+ + + + +




+ + + + +




+ + + + +




+ + + + +


            The meeting was held in Salons A, B, and C of the Hilton Washington, D.C. North, 620 Perry Parkway, Gaithersburg, Maryland, at 8:34 a.m., Dr. Sanjiv H. Naidu, Acting Panel Chairperson, presiding.

















PRESENT (Continued):










                  C O N T E N T S


Appointment to Temporary Voting Status ......... 5

Conflict of Interest Statement ................. 6

Introductions ................................. 11

Post Market Study Design, Susan Gardner, Ph.D. 13

Division Update, Glenn Stiegman ............... 19

Open Public Comment:

      Susan Krasny, Ph.D., R.A.C. ............. 26

      William Maloney, III, M.D. .............. 32

Birmingham Hip Resurfacing System, P040033

Sponsor Presentation, Marcos Valez-Duran .. 46, 66

      Derek McMinn, FRCS ...................... 47

      Tim Band ................................ 61

      George DeMuth ........................... 77

      Neal Defibaugh .......................... 86

FDA Presentation, John Goode, M.S. ........... 117

      Chang S. Lao, Ph.D. .................... 151

Panel Deliberations and FDA Questions......... 168

Open Public Comment:

      Craig Thomas, M.D. ..................... 290

               P R O C E E D I N G S

                                       (8:34 a.m.)

            MS. SCUDIERO:  Good morning.  We're ready to being this meeting of the Orthopedic and Rehabilitation Devices Panel.

            I am Jan Scudiero, the Executive Secretary of this panel and a reviewer in the Division of General Restorative and Neurological Devices. 

            First, the usual housekeeping matters.  If you haven't already done so, please sign the attendance sheets that are on the tables by the door.

            Information on today's agenda and for panel meeting minutes and transcripts as well.

            The next tentatively scheduled meeting for this panel on November 3rd and 4th is canceled because there's no agenda item ready for panel review.

            Upcoming panel meetings are announced on an advisory panel Website, the Federal Register and in the telephone information line.  Please monitor the Website for future meeting announcements.

            Finally, as a courtesy to others in the room, please, turn off or put on silence your cell phones during the meeting.

            Thank you.

            Dr. John Kirkpatrick is unable to be with us today.

            I will now read into the record three agency statements prepared for this meeting.  They are the appointment of temporary panel chair statement, the appointment of temporary voting member statement, and the conflict of interest statement.

            First, I appoint Sanjiv H. Naidu, M.D., Ph.D., a voting member of the Orthopedic and Rehabilitation Devices Panel as Acting Panel Chair for the September 8th and 9th, 2005 meeting of the panel, and this was signed by Dr. Daniel G. Schultz, Director, Center for Devices and Radiological Health, on September 7th.

            The appointment to temporary voting status:  pursuant to the authority granted under the Medical Devices Advisory Committee Charter dated October 27th, 1990, and amended April 20th, 1995, I appoint the following as voting members of the Orthopedic and Rehabilitation Devices Panel for the duration of this meeting on September 8th, 2005:  Brent A. Blumenstein, Ph.D., Jay D. Mabrey, M.D., Michael B. Mayor, M.D., and Harry B. Skinner, M.D. Ph.D.

            For the record, these people are special government employees and are consultants to this panel or another panel under the Medical Devices Advisory Committee.  They have undergone the customary conflict of interest review and have reviewed the material to be considered at this meeting.

            The conflict of interest statement:  the Food and Drug Administration is convening today's meeting of the Orthopedic and Rehabilitation Device Panel of the Medical Devices Advisory Committee, under the authority of the Federal Advisory Committee Act, FACA, of 1972

            The panel meetings provides transparency into the agency's deliberative processes.  With the exception of the industry rep. all members of the committee are special government employees, or SGEs, or regular federal employees from other agencies and are subject to federal conflict of interest laws and regulations.

            Consequently, in the interest of transparency and in the spirit of disclosure, the following information on the status of this panel's compliance with federal ethics and conflict of interest laws as covered by, but not limited to those found at Part 18, U.S. Code 208, and Part 21, U.S. Code, Section 355(n)(4).  This is being provided to the participants in today's meeting and to the public.

            FDA has determined that the members of this panel are in compliance with the federal ethics and conflict of interest laws, including, but not limited to, Part 18 U.S. Code Section 208, and Part 21, U.S. Code Section 355(n)(4).  Under Part 18 U.S. Code Section 208, applicable to all government agencies and Part 21 U.S. Code Section 355(n)(4), applicable to FDA, Congress has authorized FDA to grant waivers to special government employees who have limited financial conflict when it is determined that the agencies need for the particular individual's services outweighs his or her potential financial conflict of interest.

            Members who are special government employees at today's meeting, including special government employees appointed as temporary voting members, have been screened for potential financial conflicts of interest of their own, as well as those imputed to them, including those of their employer, spouse, or minor child, related to the discussions of today's meeting.  These interests may include investments, consulting, expert witness testimony, contracts, grants, CRADAs, teaching or speaking and writing, patents and royalties, and primary employment.

            Today's agenda involves the review of a pre-market approval application, PMA, for a hip resurfacing system in tended to relieve hip pain and improve hip function in patients who have adequate bone stock and are at risk of requiring more than one hip joint replacement over their lifetimes.

            This is a particular matters meeting during which specific matters related to the PMA will be discussed.

            In accordance with Part 18, U.S. Code Section 208(b)(3), a full waiver has been granted to Dr. Michael Mayor.  Dr. Mayor's waiver involves a patent licensed to assist a company of a competitor.  He receives less than $15,001 in royalties for the patent.

            Copies of each acknowledgement and consent to disclosure statement signed by each participant at today's meeting who received a conflict of interest  waiver along with the statement will be available for review at the registration table during the meeting and will be includes as part of official meeting transcript.

            A copy of the written conflict of interest statement may be obtained by submitting a written request to the agency's Freedom of Information Office, Room 12A-30 of the Parklawn building.

            Lastly, Ms. Pamela Adams is the industry rep. acting on behalf of all related industry and is employed by Etex Corporation, Incorporated.  In the event that the discussions involves any other products or firms not already on the agenda for which an FDA participant may have a financial interest, all meeting participants are reminded that they're required by Part 18, U.S. Code 208, to exclude themselves from such deliberations and announce their exclusion for the record.

            Finally, in the interest of public transparency with respect to all other participants, we ask that they publicly disclose prior to making any statements any current or previous financial involvement with any firm whose product they may wish to comment upon.

            Thank you.

            I would now like to turn the meeting over to Dr. Naidu, our Acting Chairman for the day.

            PANEL CHAIRPERSON NAIDU:  Good morning. My name is Sanjiv Naidu.  I am the Acting Chairperson for the Orthopedic and Rehab. Panel today.  I'm Professor of Orthopedic Surgery at Penn State College of Medicine, and I'm also a materials scientist.

            At this meeting the panel will be making a recommendation to the FDA on the approvability of the pre-market application, P040033, for the Smith & Nephew Birmingham resurfacing hip system, the BHS system.  It is a hip joint, metal on metal, semi-constrained, hybrid prosthesis, cemented femoral component, uncemented acetabular component.  It is intended to relieve hip pain and improve hip function in patients who have adequate bone stock and are at risk of requiring more than one hip joint replacement over their lifetimes.

            Before we begin, I would like to ask our distinguished panel members who are generously giving their time to help FDA in the matter being discussed today and the other FDA staff seated at this table to introduce themselves.  Please state your name, your area of expertise, your position, and affiliation.

            Mr. Melkerson, if you could start off.

            MR. MELKERSON:  I'm Mark Melkerson.  I'm the Acting Director for the Division of General Restorative and Neurological Devices.

            DR. MAYOR:  Michael Mayor, orthopedic surgeon, Professor of Orthopedics at the Dartmouth-Hitchcock Medical Center and the Dartmouth Medical School.  I'm co-director of the Thayer Engineering School, Dartmouth Biomedical Engineering Laboratories.

            DR. BLUMENSTEIN:  Brent Blumenstein, a biostatistician working out of Seattle.

            DR. MABREY:  Jay Mabrey, orthopedic surgeon, Medical Director  of the Orthopedic Motion and Sports Performance Laboratory at Baylor University Medical Center and also Chief of Orthopedics at Baylor University Medical Center in Dallas.

            DR. KIM:  I'm Choll Kim.  I'm the Assistant Professor of Orthopedic Surgery at the University of California, San Diego.  I'm the Director of the Spine Research Lab there.

            DR. SKINNER:  My name is Harry Skinner.  I'm Professor and  Chair of Orthopedic Surgery at the University of California, Irvine.

            MS. WHITTINGTON:  My name is Connie Whittington.  I'm an orthopedic clinical nurse specialist at Piedmont Hospital where I serve as the Coordinator for Orthopedic Research.

            MS. ADAMS:  I'm Pamela Adams.  I'm Chief Operating Officer at Etex Corporation.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            I would like to note that for the record that the voting members present constitute a quorum as required by 21 CFR Part 14.

            There will be two brief presentations before the main agenda topic.  First is Dr. Susan Gardner who will speak on post market study design.

            DR. GARDNER:  Okay.  Good morning.  I'm going to spend just a few minutes telling you about an important programmatic change that has taken place in the center.

            The essence of the change is a move of the condition of approval studies program from the Office of Device Evaluation to the Office of Surveillance and Biometrics.

            Briefly, the Office of Surveillance and Biometrics is involved both in premarket and post market activities.  We're involve din the premarket review because the statisticians and the epidemiologists of the center are in the Office of Surveillance and Biometrics, and we also have a major role in the post market in that all the adverse events and the post market monitoring tools, the medical device reporting system, and the MedSun program are in OSB.

            We're also responsible for analyzing the data to come in on these post market monitoring tools and characterizing the risk and working with the rest of the center to identify post market problems and take action on those.  We're responsible for coordinating the center response to health care professionals in risk communication, and we're responsible for interpretation of the medical device reporting regulation.

            The legislative 21 CFR 814.82 says that post approval requirements can include continuing evaluation and periodic reporting on the safety, effectiveness, and reliability of the device for its intended use, and this is the basis of the condition of approval studies program.

            The impetus for the change came from a study, an internal evaluation that we did in CDRH of our CoA Program.  We look at this in about 2000, 2001, and we went back and we looked at all of the PMAs that had been approved from 1998 through the year 2000.  There were 127 PMAs, and 45 of those had condition of approval orders.

            And what we found unfortunately is that we were really unable to find some of these studies, and we realized this was because we had no standardized tracking procedures for tracking the results of these studies.  As one might expect there had been a turnover in lead reviewers as people naturally moved through the organization or changed jobs, and also that in ODE with their focus on premarket, it was really difficult for them under their current resources to continue to follow these studies and give them the attention that they need.

            So we came up with a plan to change the program.  The point of the change or the goal of the change was to make sure that we could obtain this post market information at this critical period when the device enters the market and we could continue to assess the safety and effectiveness as the device moves from the clinical trial into the real world use, and this obviously would allow us to better characterize the risk-benefit profile and add to our ability to make sound scientific decisions as we continue to monitor these devices.

            We actually did a pilot that went for about two years.  So when we officially changed the program on January 1st of 2005, we were fairly well prepared, and in addition to that a number of studies already were being monitored in OSB.  We have developed and instituted an automatic tracking system that is up and working, and not only do we have studies being tracked from January 2005.  We've gone back to pick up studies that have been approved earlier.  So we're also following those.

            The point of the tracking system obviously is to acknowledge to industry when studies are received and to follow up if we don't receive the information that we're supposed to have.

            Another fundamental change is we have added an epidemiologist to the PMA review team when we think that we're going to have or it looks likely that we'll have a condition of approval study if the device is approved.  The epidemiologist is tasked with the development of the post market monitoring plan during the premarket review working with the rest of the PMA team.  The epidemiologist has the lead for developing, and let me emphasize we're really looking to develop well formulated post market questions and make sure that these studies are important if we're going to ask industry to do them.

