population and the procedure that was


      studied.  Next slide.


                What about the issue of antisepsis and


      antiseptics?  Probably, as you have heard from Dr.


      Boyce, a lot of the studies around the efficacy and


      the benefits of antiseptics really use bacterial


      count on scans and the amount of cutaneous flora


      remaining after their use as the primary outcome


      measure.  When we look at hard outcomes or harder


      outcomes in terms of patient outcomes, data becomes


      much thinner.


                These are just summarizing some data, and


      these are admittedly older studies and, you know,


      these studies to be done today are much more




      difficult for a variety of reasons, but these three


      studies summarize data looking at surgical site


      infection rate with patients receiving preoperative


      showers versus those not getting showers.  The


      earliest study was in the '70's where the rate


      among those who did not get showers was 2.3 percent


      versus 1.3 percent.  In the subsequent two studies,


      in the 1980's, the actually the difference was


      quite closer.


                Again, I think some of these studies,


      although they did not show a statistically


      significant difference, may be confounded by


      failure or inability to control for a lot of the


      factors that we mentioned up to this point.  But,


      also, I am not convinced that these studies were