            The epi person will have the lead in the design of the study protocol and in the evaluation of study progress and results after approval, and so when the results of these studies come into OSB, we will then look at the results and then turn to our colleagues in the premarket arena and go back to the PMA team, review what the progress has been and see if there's anything else that we need to do.

            So with these changes, why do we think we will do better?  Well, first of all, we think that with a real emphasis on working with industry to make sure that the questions that we're asking in the premarket arena are really important fundamental questions, and we have a good study protocol design.  Everybody will be motivated, first of all, to get the studies done, and secondly, to evaluate them and make sure that the results are important.

            For everybody acknowledgement of the work that you do and feedback on the studies motivate people to do it.  So, again, industry won't feel that their work is falling into a black hole, and we will be able to interact with them to make sure that the results are on target.

            We will be posting the status of the studies on CDRH Website, and also if studies are not done, we do have the ability under Section 522 to mandate a post market study, but again, we are hoping if people work together to have a good study and a good protocol that we won't have to go there.

            What does this mean to the Advisory Panel?  Well, first of all, let me emphasize it certainly doesn't change the standards for making sure that the product is reasonably safe and effective before it's approved. 

            However, during the approval process we will sometimes lay out deliberately and sometimes, of course, it will come up naturally in discussion questions that deal with the post market piece of the approval process, and your input, your suggestions of possible approaches for the post market pieces or questions that you're concerned about will be really important to us and certainly will be taken under consideration if the device is approved and we decide to do a condition of approval study.

            We are also committed to coming back to you, either FDA or industry, to come back to you and give you the results of condition of approval studies after the product is approved.

            Any questions?

            (No response.)

            DR. GARDNER:  Thanks.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Gardner.

            Now, Mr. Glenn Stiegman will give us a division update on the activities since 2004 panel meeting.

            MR. STIEGMAN:  Hi.  My name is Glenn Stiegman.  This is the panel update for the Orthopedic and Rehabilitation Panel of September 8th and 9th.

            Action on items from previous Advisory Committee meetings.  From the June 2nd, 2004 meeting, P040006 from the DePuy Spine, the Charite artificial disc.  This device was approved October 26th of 2004 for spinal arthroplasty in patients with degenerative disk disease at one level from L4 to S1.

            A post approval study to further document the incidence of complications, such as migration and subsidence is ongoing.

            Other significant approvals that have occurred since the last panel update, two ceramic on ceramic hips, one on December 17th, 2004, the Smith & Nephew Reflections ceramic acetabular system, which is indicated for patient requiring primary total hip arthroplasty due to noninflammatory arthritis, such as osteoarthritis, avascular necrosis and traumatic arthritis.

            On May 4, 2005, the DePuy Orthopedics, the Dura Option ceramic hip system for patients with noninflammatory degenerative joint disease, such as osteoarthritis, avascular necrosis, congenital hip dysplasia, and post traumatic arthritis.

            On December 3rd, 2004, P010029, Savient Pharmaceuticals, which has recently been bought out by Ferring Pharmaceuticals, the Nuflexxa one percent sodium hyaluronate for treatment of pain and osteoarthritis of the knee in patients who have failed to respond adequately to conservative nonpharmacologic therapy and simple analgesics.  This particular device is made by bacterial fermentation instead of being extracted from chicken products like other hyaluronate products.

            An HDE, the humanitarian device exemption, H030009 from Synthes USA, the vertical expandable prosthetic titanium rib, which is indicated for the treatment of thoracic insufficiency syndrome.  It's going to immature patients.  TIS is defined as inability of the thorax to support normal respiration or lung growth.

            Other significant 510(k) clearances, February 18th, 2005, Blackstone Medical, we cleared a laminoplasty fixation system for holding bone graft in place during laminoplasty procedures.

            From the Zimmer Trabecular Technology, the trabecular metal osteonecrosis interventional implant which is indicated to treat patients with osteonecrosis of the femoral head.

            Guidances that have been published since the last panel update, the clinical data presentations for orthopedic device applications and the clinical trial considerations, vertebral augmentation devices to treat spinal insufficiency fractures.

            The Division of General Restorative Neurological Devices staffing changes.  The new or permanent orthopedic staff in ORDB, Ronald Jean, Dr. Kristin Mills, John Holden, and Dr. Khan Li have all joined our staff since the last panel update on a permanent or new basis.

            Also, no longer with the Orthopedic Devices Branch or FDA, Barbara Zimmerman is now the Deputy Director of the Division of Cardiovascular Devices.  Dr. Celia Witten is now the Director of the Office of Cellular Tissue and Gene Therapies at CBER.  Dr. Martin Yahiro is moving on to industry on September 17th.  Genevieve Hill is returning to school and Dr. Michael Schlosser is returning to private practice.

            Current division staff changes, Mark Melkerson is the Acting Division Director on a four-month detail.  Ted Stevens and Barb Buch are both Acting Deputy Division Directors for DGR&D.  Myself, Glenn Stiegman, is the new Branch Chief of Orthopedic Devices Branch, and Aric Kaiser is currently on a four-month detail on the Office of Surveillance and Biometrics.

            And lastly, in conclusion, we'd like to take a moment of silence to remember one of our bright, young, highly esteemed orthopedic device reviewers, Jonathan Lim who passed away last week.  More than just a colleague, Jon was a good friend to many and will be sorely missed.  Jon's spirit will always live in the pleasant memories that he created in our hearts.

            (Pause in proceedings.)

            MR. STIEGMAN:  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Stiegman.

            We will now proceed with the open public hearing portion of the meeting.  Prior to the meeting, two people asked to speak in the open public hearing.  They will speak in order of their request to speak.

            We ask that you speak clearly into the microphone as the transcriptionist is dependent on this means of providing an accurate record of this meeting.

            Please state your name and the nature of any financial interest you may have in this or any other medical device company.  Ms. Scudiero will now read the open public hearing statement.

            MS. SCUDIERO:  Both the FDA and the public believe in a transparent process for information gathering and decision making.  To insure such transparency at the open public hearing session of the Advisory Committee meeting, FDA believes it is important to understand the context of any individual's presentation.

            For this reason, FDA encourages you, the open public hearing or industry speaker, at the beginning of your written or oral statement to advise the committee of any financial relationship you may have with the sponsor, its products and, if known, its direct competitors.

            For example, this information may include the sponsor's payment of your travel, lodging or other expenses in connection with your attendance at the meeting.

            Likewise, FDA encourages you at the beginning of your statement to advise the committee if you do not have any such financial relationships.  If you choose not to address the issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

            I would like to note for the record that prior to the meeting three letters with general comments on the agenda item were received.  Copies of these letters were given to the panel and the PMA sponsor this morning.  These are now part of the record for this meeting.

            Dr. Naidu.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            The first open public hearing presented is Dr. Susan Krasny, Vice President of the Orthopedic Surgical Manufacturers Association.

            Dr. Krasny, you have ten minutes.

            DR. KRASNY:  Thank you.

            I am the Senior Director of Regulatory Affairs and Clinical Affairs with Stryker Spine, and I have no financial interest in the sponsor today.

            Good morning.  My name is Susan Krasny, and I speak here today representing the Orthopedic Surgical Manufacturers Association, OSMA, of which I am Vice President.

            OSMA, a trade association of over 30 member companies, welcomes this opportunity to provide general comments at today's Orthopedic Advisory Panel meeting.  OSMA's comments should not be taken as an endorsement of the products being discussed today.  We ask instead that our comments be considered during today's panel deliberations.  These comments represent the careful compilation of member companies' views.

            OSMA was formed over 40 years ago and has worked cooperatively with the FDA, the American Academy of Orthopedic Surgeons, the American Society for Testing Materials and other professional medical societies and standard development bodies.  This collaboration has helped to insure that orthopedic medical products are safe, of uniform high quality, and supplied in quantities sufficient to meet national needs.

            Association membership currently includes over 30 companies who produce over 85 percent of all orthopedic implants intended for clinical use in the United States.

            OSMA has a strong and vested interest in insuring the ongoing availability of safe and effective medical devices.  The deliberations of the panel today and the panel's recommendation to the FDA will have a direct bearing on the availability of new products.

            We make these comments to remind the panel of the regulatory burden that must be met today.  We urge the panel to focus its deliberations on the product safety and effectiveness based on the data provided.

            The FDA is responsible for protecting the American public from drugs, devices, food and cosmetics that are either adulterated or unsafe or ineffective.  However, FDA has another role, that to foster innovation.

            The Orthopedic Devices Branch is fortunate to have available a staff of qualified reviewers, including Board certified orthopedic surgeon to evaluate the types of applications brought before this panel.  The role of this panel is also very important to the analysis of the data in the manufacturer's application and to determine the availability of new and innovative products in the U.S. marketplace.

            Those of you on the panel have been selected based on your expertise and training.  You also bring the view of practicing clinicians who treat patients with commercially available products.

            OSMA is aware that you have received training from the FDA on the law and regulation, and we do not intend to repeat that information today.  We do, however, want to emphasize two points that may have a bearing on today's deliberations:

            One, reasonable assurance of safety and effectiveness, and

            Two, valid scientific evidence.

            There is a reasonable assurance that a device is safe when it can be determined that the probable benefits outweigh the probable risks.  Some important caveats associated with this oversimplified statement include valid scientific evidence and proper labeling, and that safety data may be generated in the laboratory, in animals and in humans.

            There is reasonable assurance that a device is effective when it provides a clinically significant result.  Again, labeling and valid scientific evidence play important roles in this determination.

            The regulation and the law clearly state that the standard to be met is reasonable assurance of safety and effectiveness.  "Reasonable" is defined as moderate, fair, and inexpensive.

            The regulation states that well controlled investigations shall be the principal means to generate the data used in the effectiveness determination.

            The following principles are cited in the regulation as being recognized by the scientific community as essential in a well controlled investigation.

            One, a study protocol;

            Two, method of selecting patients;

            Method of observation and recording results;

            And, four, comparison of results with control.

            The panel has an important job today.  You must listen to the data presented by the sponsor, evaluate the FDA presentations, and make a recommendation about the approvability of the sponsor's application.

            We speak for many applicants when we ask for your careful consideration.  Please keep in mind that the standard is reasonable assurance, balancing the benefits with the risks.  The regulatory standard is now proof beyond a shadow of a doubt.

            When considering making recommendations for further studies, remember that the FDA takes these recommendations seriously, often as a consensus of the panel as a whole and they may delay the introduction of a useful product or result in burdensome and expensive additional data collection.

            Therefore, you play an important role in reducing the burden of bringing new products that you and your colleagues use in treating patients to the market.  Please be thoughtful in weighing the evidence.  Remember that the standard is a reasonable assurance of safety and effectiveness and that there is a legally broad range of valid scientific evidence to support the determination.

            OSMA thanks the FDA and the panel for the opportunity to speak today.  Our association trusts that its comments are taken in the spirit offered, to help the FDA decide whether to make a new product available for use in the U.S. marketplace.

            OSMA are present in the audience and are available to answer to answer questions any time during the deliberations today.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Krasny.

            The next presenter is Dr. William Maloney, III. 

            Dr. Maloney.

            DR. MALONEY:  Thank you, and good morning.

            My name is Bill Maloney, and I'm the professor and Chairman of Orthopedics at Stanford University School of medicine.

            And while we're getting up my presentation, I'll do my conflict disclosure.  I design hip implants for Zimmer, for which I receive royalties.  I am not involved in any surface replacement design.  I design knee implants for Wright Medical for which I receive royalties.  I do not own stock in any orthopedic implant companies.  My travel expenses have been reimbursed by Wright Medical, and currently my time is not being compensated to come here.

            So I'm here to make some comments on this submission, and for those of you on the panel who do not know me, in addition to being professor and Chairman of Stanford University, I'm currently Chief of Joint Replacement Service at Stanford.

            Prior to this I was the Chief of the Joint Replacement Service at Washington University School of Medicine.  I'm a member of the North American Hip Society and International Hip Society.  I led the American Joint Replacement Registry effort in trying to get that off the ground in this country and chaired that committee for two years, and I'm currently a member of the Quality Improvement Program for Medicare working with Dr. David Hunt.

            I want to say quite clearly what I'm not here to do.  I'm not here to indict Smith Nephew.  It's a well established orthopedic company, a great reputation.  I'm not here to indict Total Resurfacing Arthroplasty or the specific implant, and I'm certainly not here to indict Derek McMinn, who is a colleague of mine and certainly a well known arthroplastic surgeon.

            What I am here to do is to make some comments on the study methodology which I think appears to me as a problem, to talk a little bit about conflict of interest.  Obviously that's an important thing in this country and one as a chairman of an academic department at a major medical school I deal with on a regular basis, and make a comment on compelling medical need.

            When you look at studies, there are a variety of studies in the orthopedic literature.  Most of them are retrospective.  There's prospective data collection with retrospective data review, like registry as the Mayo Clinic Registry would be a good example of that.

            There's a prospective study.  In a prospective study you generate a hypothesis.  You design the study to test the hypothesis.  You get IRB approval.  You recruit patients, and you consent the patients.  You collect the data, and you analyze the data.

            Then you have an FDID study, which is a prospective study that goes beyond those requirements.  There's site monitoring.  There's concurring data verification against source documentation.  There's adverse event documentation and reporting, and there's FDA inspections, akin to an IRS audit.

            When you look at the norms for an IDE study protocol, these have been well established.  The protocol is predetermined.  The safety and efficacy endpoints are well defined.  The control group is homogeneous with the treatment group.  There's well established inclusion and exclusion criteria.  There's predefined follow-up intervals and direct patient evaluation.  The conclusions are based on a predetermined, valid statistical plan, and there's substantial patient site monitoring which is mandatory.

            These trials are usually multi-center under a common protocol.  There's contemporaneous accountability for all adverse events and complete documentation of all protocol deviations, regardless of severity.

            The data that we're currently or you're currently reviewing in this current PMA submission is a combination of retrospective data and data collected prospectively and reviewed retrospectively. In that submission it states there were not predefined follow-up time windows, standardized clinical evaluations, adverse event report forms or standardized radiographic evaluations.

            So clearly, that doesn't meet the standard that we're used to in the evaluation of Class 3 devices.

            When you look at critical study elements, the survivorship data is from a single surgeon.  It's from Derek McMinn's data, and the justification of that is as follows.  The Oswestry Outcomes Center records information regarding complications, deaths or revisions, but the data is not verifiable.  You cannot verify the primary source data.  Therefore, the only verifiable data is Dr. McMinn's data.

            They further go on to state, "Comparison of data from the Oswestry Outcomes Center database to the data from Mr. McMinn's clinical records provides additional assurance that an accurate, up-to-date survivorship data is known."

            But the problem here is "additional."  The only data we have here on survivorship that can be verified is Dr. McMinn's data, and it's not done concurrently and not by an independent study monitor.

            The single surgeon series, as we all know, has significant disadvantages.  Derek is an expert surgeon.  He's been doing this operation since 1989.  He is the pioneer in this operation and deserves a lot of credit, but the question is:  were his results be optimal to the surgeon community at the U.S. at large?

            Clearly, I don't think that's the case.  This does not represent U.S. surgeon data or U.S. patients, and this is a new operation in the United States.  We have no experience with it, and surgical training is going to be critical.

            When we look at the specific cohorts in this study, there are three.  There's an X-ray cohort.  There's the Oswestry cohort, and there's the McMinn cohort.

            The Oswestry Outcome Center, which it's unclear how that's funded, is a self-administered patient questionnaire.  There's no direct patient contact, which is, of course, required in ID studies.  They report an outcome measure which we're not familiar with, an OSHIP score which appears to be a combination of a Harris HIP score and a Merle D'Aubigne score, and presumably  there's patient consent in this study, although it's not well documented.

            In the McMinn cohort, it appears to have no functional outcome data.  It is specifically just survivor data, as the outcome data was not available for review of this submission, and it doesn't appear that those patients in the McMinn cohort were actually consented for a research study.

            They certainly were consented for surgery.  The PMA submission clearly states that they were given options as it relates to hip arthroplasty, but that's fundamentally different than being consented for a research study.

            What about the radiographic analysis?  This, again, appears to be primarily a retrospective radiographic analysis.  Clearly at the beginning of this PMA station they stated that there's no radiographic protocol. 

            If you look at the numbers in this cohort, there's 124 hips.   Two hips were excluded because of patient death in one patient.  Four hips were revised.  So they left with 118.  Ten of those we then lost to follow-up.  So you're down to 108, and then 19 hips had no immediately post-op X-ray.  So there's 89 hips which were potentially valuable.

            It goes on to state there were immediate postop films on 89 of 108 procedures with five irradiographs, but the sponsor stated that these films were of low quality, portable films and unusable for the purposes of precise postoperative comparisons.

            Therefore, baseline films for the purpose of comparisons were made in each of the 108 cases in the postoperative time period, usually within three months, but eight of the 108 procedures had baseline evaluations performed at the time points ranging from 110 to 860 days.

            So in some cases the baseline X-rays were quite a bit of time after the deemed X procedure.  In a prospective ID study, these would all be protocol violations and could be potentially excluded.  So you could actually end up with a cohort in the X-ray study of zero if you had a strict analysis of the data.

            The next topic I want to briefly discuss, and it has already been a significant topic here this morning is one of conflict of interest.  In this country, we are under great scrutiny as it relates to our conflict, and as a department chairman, again, I deal with this on a regular basis.

            Conflict of interest exists when an investigator and/or his or her family has the potential for financial gain.  Financial conflicts in human subject research at least at our institution are of special concern, and I'm sure they are across this country.

            How, Derek is the product designer.  He's the pioneer here, and he's one of the co-founders of Midland Medical Technology, who we do not have disclosure in the PMA submission as to what the exact conflict is here.  This company was purchased by Smith and Nephew for 67 million pounds, and that's a fair sum of money in anybody hands, and it will pay another 33 million pounds for the Food and Drug Administration or if the Food and Drug Administration allows this procedure into the United States.

            The company's history, this has been an extremely successful company, and I wish I was involved with it.  It was incorporated in 1996, and in 1997 they issued about 100,000 shares, 60,000 of which were immediately transferred to Sky Fall, Limited and 30,000 to Maldeney, Limited.  Who owns those we have no idea, and who has gotten the 66 million pounds needs to be addressed because we have to do that in this country, and the playing field needs to be level.

            Now, Derek has got a five-year contract based on the Daily Mail with his new employers.  He told the Daily Mail, "I will stay on until I retire and drop.  I am completely addicted to this product."

            Now, if that's the case, if he's an employer of the sponsor, and again, we need to have clarification there, it appears that the data in this PMA is entirely from the clinical practice of a Smith & Nephew employee.

            In this country and for the FDA for the ID studies, you write "primary" on the individual's IRB to oversee issues of patient protection.  This includes the study consent form and disclosure of financial conflict of interest. 

            At my institution I would not be able to do an IDE study that involved a Zimmer product.  I would not be able to do a clinical study that involved a Zimmer hip product.  It's simply not mitigatable.

            The academic community in this country has weighed in on this issue.  The NIH has weighed in on this issue, and the federal government has weighed in on this issue.  Significant financial conflicts cannot be mitigated.  It doesn't mean that the data is bad, and it doesn't mean that the investigator is dishonest.  It just means that the appearance of significant financial conflicts currently are not felt to be mitigatable.

            The last point relates to compelling medical need.  Now, some people have said, well, these products are great we need to have them on the market, but clearly with resurfacing arthroplasty that's not the case.  Conventional hip replacement is a good operation.  It has an extremely long track record, and currently in this country there are three IDEs being performed and two PMAs that are pending before the FDA.

            Corin and DePuy and Wright Medical all have IDE studies in this country.  So clearly, we're going to have a little bit higher quality data to review in the relatively near future.

            So how do I summarize my evaluation and those of us who are involved in the American Joint Replacement Registry? 

            This is a PMA device.  It's the first of its kind in the United States.  It's supported by data, and this is a quote from the PMA submission, from essentially one surgeon, Derek McMinn, who is a very qualified surgeon and very expert surgeon.

            That source of data is also the product designer and the founder of the company and the things they  implant.  IDE conventions are not followed.  The data set does not reflect the high bar set by the FDA for approval of Class 3 orthopedic devices, and I remind you that the last time this type of data was accepted at least that I could find was back when the Mittelmeier hip was approved.  This was a ceramic-on-ceramic total hip replacement that had excellent clinical results in Europe, principally by Dr. Mittelmeier.  There was no U.S. IDE.  It was a clinical disaster in this country, and it doomed ceramic-to-ceramic hip replacements in this country for years.

            This would be precedent setting in terms of how all Class 3 devices are evaluated for approval in this country, and I'll tell you I was involved, and I was in the orthopedic implant side.  If this goes forward, this would be the last IDE done in this country.  There's no reason to do an IDE if retrospective data from another country is acceptable.

            And maybe that's the way we should go, but clearly, it's a significant change in our current standards.  Maybe the bar is too high, and maybe it is too burdensome, and it probably is, and I probably need to have some changes made, but this is a big jump from what we're currently used to, and it potentially is a significant patient safety issue.

            Thank you for your attention.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Maloney.

            Is there anyone else in the room who would like to address the panel at this point?  If so, please raise your hand and come forward and state your name, affiliation and whether you have any involvement in a medical device firm.

            (No response.)

            PANEL CHAIRPERSON NAIDU:  I don't see anyone.  Please note that there will be a second open public session in the afternoon.  If anyone else would like to address the panel about today's agenda topic, you may speak in the afternoon.

            We will now proceed to the sponsor presentation for the Birmingham hip resurfacing device.  We will then have a short break and proceed with the FDA presentation.

            After lunch the panel will deliberate on the approvability of the PMA.  Before the panel votes on the approvability of the PMA, there will be a second open public hearing and FDA and sponsor summations.

            I would like to remind public observers at this meeting that while this meeting is open for public observation, public attendees may not participate except at the specific request of the panel.

            We will begin with the sponsored presentation.  The first Smith & Nephew presenter is Mr. Marcos Velez-Duran, Vice President of Clinical and Regulatory Affairs and quality assurance.  He will introduce the other Smith and Nephew presenters.

            Mr. Duran.

            MR. VALEZ-DURAN:  Yes.  Good morning.  My name is Marcos Valez-Duran, and I am the VP of Regulatory and Clinical Affairs and Quality for Smith & Nephew.  I'm here today to present to you a summary of the data presented in the pre-market approval for the Birmingham hip resurfacing product.

            I will take this opportunity to thank the panel and FDA for the time and effort that they have put in into the review of this PMA.  In particular, I would like to thank FDA for embracing the spirit of the regulation.  As demonstrated by the then interactively review in this PMA and concerning the data that will be presented today as meeting the requirement of valid scientific evidence as outlined under the law.

            The presenters for this morning as me, Marcos Valez; Mr. Derek McMinn will present design history; Mr. Tim Band will present on preclinical information.  I will come back and present the summary of the clinical data.  George DeMuth will present on the statistical issues associate with this PMA and data analysis, and Neal Defibaugh will review the labeling, plus approval studies and plan for training.

            In addition, we have a number of people supporting our efforts today that will get up and answer questions as necessary, that is, Dr. Cecil Rorabeck, Professor James Richard from the Oswestry Center, Mr. Joseph Daniels, Professor Anthony Unsworth, Dr. Roger Rogerson, Marie Marlow, Dr. Marc Thomas, and Sally Maher.

            Now I would like to introduce Mr. McMinn to give the next presentation.

            DR. McMINN:  Thank you, Marcos.

            Good morning, ladies and gentlemen.  I'm Derek McMinn.  I'm an orthopedic surgeon from Birmingham, England, and I'm the co-inventor of the Birmingham hip resurfacing.  I do have a financial interest in the Birmingham hip resurfacing.  I'm a consultant for Smith & Nephew, and I'm a non-Executive Director of Smith & Nephew.

            For the nonclinicians, I just want to do a very quick review of the indications for hip arthroplasty.  On the left there you see a normal hip, and on the right you can see an osteoarthritic hip with loss of some sippier (phonetic) cartilage.


            When you look at an arthritic femoral head, there is the retained cartilage, and there is the variable on the femoral head.  So the difference between a pain free hip and a very painful hip with a disabled patient is the loss of a few millimeters of articular cartilage.

            Now, for that lesion, we take a really aggressive approach to it.  We resect the femoral head and neck, insert a cup in the stem to the shaft of the femur called a total hip replacement.  Why do we do that?   Because we've all been trained to do it and because for the vast majority of patients it works extremely well.


            Here's a patient of mine, a total hip replacement freshly done on the left.  A few years later unfortunately the patient has developed loosening of the femoral component with substitutes.  It had to be revised.


            This is another patient of mine with two hip replacements and both sides have failed for different reasons.  On this side, a typical total hip replacement.  Where is the polyethylene, and that is caused a linear pattern of osteolysis and loosening of the soffit.

            On the other side we've got a hybrid total hip replacement and an uncemented cup, and again, there's been wear of the polyethylene, and this patient has nasty pelvic osteolysis, and on the femur side it's removed most of Zone 7 in the femur.

            The problem that we all face now is largely that set-up with pelvic osteolysis thanks to wear of polyethylene, and the problem is made worse the younger and more active the patient is.


            So to summarize the problems with total hip replacement, there's an incidence of dislocation.  In our country it's three to four percent.  There's a leg lengthening issue with total hip replacement, which is a common cause for patient unhappiness.  There's acetabular and femoral loosening which is reducing with time, but particularly acetabular and femoral osteolysis is a big problem.  There's the inevitable stress shielding of inserting a stem into the shaft of the femur and a revision surgery, as we all know, is difficult and expensive and in young patients may be recurrent. 

            So why can't we just replace the worn out acetabular and femoral articular cartilage?  The answer is we can, and that's a typical model for a hip resurfacing where we put a thin shell over the bowl and a thin shell into the socket to replace the worn out articular cartilage.

            Is that new?  Not at all.  This was Sir John Chamley's first attempt at hip arthroplasty, and he used Teflon shells both on the socket and the femoral head, and here's some removed two years with complete wear through of the socket and the head.

            Then came the 1970s generation of his resurfacings using the materials available at that time, polyethylene on the socket, metal shell on the femoral head.

            Here's a case of ours, loose cup, a common complication of these cemented resurfacing models.

            Here's another one I revised, loose socket, femoral component solid.  When we sliced that component there are holes in the femoral head.  Why are they there?

            Here you can see an entry hole at the head-neck junction.


            Here you can see a granuloma on histology and on polarized light microscopy this head is stuffed full of polyethylene debris.  So the polyethylene debris problem has loosened the cup, and it's starting to erode the femoral head.

            Is hip resurfacing new in the United States?  Not at all.  There's some extremely good innovators in the United States.  The first two models of hip resurfacing are cemented, and the problems are largely as I outlined already in the previous models.

            The second two models are uncemented, and they did, indeed, address the issue satisfactorily of component loosening.  But of course, as you can see from those sample radiographs, they did not address the problem of tremendous osteolysis because of polyethylene debris generated as an inevitable result of the large head articulating on polyethylene.

            So to summarize the 1970s and '80s resurfacing, there's a small instance of femoral neck fracture and collapsed femoral heads.  That was manageable.  It wouldn't have finished resurfacing in the '70s and '80s.  The thing that finished it was the large head articulating on polyethylene gave a massive debris of polyethylene, and that loosened cemented cups and cause osteolysis with cementless components.

            So the conclusion of that was that hip resurfacing was not a viable operation with varying materials available at the time.  Our experience in Birmingham with the '70s and '80s resurfacings unfortunately was no better than anyone else's, and this is a survivorship curve with the cases we did, and you can see that we've got a 50 percent revision rate at six to seven years.

            As a trainee I put some of these in, and then as an orthopedic surgeon in the late '80s, I got to revise most of these cases.  So I was seeing patients turning up in my clinics with resurfacings that had failed because of polyethylene debris generated from large femoral heads.

            Also in my clinics, I was following up my predecessor's cases of metal-metal total hip replacements, and the first total hip replacement of a metal-metal nature in our hospital was done in 1966.  So I was seeing large headed implants, metal-metal, that were surviving well over 20 years.

            To summarize the metal-metal total hip replacement experience in use sine 1960, osteolysis is rare and severe osteolysis is exceedingly rare.  Peter Walker showed a few decades ago that equatorial bearing must be avoided.  Polar bearings work well, and large heads work well. 

            There you can see the difference between an equatorial bearing with the head jamming with increasing load and the successful polar bearing where the ball is slightly smaller than the inside of the cup.

            Now, it eventually dawned on me that what we needed to do to resurrect hip resurfacing as a viable operation was to take the concept of resurfacing and put with it the large headed metal-metal that had been clinically proven since 1960.  And interestingly, that thought didn't just occur to me.  It occurred also to the late Professor Heintz Wagner from Germany, and without knowing what each other was up to, we ended up both putting in our first metal-metal hip resurfacings in February 1991.

            So to summarize our first six years' experience with metal-metal, we started with a loosening problem.  We then had to change the component design, but cementless fixation of the socket was best.  Cemented fixation of the femoral head was best.

            The operation eventually was extremely reliable.  Hip function was good, and it satisfied all of those goals that we had talked about in the '70s, namely that where failure occurred revision to a total hip replacement was easy because the femoral  canal was not violated.

            We had three cups available in those early systems.  There are now 23 cuts available in the Birmingham hip resurfacing system, and we designed that and this first use was 1997.  The implant has a porous surface on the cup with a hydroxyapatite coating.

            And there you can see a single layer of beads integrally cast with the substrate metal, and that means that you don't have to heat the implant by centering to glue on the beads.  That has two effects.  The beads are highly unlikely to come off, and second, because it's not subjected to the heat of centering it does not destroy the carbide microstructure of the metal, which you can clearly see and which  we believe is very important for satisfactory metal-metal long-term bearing performance.

            This is a case that I had to remove for hematogenous infection a number of years after implantation, and you can see excellent bone ingrowth and on growth onto that socket. 

            This is another case, and you can see good ingrowth into the porous surface.  So this device appears to be acting as intended.

            On the femoral side, we've gone for a cemented component because that's what I've done for the last 13 and a half years, and that's, therefore, what we did with the Birmingham resurfacing.

            On the femur, we've got microinterlock of cancellus bone into the peripheral femoral head bone and a cementless stem.  So we have a specific design to line-to-line contact of implant on bone, and as the implant is inserted, the high pressure drives low viscosity cement into the peripheral femoral head giving good microinterlock in succession.

            On the right I'm not sure if you can see it, but that's blown up in an attempt to show you that that's a tapered stem put into a parallel hole, and from this point down the stem has a gap between it and the bone, and I did that deliberately to try and avoid distal loading and proximal stress healing of the implant.

            This is the first Birmingham hip resurfacing that I carried out, that was carried out ever on the 30th of July 1997 in a 38 year old man.  I can't quite read it from there.  That is his postop X-ray, one year and two-year X-ray.


            Five-year X-ray.  Now, what can we tell from this and other X-rays?  Well, he's got no heterotopic ossification.  He's got no radiolucent lines.  That's not too surprising.  Bach's radiographic study published from Melvin in the Journal of Bone and Joint Surgery showed that they had no radiolucent lines.

            There's no gross migration of this component, but what about minor migration?  Because we know from our Swedish colleagues that early minor migration can herald late loosening.  Can we tell about minor migration from a plain X-ray with an accuracy of plus-minus three millimeters?  No.

            Can we tell about stress shielding?  No, because you need a 30 percent change in bone density to be able to see it on a plain x-ray.  So we need special studies.  So we've done the RSA with our Swedish colleagues from the Carl Linsek Institute in Stockholm, and there's also been an RSA study published in the Journal of Bone and Joint Surgery also from Oxford, England.

            And on the femoral side, there is no measurable migration out to two years.  On the acetabular side, there's .2 millimeters of migration of the socket within the first two months, and then there's no further migration out to two years.

            If we're looking at density, we need to do a longitudinal Dexter study.  This has been done by Kishita (phonetic) and again published in the Journal of Bone and Joint Surgery, and his control group was a standard proximate porous coated cementless stem, and on the cementless stem the bone density in Zone 7 goes down.  That's what we all expect.

            Interestingly, with the Birmingham resurfacing in Zone 7, the bone density goes up.  So not only are we retaining bone in the femur, but when we load that femoral head normally in the absence of pain, the density in the upper femur goes up.

            So hip resurfacing is not new.  Metal-metal total hip replacement is not new.  The new thing is combining metal-metal with resurfacing.

            Now, all we need to know is:  is the metal line exposure to patients from a metal-metal Birmingham hip resurfacing any different to a contemporary metal-metal total hip replacement?

            We've done a number of studies.  We've included some of these for your interest.  This is whole blood cobalt analyzed by high resolution ICP mass spec, which is the most accurate means we have of measuring these.

            In our one-year Birminghams, with 50 and 54 heads, the levels and mean and standard deviations are shown.  The one year metasuls are shown, and they were 28 millimeter heads and the five-year Birminghams are show, and there's no difference between the large headed Birmingham metal-metal, and the 28 millimeter metasul metal-metal.

            For chromium, again, there's no difference between the large headed Birminghams and the 28 millimeter metasuls.

            Now, we were interested in the total amount of metal lines produced.  So we've got 24-hour output of cobalt on a lot of patients, and here is a longitudinal study of the Birminghams at two years, and the amount of metal line produced per day in the Birmingham with a 50 and 54 head is no different to the metal line produced in the 28 millimeter metasuls.

            At five year the Birminghams metal line production is no different to the 28 millimeter metasul production. 

            So to summarize, hip resurfacing is not new.  Metal-metal total hip replacement is not new.  With the new part of combining metal-metal with resurfacing, I've shown you that fixation of the femur and the acetabular component of the Birmingham hip resurfacing is good as demonstrated by two published RSA studies in the Journal of Bone and Joint Surgery.  The loading of the proximal femur is favorable as judged by DEXA and, again, published in the Journal of Bone and Joint Surgery.

            I've just shown you that the metal line exposure to patients is no different with a metal-metal Birmingham hip resurfacing compared to a 28 millimeter headed metasul bearing total hip replacement.

            So the conclusion is that the Birmingham resurfacing device is based upon the lessons learned from previous resurfacing designs and historic metal-metal bearings.

            Thank you.

            I'd now like to pass over to Tim Band, who was a young, fresh faced metallogist when I first met him, but I've grown him then by totally unreasonable demands in the last ten years.

            MR. BAND:  Thank you, Derek.

            Good morning.  My name is Tim Band.  I'm an employee of Smith & Nephew.  I'm going to present on the device description and preclinical testing.

            Before I do that, with the Chairman's permission, I'd like to present the panel with a physical example of the component for your physical consideration.

            PANEL CHAIRPERSON NAIDU:  That would be great.  Thank you.

            MR. BAND:  The Birmingham hip resurfacing device, as we've heard, is a device comprising of two components, the femoral head and acetabular  cup.  Both components are produced by high carbon caskable chromium material.  The femoral head is available in six sizes of four millimeter increments.  These are described by the external diameter of the femoral component and are in sizes between 38 millimeters and 58 millimeters.

            The femoral head achieves fixation and stability through the use of bone cement, and on the inside of the femoral head, there are six recesses to assist in the stability.

            There are 12 sizes of acetabular cup which means there are two sizes of cup per femoral head component.  Again, they're described by their external diameter, and they start from 42 millimeters in diameter through to 66 millimeters.

            So, for example, for a 38 millimeter femoral head component, you could use either a 44 or a 46 millimeter diameter acetabular cup in the event of any mismatch between the femoral device and the recently prepared acetabular.

            The acetabular cup is a porous coated, HA coated, cementless device, and they're also a dysplasia option and bridging cup with screws to provide primary stability.

            So to summarize the technical characteristics of this system, we can say that the material is a high carbon, as caskable chromium material.  The specification for the components allows the femoral head to be smaller than the acetabular cup to provide a polar bearing, as described by Mr. McMinn in his lessons learned presentation.

            As I showed you, the range of devices I described are appropriate to meet the anatomical sizes of presenting patients.  The acetabular cup is an uncemented HA porous coated device, and the femoral component achieves stability with the six recesses inside the femoral head and the use of bone cement.

            So both components are produced from cobalt chrome moybderium alloy conforming to ASTM F-75, an ISO 5832 specifications.  And the biocompatibility of this material has been proven by its benign clinical use since as early as the 1930s.

            The first specification for this material grade and composition was actually described in the early 1960s, but interestingly the material grade has remained fairly unchanged to the present day.

            So on the slide here you can see an example of a first generation metal-metal bearing which had survived for over 30 years in vivo.  Its failure mode was typical of this type of device, which was for tissue growth around the smooth cobalt chrome surface.

            We know, of course, today that this is not a very good fixation surface, but the slide on the right is a micrograph.  This is a section of the component which has then been polished and chemically etched to reveal the metallurgical phases in the microscope here.

            And as we can see, it's a biphasic material, which means there are two microstructural phases in the material.  The light background is the matrix, which is essentially cobalt, chrome, and molybdenum maximums, and the small, dark particulate component is a precipitate carbide which is made of chromium and molybdenum and carbine, which forms during the solidification of this alloy during the investment casting process.

            The Birmingham hip resurfacing which is shown below has its microstructural shown on the right-hand side, and as you can see, its specification was based upon the extensive forensic study of those first generation metal-metal bearings.  The microswitches are comparable.

            So to talk about our preclinical studies which were submitted to the FDA and summaries have been made in your panel packs, all of the studies were conducted in accordance with the FDA guidance documents and were provided for the components, the beaded surface and the H-8 coating.  The component testing included wear testing, friction testing, federal standard fatigue testing and kinematics were assessed by simulating range of motion.  Metallographic microstructure metrology examinations were also carried out to characterize the device.

            The preclinical studies carried out on the beaded surface included static shear, shear fatigue and static tensile strength testing, while on the substrate yield ultimate tensile elongation and abrasion testing were carried out.

            Finally, the preclinical studies on the hydroxylapatite coating included environmental stability, coating thickness, static shear and tensile strength, and analysis of the chemical and crystallographic characteristics of the coating.

            So, in summary, the device which was designed from the lessons learned that Mr. McMinn presented in his earlier presentation have all been evaluated on the preclinical data, and the results confirm that the device showed performance intended in vivo.

            Thank you.

            I'd now like to hand over to my colleague, Mr. Marcos Velez-Duran.

            MR. VALEZ-DURAN:  Marcos Valez-Duran with Smith & Nephew, and I'm to present the summary of the clinical data and clinical studies.

            This product was first available in Europe in the U.K. specifically in July of 1997.  Since then it has been currently marked in 23 countries.  Some of those countries include Canada, Australia, Japan, and all over Europe.

            There have been more than 33,000 implants implanted worldwide at the time of the PMA submission.

            The evidence of safety and effectiveness presented in this PMA is based on a consecutive series of 2,385 cases, surgeries that occur from July of 1997 to 2004.  The safety data that's presented in the PMA, and you have available in your packet, is a review of all 2,385 cases in this study.  The effectiveness data is based on a total of the first consecutive series of 1,626. 

            The effectiveness data is based on an independent review and follow-up by the Oswestry Center Registry, and those are surgeries from July 1997 to March of 2002.

            The radiographic study that you also have in your packet is based on a total of 124 cases.  Those are the first consecutive 124 cases conducted by Mr. McMinn.

            The study design has been described as prospective, and it's a consecutive series of like I mentioned, 2,385 procedures on the BHR, and no further design changes are in the series.  So 100 percent of the products in this series are of the same design.

            It's a single center by surgeries conducted by Mr. Derek McMinn and their Birmingham hospital using a proven surgical technical.

            There is five years' follow-up on the series, and we include safety, survivorship, pain or function assessment, and also patient satisfaction.

            The strength of this data is that one is consecutive clinical series, and for the safety assessment there was a 100 percent audit of all cases at patient records and at the Oswestry Center.

            On the effectiveness side, the prospective registry was independent of Mr. McMinn, was performed by the Oswestry Outcomes Center and include, like I mentioned previously, 1,656 cases, and in addition to the patient self-assessment of pain and function, there is also a patient satisfaction question.

            Also, as part of effectiveness, there's an independent radiographic evaluation at five years using a prospective protocol on 124 cases.

            There have been some questions about the comparability of the U.K. and the U.S. in terms of patient populations and practice of medicine, and we will surely talk about this later in this meeting.  Our observation is that there are similar target populations in both countries, and joint surgeries are performed in the same matter in both countries and similar in hospital procedures.

            There are three specific cohorts that are mentioned in the PMA.  There is the X-ray cohort, Oswestry cohort, and McMinn cohort.  I would like to make the clarification that they're all related to the very same patient, the total of 2,385 cases.  It just happened that the X-ray cohort and the Oswestry cohort were patients that were followed by the Oswestry Center registry, but all three cohorts were included in the safety data.

            The effectiveness measurement, the primary effectiveness measure for this PMA, incident by survivorship and secondary effectiveness measures include the Oswestry modified Harris Hip Score, which we will refer to through the presentation as OSHIP patient satisfaction.

            In addition, there is the five-year radiographic assessment.

            Safety measures include the primary safety measure is the number of revisions and percentage of the population study, as well as all other adverse events.

            The study population of these cases include mostly men with osteoarthritis.

            In terms of accountability, we have an excellent patient follow-up at five years of 90.8 percent, and at five-year survivorship of 98.4 percent.  The survivorship is also available and is consistent with publish reports in British Journal of Bone and Joint.  You can see series from Australia, other parts of Europe and also U.K.  The survivorship estimate is consistent to the one we observe in this PMA. 

            In addition, the Oswestry Center followed a total of 5,000 cases, including Mr. McMinn.  So 140 surgeons collected 3,374 cases in over five years, a maximum of five years to survivorship is 96.3 percent.  So it is consistent with what we see with Mr. McMinn.  So the success, other clinicians are able to repeat the same success.

            In terms of clinician rates as compared to a comparative group presented in the PMA for the BHR on all cohorts or the specific group that's X-ray, Oswestry cohort, it's the primary population group for effectiveness.  You can see that the percent of revision for BHR compares very well with published articles on existing total hip replacements.

            The Oswestry modified Harris Hip Score was developed by the Oswestry Outcomes Center.  It is validated and compares well to the Harris Hip Score. It's a patient self-assessment of pain and function, and it took questions and scored similar to the Harris Hip Score, with a difference that deflection and extension questions are different from the Harris Hip Score in that the OSHIP score asks relevant questions to patients and activity of daily living.

            There's also patient self-assessment of satisfaction that's very simple and we will present it later in this presentation.

            To compare the Harris Hip Score to the Oswestry hip score, I put together the two scoring systems, and as you can notice, the only difference is in the section of shoes and socks and deformity and range of motion.  Those have been captured in the Oswestry score in the area of movement, and it equals 13 points.  So at the end of the day both scoring systems are up to 100 points.

            The result of the OSHIP scores is as follows.  There is a baseline.  There is an outreach of 59.8 points and at five years, 95 points, a significant improvement over time.

            And if you were to define success based on the scoring system as patients that have greater than or equal to 80 points at five years, 93.2 percent of those patients are considered successful.  Even if we were to define success on the total score as greater than or equal to 90 points, at five years our success rate would be 85.5 percent.

            The second, in the patient satisfaction, these are the questions that were asked of the patients.  You can see they're very straightforward questions, no room for misinterpretation of the question.

            And at five years, patient satisfaction was 99.5 percent of extremely pleased or pleased with the operations.

            On the radiographic data, there were predefined failure and success criteria.  The 35 failure criteria is presented here as presence of the incomplete or complete radiolucencies or a radiolucency in all zones and migration of components greater than two millimeters or a change in acetabular orientation of greater than or equal to five degrees.

            The five-year radiographic success result is 97.2 percent.  There were three out of the 108 that were radiographic failures, representative of 2.8 percent, and we want to note that no radiographic failure in the series studied in revision.

            Those radiographic findings are consistent with what Mr. McMinn presented previously on ISA studies, which are the best ways of assessing migration and other radiographic observations.

            The primary safety outcome with revision and of the total, 2,385 cases there were 27 revisions and probably as previously explained, that represented 1.1 percent of the total population.  The majority of the revisions were on the femoral neck fracture, and they occurred as we can see by the average time to revision, early, followed by infection and collapsible head, which happen between two and three years on an average.

            We wanted to spend a little time explaining the method of data collection for our first event.  We collected all first events in patient charged and the OSHIP questionnaires.  There was no differentiation at the time of the collection between clinical observation and actual event.  The collection of the adverse event was conducted by separate consulting group.  Mr. McMinn was not included in that assessment or review of the chart.

            Some hips have more than one adverse event over time, and adverse event documented multiple times were treated as separate events.  It was a very comprehensive review of everything that was in the patient chart.

            In terms of result and device related events for the overall cohort, the AND, actually all of the device related related events were at one percent or less.  For AND, there's one percent for which were at grade the one year postdoc, and 31 out of the 35 total observations of AVN were interoperative observations.

            On the femoral hip collapse, there were 15 total reported events.  Eight were evacuated.  Seven were interoperative observations, again, femoral neck fractures less than one percent, component migration less than one percent.

            In summary, we have a very large series of hips, 2,305, followed for a long term, a long-term follow-up for five years.  Excellent patient follow-up of 90.8 percent at five years. 

            There's an independent assessment of pain and function by a separate research center that was presented and not related to Mr. McMinn.  There's an assessment of patient satisfaction, independent critique or evaluation, confirmation of the PMA there resolved by other BHR series, and literature result for a total hip replacement. 

            In the effectiveness side, the survivorship was 98.4 at five years.  The Oswestry score was an average of 95 at five years; patient satisfaction, 99.5, or extremely pleased with the operation at five years; radiographic evaluation, 99.2 success at five years.  In terms of safety, revisions are 27, which represent only 1.13 percent of the total population and very low instance of the adverse event, like I mentioned, all of them less than or equal to one percent.

            In conclusion, we have talked before in the presentations before me that the BHR device design was based upon the lessons learned from previous resurfacing design and historical metal-on-metal bearings; that the clinical data confirms that the device should perform as is intended in vivo, and now the clinical data offer a reasonable assurance of safety and effectiveness.

            Thank you.


            MR. DeMUTH:  Hello.  I'm George DeMuth.  I'm a consultant to Smith & Nephew, and I do not have an interest, financial interest, in the product or the company.

            So I'm going to provide some statistical commentary and go to the next slide.

            I had a few background comments, and then I want to touch on the accountability and efficacy in terms of surviving OSHIP and have some conclusions.  Go ahead.

            I want to just start with some interpretational issues and we'll come back to them later, but the OSHIP has actually been collected prospectively, but we're analyzing it down the way now, and that constitutes some independent follow-up, but we do have the single site issue and trying to figure out what to compare against, which is part of what I think the discussion will be today.

            The FDA raised questions about pooling of the data.  My first reaction was it's a single site, single physician.  People are being handled the same way.  So I'd like to be inclusive, but it's also of interest to see if we can see if some of the subsets of patients are behaving differently.  They're better or worse than expected.

            And I want to take this as a brief note to thank the FDA because they provide some nice commentary and, I think, added to the analysis.

            So the data I'm going to present will actually be for unilateral hips in the common tables.  I don't have an n here.  It was actually more than 1,100, and it will be the X-ray and Oswestry cohorts.

            Go to the next slide.

            And the reason I want to do the X-ray and Oswestry cohorts is because there we have the data follow-up coming off the OSHIP data and we can use that for censoring.  So in this case we know we've gotten contact from the patient and have a much better feel for whether they, you know, are being aware or we know what they're doing, and for the McMinn cohort, those patients we would have to use a cutoff date, and that would be conservative.

            So we'll focus on the group that we know the best about.

            Go ahead.

            All right.  So this is actually the OSHIP thing.  I want to start and talk about here on the bottom of the slide because in terms of survival we want to know what their last visit is, and they may have some intermediate missing OSHIP data, but we want to know at the end how they're doing.

            So we had a very nice, 91 percent follow-up at five years across the cohorts, and very good, 88, greater than 85 percent three and four.  So I think in terms of survival in that kind of follow-up we're very happy about the data.

            Now, there's some missing data at the one and two year data on the OSHIP insert, some missing baseline that doesn't show up here as well, and so in terms of evaluating the OSHIP, we have to have that in mind.

            Now, sort of back up to the top, there wasn't a flag in the Oswestry Outcomes Center that said patients were discontinued.  So what we did is just to try to get some idea.  That was to look at whether patients had missed their last two expected visits, and that doesn't even mean that they won't come back in, but in this case, it was 84 out of the 1,626 hips.  That's just slightly more than five percent.

            So I think in terms of the survival endpoint, that's good as well.  So let's go ahead.

            I don't want to spend a lot of time here, but this just a report.  Really the rate here for this cohort, the unilateral X-ray and Oswestry cohorts:  98.3 percent survival, 1.7 percent revision rate.  The upper bound is 3.1 percent of revision rate.  So it looks solid as we expect with a lot of patients.

            So let's go.

            So without trying to do some of these covariate analyses, I think there will be some discussion later about P values.  I'm just presenting both here.  None of these are significant for the cohort and the gender and the age at five years.  There's, you know, a very small survival difference.

            I wanted to pick up and just look at the baseline data, people that had baseline and didn't.  It's not different, but the people that didn't have baseline available were just slightly worse, 1.7 percent -- 1.3 percent.  Sorry.

            Let's go ahead.

            We looked at the diagnosis.  This is the reason for resurfacing.  We saw a significant or marginally significant effect for AVN.  Nothing else was different.  All of these P values are pair-wise comparisons of the osteoarthritis group.  So I don't think that's an expected result.

            Let's continue.

            This just shows the all hips, unilateral versus bilateral.  Bilateral is still a little bit better, but again, good across the board.

            Go ahead.

            So the X-ray and Oswestry cohort, I think, is a good focus here.  It's the best follow-up information we have, and it looks very good in terms of the data, the file we have and the revision rates, less than two percent at five years, 3.1 percent upper bound.

            There's a limited number of revisions.  Probably doesn't have a lot of power to see subgroup effects, and there was one significant effect in terms of AVN.

            So go ahead.

            This is the OSHIP data, and this is observed data, and I don't want to spend long here because you've seen it, but a big increase from 60 to mid-90s.  That's basically on average 35 to 36 points as an increase, and it's just very large.  So let's go to the next slide.

            I think this probably to me is a little bit better because you see the majority of patients less than 70, and then everybody is piled up to be greater than 90 or greater than 80.  So it looks like you have a very strong response in the OSHIP.

            And so this summary is just a big response, a good response for five years.

            We did model the OSTET (phonetic) two-plus year results.  That's basically figuring if they had a two-year point, you took that one or you took the next one after it.  Almost everybody has that, and there were some significant effects there.  I think we just have a lot of power because now we're talking about, you know, 1,100 hits and effects on the order of three points, you know, would be significant, three to four points.

            And in terms of a 35 point increase, I'm not sure that we have to worry about that a lot.

            So if you touch on this missing data issue back here, a couple of cuts of it and a couple of ways to look at it.  One, it summarized patients that had missing baseline, looked at their post treatment, and you just compared that.  You wrote sort of descriptively to the other patients.  Those differences are less than three points, around two points, indicates that those patients may have had a slightly lower mean.

            Didn't repeat a measured mile, but there's very little difference there, but that's probably dominated by the observed data.

            And then the other point is that if you look at the two-plus data, it looks like, again, that has very high value.  So I don't think you can rule out an association here, but it looks small in respect to the big improvement we're seeing on the average.

            Let's go ahead.

            This is just patients and just to point out and we saw it before; all most everybody is pleased, and this coincides with low survival, low revision rates and the very big improvement in OSHIP.

            So let's go ahead.

            This is going to touch back.  I hope to capture here a compendium of the issues that will be raised later on about the study design, a single investigator, no control.  There are no a priori sample size.  I think some of these I answered, and then we have a very large sample size.  I think the OSHIP was very sensitive to reasonably small differences relative to the treatment size. 

            I think we had very good follow-up in that.  We got very tight confidence bounds around the survival.  I think the survival is a very objective endpoint here, and so I think that's favorable.

            We have external radiographic review.  It helps, I think that the efficacy data is being collected in terms of OSHIP externally and an independent safety review.  So I think it's important to put this study sort of in the context of the other studies that we had, the published literature where we see similar survival rates and nice or, you know, low survival rates and high OSHIP satisfaction look internally consistent.

            So second page.

            There aren't too many comments about this.  There is a question about the study design for the validation work, and I don't have too many comments.  Just say there were some usually high correlations.  There was a difference in the mean OSHIP and HHS.  At least it's in the direction that the OSHIP is lower, which would make it seem more conservative.

            I think, too, we have to go back and look at OSHIP in terms of patient satisfaction, and that consistency as well.

            All right.  Go ahead.

            So in summary, back to sort of what I thought were data points here, there's a large number of procedures.  We have, I think, very good follow-up.  Survival rate in the X-ray and Oswestry cohort for the hips of 83. -- 98.3 percent, sorry.  Large OSHIP improvements.  Patient satisfaction look very good, and to the extent we did a sensitivity analysis, our evaluations of the missing baseline data, missing data don't lead me to believe it should interfere with a major interpretation of a very large increase from baseline.

            This brings me back to the slide we've been building on.  I think we've started with the lessons learned, and there's a history about how the device is what it is and how we should have expectations of good survival for this device.  In the preclinical data, overview of the clinical data and now some statistical analyses showing, you know, very good effects and a lot of data with good follow-up.

            So I'll turn it back to Neal.

            Thanks very much.

            MR. DEFIBAUGH:  Hello.  My name is Neal Defibaugh, and I'm an employee of Smith & Nephew.

            I would like to spend just a few minutes talking about some approval considerations, namely our proposed labeling, training and post approval study.

            Next slide, please.

            We believe the data that we have shown you today demonstrates that the BHR device is intended or can be used for patients who are at risk of requiring more than one hip joint replacement over their lifetimes, and while it's impossible to predict exactly who may require a future revisions, some factors that are known to increase the risk of revision or include patients of a young age and anybody less than 55 years and/or high activity level.

            Next slide, please.

            Therefore, specific indications for use that we believe the data supports include noninflammatory degenerative joint disease such as osteoarthritis, AVN, dysplasia, DDH, as well as inflammatory DJD, such as rheumatoid arthritis.

            Next slide please.

            Obviously there are contraindications to the use of the BHR device.  General complications you would see for any type of hip replacement include things like infection, patients who are skeletally immature, and patients with conditions that would compromise the implant stability or the post operative recovery period. 

            Review of clinical data also indicates that very clearly patients with inadequate bone stock, such as severe osteopenia, patients with avascular necrosis, greater than 50 percent involvement of the femoral head, and patients with cysts, multiple cysts in the femoral head greater than one centimeters.  These types of conditions are not appropriate for use of the BHR system.

            Additionally, out of an abundance of caution, given the metal-metal nature of the device, we believe that women of child bearing age and patients with renal failure should be contraindicated.

            Just briefly, I'll touch on training.  It's our intention to send a group of what we call core surgeons to view live surgery in Birmingham, U.K., as well as receive additional lecture and workshop interactions.  These surgeons would then be the ones who would come back to the U.S. and ultimately after approval train other surgeons who have an interest in the BHR system.

            This training would be held in North America, and there would also be additional information available for surgeon resources, such as surgical techniques and Website information with advice and technique information.

            FDA requested us to provide a post approval study protocol, and although we do not believe necessarily that a post approval study is needed due to the long-term follow-up and large patient population we already have, we certainly submitted a protocol that was based upon a template of a previously approved device, Class 3 hip device that we had provided the FDA earlier.  So that's the template for this, generally a prospective, nonrandomized survivorship study with clinical radiographic evaluations through five years.

            Next slide.

            In years six through ten, we will do postcard follow-up.  Any explants returned to the sponsor would be analyzed as appropriate, and we would make appropriate progress reports to FDA obviously.

            So to round out the theme we've been building on, the device design of BHR is based upon lessons learned from previous resurfacing designs and metal-metal bearings.  Extensive preclinical testing confirms the device should perform as intended in vivo.  The clinical data confirms there's a reasonable assurance that the safety and effectiveness of the device.

            The statistical analysis demonstrates robustness of the efficacy results, and we believe, finally, that the clinical results support the proposed labeling.

            I would like to thank the panel and FDA for all of their efforts on behalf of this PMA.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            I'd like to thank the sponsor and the representatives for their presentations.

            I'd like to address the panel now.  Remember that you'll have time for questions this afternoon, but nevertheless, if you have extensive questions for the sponsor at this point, please ask it so that they'll have some time to prepare for the afternoon session.

            Does anybody on the panel have any questions at this time for the sponsor?  Dr. Mabrey.

            DR. MABREY:  With regards to the surgical training of the core surgeons and then the subsequent training of surgeons within the United States, if you could provide us some information on the length of that training, the number of cases that they would see, whether or not observation of the trained surgeons within the United States would be performed, and how they would be evaluated as to their proficiency in implanting the device; whether or not that would be a requirement for using the device or simple attendance at the training course.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Anybody from sponsor willing to tackle that question?

            MR. VALEZ-DURAN:  I just wanted to make sure.  So you want us to address now or later, but we're ready to address it now.

            PANEL CHAIRPERSON NAIDU:  Yeah, we have a little time.  So we can go ahead and we've got 15 minutes before we break.

            MR. VALEZ-DURAN:  To discuss the issues of training we have brought a couple of people that are working very closely on the plan, and I would defer that question to one of my colleagues, and also we have some clinicians that have also participated in that discussion.

            So Marc.

            DR. THOMAS:  My name is Dr. Marc Thomas, and I'm an employee with Smith & Nephew.

            The training of the core group of surgeons will initially be a two-day, on-the-ground course associated with didactic issue and hands-on component which will be in the form of sawbones (phonetic) and also viewing live surgery.

            After that course that they will attend, we had to time it within an acceptable period of when the people or the surgeons will get the device in their hands.  Before that occurs, we will have another meeting where, once again, there will be follow-up lectures to solidify the information for that core group of surgeons.

            There will be another opportunity for a hands on component, and there will be another opportunity to view our surgery to see any tips or tricks and have an opportunity for a last question and answer session with one of the designing surgeons from the U.K.

            At that stage if the FDA allows, they will be allowed access to the device and work on their own proficiency through performing the surgery on their own patients for a period of time until they feel that they have the self-proficiency to stand up and teach their colleagues.

            DR. MABREY:  Will there be an opportunity for cadaveric dissection as well or will the entire procedure be conducted on sawbones?

            And just for the audience's clarification, sawbones are basically plastic representations of the bones without soft tissue attachments versus cadaveric, which includes all of the soft tissue and muscle attachments.

            DR. THOMAS:  There will be an overture in the United States for use of the instrumentation on cadaveric specimens.

            DR. MABREY:  Thank you.

            MR. VALEZ-DURAN:  I would also like to take the opportunity to introduce Dr. Rogerson.  He is a U.S. clinician from Wisconsin, and he has reviewed also our plan for training and we'd like his comments on that.

            DR. ROGERSON:  Thank you.

            I'd like to preface this that I am an orthopedic surgeon in Madison, Wisconsin, specializing in the treatment of shoulder, hip, and knee from arthroscopy to joint replacement.  I've been very active in the Arthroscopy Association, presently on the board of directors, and vice chair of the Orthopedic Learning Center  where we do do cadaveric courses and training for arthroscopic surgeons learning shoulder, ship, and knee arthroscopy.

            I have no financial interest in this whatsoever, although my air fare and hotel room are being reimbursed by Smith & Nephew, and my main interest in this proceeding is in trying to help facilitate the introduction of this technology in the United States.  Some of that is a selfish, personal reason.  I have some arthritis brewing in my hip, and I have been very interested in looking at alternative to conventional total hip arthroplasty, as a lot of my patients.  And I have personally send about 30 patients over to Europe to have Dr. Kunda Schmidt perform the Birmingham replacement, and I have been basically amazed at the results of these patients coming back and their activity level.

            I've also visited Dr. Schmalzeried, Michael Mont and staff here in the United States.  So I have no particular allegiance to Smith & Nephew, but I do feel that the technology that is being presented to you is extremely important, and I hope that it at some point will make its way through the FDA.

            In terms of the training program, which I think is imperative because when you look at this replacement, it is a different animal than a conventional total hip replacement.  You actually do not have the exposure that you'd have by cutting off the head and neck to evaluate and to reinasutablum (phonetic).

            The soft tissue dissection is critical and the learning curve for this procedure is going to be much greater for surgeons in any setting for this procedure.

            Smith & Nephew, as you've seen, has had extensive experience with this procedure worldwide.  They've developed a well structured, educational experience that has previously been employed in other countries, and what it really consists of is graduated release of this prosthesis into the market so that it basically guards against a full release, a full lodge with poor training and ultimately poor results.

            The core group of surgeons that would go to Birmingham for the two-day intensive training would be subjected to educational didactic lectures, which would go through the history of the development, the tribology (phonetic), the metallurgy, the surgical technique, the indications and contraindications for the procedure, and particularly the surgical technique, and that would be augmented by surgical observation of Dr. McMinn or Dr. Treacy, and also work with sawbones in Europe would have a hands-on laboratory in that setting.

            That would hopefully be about two months before the introduction of the device on the market, and then right before the introduction of the device on the market, there would be another intensive training session for that core group of surgeons, and they would go to a learning facility, again have didactic lectures, but more importantly, more exposure to the surgical procedure both in terms of DVD/CD education.  Sometimes you can actually see more.  My experience at the learning center is that although the cadaveric dissection is critical, sometimes a well developed and particularly well executed DVD of the procedure where you can zoom in and show the critical parts of that procedure are equally as valuable as the cadaveric dissection.

            So during that, right before the release those surgeons would be, again, educated as to the indications, contraindications, but more importantly have a very fresh experience in terms of the operative exposure and operative technique.

            They would then go back to their local setting and start performing these operations on patients.  The desire of the company is to have field representatives who are very experienced in this technique, go out to this core group of physicians and observe and help in the first ten cases and to make sure that they are getting through their learning curve, and when that learning curve has -- and that may vary depending on the core group of surgeons -- but when that learning curve has bene attained, then those core group of surgeons would become the teachers of the next group that would occur at the full launch.

            So that when it comes time for the full launch for the rest of the United States, there would be a number of regional facilities where you have this core group of surgeons that have excellent experience.  The participants would then go to a lab setting in the United States where they would go through a very similar procedure that the core group of doctors went through in England, do that in the United States and then have the ability to have monitoring come into their OR with the field representatives, but also have the opportunity to visit the core group of surgeons, observe surgery and get the same type of course that the core surgeons got in Europe.

            So I think that from my standpoint in terms of learning how to do a procedure that Smith & Nephew has thought this out.  There seems to be a well structured program that has, I think, been effective worldwide up to this point, and I think that it will be even more defined and meticulous in the United States because of the wider launch that will occur here.

            DR. MABREY:  How many surgeries do you anticipate being part of this initial learning curve for your 15 core surgeons?

            DR. ROGERSON:  I think that the level of expertise in terms of comfort level in the learning curve, I would think that you would need to have probably 30 surgeries under your belt before you would start to get through the learning curve.

            That could vary depending on the surgeon and what his experience has been.  For example, I've up to this point been doing a reasonable number of metal on metal, big thermal head stem prostheses.  So I've got very good experience with the acetabular components of a joint replacement when it's a metal on metal.

            So the learning curve of the acetabular might be easier for one surgeon versus another, but I think that it would be hard to put a number of cases on it, but I think the comfort level would be documented by the company and by the field representatives when somebody has gotten through their learning curve.

            PANEL CHAIRPERSON NAIDU:  Thank you. Dr. Blumenstein.

            DR. BLUMENSTEIN:  How many deaths were in these cohorts?

            MR. VALEZ-DURAN:  I'm sorry.  Could you repeat the question?

            DR. BLUMENSTEIN:  How many patients died?

            MR. VALEZ-DURAN:  If you'll just wait a minute, we have the data.  We're looking for it.

            DR. SKINNER:  Could I comment?  I believe there are 20 deaths, and that was one of the questions I was concerned about because an average age of 53 years, 20 deaths in five years in that age group sounds like a lot a deaths.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            MS. MARLOW:  As far as the number of -- I'm sorry.  My name is Marie Marlow.  I'm a consultant to Smith, and Nephew and as such I'm paid for work I do for them.

            In your panel packets, there are narratives for each of the deaths that occurred in this trial, the reasons for them, length of time after surgery.  We can get a summary table together for you and  discuss that in greater detail if you'd like.

            DR. BLUMENSTEIN:  How were the deaths handled with respect to the estimation of survivorship?

            MS. MARLOW:  I'll pass that to a statistician.

            MR. DeMUTH:  Yes.  Actually -- George DeMuth -- we censored at the time of death if there was a death available.  So if there was a revision, well, if they had revised the packet in the revision, but we just censored at the time of death those patients.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            DR. SKINNER:  Could I follow up on that, please, Dr. Naidu?


            DR. SKINNER:  I'd be interested to know what the number of expected deaths in a group of 2,385 individuals would be over the following five years with an average age of 53 to see if that's actually an expected death number of 20, which is the number of deaths in it.

            I looked at the packet.  It doesn't look like any of them were caused by the hip.  so it seems like an awful lot of deaths in a group of 53 year old individuals.

            DR. RICHARDSON:  My name is James Richardson, Professor of Orthopedics at Oslostree (phonetic), U.K..

            I'm an orthopedic surgeon.  I've been studying hip resurfacing for eight years now, and I'm not affiliated with Smith & Nephew.  I have got funding in order to carry out my research, but I have no personal benefit.

            I've been interested in this question, and using the data from the Outcome Center to try and find a solution to this.  The average death rate of 20 out of 2,385, you can check my mathematics, but it's in the order of, as a percentage, one percent, but that's over three years, the average 3.3 years.  Again, I'm just giving general figures and I will explain why.

            The death rate per year then comes to about .3 percent.  Now, studying death rates is tricky because each age band, I'm afraid we all have a different expectation of death in each year, but for the age group within the U.K., 55 to 59, information I have is that for ladies who are the tougher of the species, the expected death rate is .47 percent, whereas for us feeble men it's .74, almost twice.

            So per year, you can see that the death rates in the hip replacement group are actually slightly below what might be predicted.

            I qualify my comments by saying that if you really want to analyze this in detail, then looking at simple averages it not sufficient.  You really need to break down all the data and look at the death rates for each age band and it becomes more complicated.

            But at least from what I can find out, I don't think it's a major concern.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Dr. Mayor, you had a question?

            DR. MAYOR:  Yes, I have a couple of questions for Derek McMinn.

            Congratulations for a prodigious body of work, but I wonder if you could clarify a couple of issues for me.  You suggested that revision is a fairly straightforward undertaking, although it might not be the most enthusiastic problem that the patient might perceive.  How do you handle the acetabulum in the process of doing the revisions that you've done, particularly, say, for femoral neck fracture?

            And then the second question, if I can present them both together, I was confused by the description of femoral head collapse and identified as distinction from a vascular necrosis and wondered how that distinction was drawn.

            DR. McMINN:  I'll try to bring up the slide.

            MS. MARLOW:  Chairman Naidu, with your permission, may I ask Dr. Mayor a question?


            MS. MARLOW:  Thank you.

            Dr. Mayor, in addition to Mr. McMinn, Dr. Rorabeck is here because he has experiences with these cases in Canada also.  So if you would like another perspective in addition to Mr. McMinn's, please feel free to ask Dr. Rorabeck about his experience.

            DR. MAYOR:  I am always happy to have Dr. Rorabeck's perspective.

            DR. McMINN:  Derek McMinn again.

            On the question of what we do with the acetabulum in the event of, let's say, a femoral neck fracture, the acetabulum is left alone, provided the acetabulum is in a good position and at surgery there's no adverse effect like, for example, evidence of age loading or wear of the component.

            So you put in cemented or cementless stem, according to your preference as a surgeon, and then put a modular bowl on it that matches the existing cup.  So that's quite straightforward.

            I'm happy to answer anything else on that if you have further questions.

            Trying to differentiate between a collapsed head and avascular necrosis, that's a good question because it's a bit like trying to identify the cause of a fire when you house has burned down.  I've been there 20 years ago, and when you're going through the ashes, it's quite difficult and tough to decide on the cause of the fire.

            If you get collapse of the head, then it's the devil's own job to differentiate between collapse from, let's say, squashing of a lot of cysts or a vascular necrosis that has occurred.

            The only ones where you're pretty clear that it's avascular necrosis is if no complete collapse has occurred, and histology then shows you a segment of the head distal to the collapsed area that has a vascular necrosis changes in it.

            But I absolutely agree that some of these cases could have been categorized in avascular necrosis or collapsed head, and when the collapse has occurred, it's very difficult to know what the original pathology was.

            DR. MAYOR:  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. McMinn.

            Any other questions from the panel?  Dr. Skinner.

            DR. SKINNER:  Harry Skinner.

            I wanted to also ask Mr. McMinn some questions if he could answer.

            First of all, could I follow up on that femoral neck fracture thing?

            I went through the data, and it looks like most of the femoral neck fracture, head collapse things happened earlier on since '97, and the package insert is going to address that by limiting the indications or contraindications to less than 50 percent of femoral head involvement in avascular necrosis.

            Do you think or does the data show that the number of femoral head problems has decreased since you have sort of instituted those changes in your selection criteria?

            DR. McMINN:  It's true -- Derek McMinn again -- it's true what you say about avascular necrosis being a major problem for femoral head collapse.  The highest percentage group of all in our femoral head collapses are avascular necrosis.  So in those with a preexisting diagnosis of avascular necrosis, they had a four percent collapse rate.

            So if you have extensive AND and want to do a resurfacing, what we have subsequently found out is that whether you get collapse or not (a) depends on the magnitude of the original femoral head lesion and (b) whether the pathology is recurring or not.

            So if you had, for example, a traumatic dislocation and it was a once an event time, then the chances of a subsequent collapse of that head are small.

            However, if you have avascular necrosis caused by, for example, alcoholics, increased alcohol intake, they have a bad postop record with further collapse of their heads.  Mind you they also have a bad postop record with total hip replacement and falling out of bed and dislocating their hips, et cetera.

            There was one other point which I've just forgotten.  Could you remind me of the other point in your question?

            DR. SKINNER:  I think you've already answered it.  You think that the changes in the labeling basically, the indications/contraindications, will address this because you've changed your criteria for selection, I gathered from the data given to us.

            So you think that the incidence of femoral neck fracture collapse is going to be decreased as time goes on, at least in your series.

            DR. McMINN:  The answer to that is it has reduced with the passage of time, but that really is based on the light of experience and understanding that there's no point in trying to attempt the impossible, and if you try and fix onto a femoral head remnant with virtually no bone in it, it's going to fall to bits.

            So I doubt a break with a sense of realism as I grew older took part here.

            DR. SKINNER:  Another question.  I would like to get a sense of your practice of orthopedic surgery to get an idea of what the selection process for these 1,685 patients was, 2,4085 patients.  Do the patients come into your place, you select them.  They either have enough arthritis to have surgery and you do a surface replacement on them or if they don't have enough and you send them away or they have enough and you decide that they're a good candidate for a surface replacement or they're a bad candidate, and then you do a total hip on them.

            I mean, how does that go?

            And a second thing.  Is your practice a referral practice with only non -- I'm not sure about the British system -- but non-health service patients or whatever the practice is?  Is there a selection process there?

            DR. McMINN:  Okay.  Mine is mainly a referral practice, but I do get through family practitioners referred local patients.  So typically in my clinic out of 20 patients I may see, 18 of them would be referrals from outside my city, typically referred by another orthopedic surgeon because the patient was young, active, had good bone stock, and wanted a resurfacing.

            So a lot of the patients that I see have come specifically for a Birmingham hip resurfacing sent there by their orthopedic surgeon, and quite a number come from other countries.  For example, we were looking through the data.  Sixty-five of my patients included in the information are from North America.  So it's an international practice.

            However, if I get the local family doctor sending a patient from Edgebastin (phonetic), particularly a lady in her 70s with an arthritic hip, the question of hip resurfacing will never get mentioned, and I will do a total hip replacement on her without discussing what my main activity is, namely, hip resurfacing.  Because that's on a particular group of patients who are younger, more active, and are likely to cause a failure of a total hip replacement.

            DR. SKINNER:  Roughly how many total hips have you done in the last seven years?

            DR. McMINN:  I would do somewhere between 50 and 100 hips a year.  That number has decreased as I've got a little older and not doing quite as much, and my younger colleagues, I tend to try and pass on total hips, and particularly revision total hips to, but resurfacing, I still enjoy doing that.  So I don't pass those on.

            DR. SKINNER:  Now, a little bit of the clinical data.  I noticed in your data that you had a fairly high pain rate.  Something like 15 percent of your patients reported pain compared to some of the ceramic studies.  There was quite a bit lower rate of pain in those. 

            Is there a particular reason for that?

            DR. McMINN:  Well, you have to understand that this is a review of my notes.  You're holding the microphone.  Do you want to interrupt me?


            DR. McMINN:  This is a review of my notes by an outside group of consultants who I understand were detailed to go through that with a fine toothed comb and record everything.

            Now, when you look longitudinally at the data, it's pretty bizarre because postoperatively small percentage of patients recorded pain.  Was there any mention of pain?  Yet at a year, there was a higher percentage, and you'd think that's completely bizarre.

            But this was from the records and there was only mention made of pain if the pain was worse than either the doctor or the patient anticipated.  So that's why postoperatively there's a low figure, percentage, for pain and later on there's a higher percentage for pain.

            In other words, the patients said, "Hey, what's this pain?"  So any level of pain that was recorded.

            DR. SKINNER:  Well, I happen to think that the 15 percent is more like a realistic number, to be honest, than the three percent or whatever, but how about limp?  A fair number of your patients also seem to have a limp even I think a year out it was.

            MS. MARLOW:  Again, Marie Marlow.

            We were the company that Smith & Nephew -- sorry.  I'm involved with the company that did the audit for Smith & Nephew of these data, and when we deployed the auditors, we told them that since this is a retrospective review, that every single incident had to be recorded.

            What we did also was combined comments that we found in the Oswestry database, along with the comments that we found in Mr. McMinn's series.  So there's an area on the Oswestry form, for example, for a patient to make notes, offer comments.  If a patient made a comment in that field, we didn't censor it; we didn't filter it.  If they said, "I have a mild limp," we recorded that as an adverse event.

            DR. SKINNER:  Okay.

            MS. MARLOW:  Is that helpful?

            DR. SKINNER:  Yes, that's very helpful.  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.  Thank you, Dr. McMinn.

            Yes, Ms. Whittington.

            MS. WHITTINGTON:  I have a question.  There seemed to be an extremely high number of patients with wound exudate, and if that data is taken from office notes after the fact, that's even more disconcerting to me since that's a primary indicator for potential late infection.

            MS. MARLOW:  Again, Marie Marlow.

            Thanks for that question.  Part of the office notes or part of the records in the patient files at the McMinn Center included the discharge notes, as well as the op notes.  So if there is a comment on the discharge note about a patient being instructed as to dressing changed, wound care, that was listed as an adverse event because, again, the auditors were instructed not to filter anything, not to make interpretations about whether something was an adverse event or not.  If there was a comment there, it would have recorded it.

            MS. WHITTINGTON:  Was there a correlation to those patients with the patients who had revision or later infections as noted in the --

            MS. MARLOW:  I'm sorry.  I didn't answer your question about the late one.

            MS. WHITTINGTON:  Right.

            MS. MARLOW:  Postop the cutoff was 30 days.  So the column that you're looking at in the adverse events table, that's a 30-day cutoff.  Anything after 30 days got moved into the one-year column.

            MS. WHITTINGTON:  Okay.  Was there a correlation to those patients to the patients who did develop late wound infections and had some of them revisions that looked like later?

            MS. MARLOW:  I don't know from the data I have in front of me.  I'll see if we can get that answer for you.

            MS. WHITTINGTON:  Okay.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Why don't we take a short break now for 15 minutes and we'll come back with the FDA presentation then?

            Thank you.

            (Whereupon, the foregoing matter went off the record at 10:46 a.m. and went back on the record at 11:03 a.m.)

            PANEL CHAIRPERSON NAIDU:  We will now have the FDA presentation on this PMA.  The first FDA presenter is Mr. John Goode, the review team leader for this PMA.  He will introduce the other FDA presenter.

            Mr. Goode.

            MR. GOODE:  Thank you, Dr. Naidu.

            Good morning.  My name is John Goode.  I'm a biomedical engineer and reviewer in the Orthopedic Devices Branch and the lead reviewer for the Smith & Nephew premarket approval application for the Birmingham hip resurfacing system.  I will be presenting the device description, preclinical and clinical information, and FDA statistician Dr. Chang Lao will summarize the statistical information in the PMA.

            But first I would like to identify some of the reasons for the panel meeting and the topics for which we are seeking panel input. 

            The Birmingham hip resurfacing system, or BHR, is a first of a kind device in the United States.  That is the first total hip system with a resurfacing femoral component and metal-on-metal articulating surfaces.

            The PMA is supported by clinical data essentially from one source, the surgical experience of Dr. Derek McMinn, who implanted devices primarily at the Birmingham Neffield (phonetic) Hospital, City of Birmingham, United Kingdom.

            The PMA includes safety and effectiveness data from an uncontrolled case series of all 2,385 procedures implanted with the DHR device from July 1997 through May 2004.

            FDA requests expert clinical opinion on the following topics:  the way in which the safety and effectiveness data were collected; the results of the study; and the applicability of data collected outside of the United States by a single investigator to the target U.S. population, U.S. practice of medicine, and U.S. orthopedic surgeon population. 

            This slide includes the outline for the rest of FDA's presentation.  It is our goal to summarize the information in the PMA to help the panel address FDA's questions.  I will briefly present the device description and preclinical testing information.  I will then discuss the sponsor's clinical data, focusing on the way in which the patients were selected to receive the BHR, the indications for use, the way in which the data was collected.

            Then I will present a proposed post approval study.  Dr. Chang Lao will then summarize the statistical information in the PMA and that will conclude FDA's presentation.

            After lunch, I will present seven FDA questions for panel discussion. 

            First, the device description.  I believe that the sponsor has adequately summarized the device description in their presentation, and I just had one clarification regarding the acetabular shells.  There are three styles of acetabular shell.  One is the standard cup.  The other is the dysplasia cup, and a third being the bridging cup, and the dysplasia and bridging cups are both for dysplasia indications with the bridging cup just being slightly thicker than the dysplasia cup. 

            I believe the rest of the information was adequately covered by the sponsor.  The sponsor provided preclinical information that included the evaluations listed on this slide and others.  Some of this information was summarized by the sponsor in their presentation, and FDA believes the sponsor has adequately addressed the preclinical issues for the BHR device.

            Now I'll discuss the way in which the patients were selected for this study.  Please note that one of the FDA questions will ask for your comments on the sponsor's device labeling.

            Generally, prospective clinical investigations predefined the study population with specific inclusion and exclusion criteria.  This theoretically allows the study results to be generalized to that diagnostic group.  In case series studies it is more difficult to generalize the results to a defined population because the patients were not enrolled for predefined conditions.

            This is the case for the clinical data provided in this PMA submission.  The clinical data were derived primarily from the surgical experience of a single surgeon.  This surgeon did not predefine a set of diagnostic indications for the device, but instead provided a list of the diagnostic indications for the patients implanted with the device.

            During the same time period Dr. McMinn implanted the BHR devices, he also had patients who either had no surgery or conventional total hip replacement.  However, a complete review of these patients was not presented in the PMA.  With this information, it might have been possible to retrospectively determine what criteria, if any, were used to select candidates for the BHR.

            As an alternative and in order to retrospectively develop the indications for use in physician labeling, the sponsor provided a list of the factors that contributed to Dr. McMinn's decision to perform a total hip replacement, or a THR, in certain patients rather than the BHR hip resurfacing procedure. 

            These factors included advanced age.  Patients of an advanced age, especially those with low activity levels, were typically candidates for THR rather than BHR.  Only 8.1 percent of the 2,385 cases were greater than 65 years of age.  In these cases the BHR were selected despite advanced age if the patients had high activity levels and had good bone stock of the femoral head.

            Low activity level.  Patients with a low activity level were considered at lowered risk for future revision and, therefore, good candidates for THR and not BHR.  Low activity level was characterized by no participation in sports activities, no heavy work required by job, a sedentary or retired lifestyle, or co-morbidities that precluded a high activity level, such as severe arthritis in other joints or severe VFR disease.

            Poor bone stock.  Patients with poor bone stock were selected for THR rather than BHR because they were considered at risk for femoral neck fracture or femoral head collapse.  With a resurfacing procedure poor bone stock was retrospectively defined and is a contraindication for the BHR.

            The sponsor stated that Dr. McMinn's preoperative evaluation was typically sufficient to screen candidates for BHR versus THR, and that only in rare instances would the planned surgical procedure be revised intraoperatively.  Because of the potential for a change in the preop plan, patients were consented for both a BHR and THR procedure.

            Based upon the population studied and the factors just mentioned and also an analysis of the BHR revisions, which included femoral neck fracture, femoral head collapse, dislocation, AVN, and infection, the sponsor proposed the following indications for use for the device.

            The BHR system is intended for patients requiring primary hip resurfacing due to noninflammatory arthritis and inflammatory arthritis, such as rheumatoid arthritis.  The BHR hip resurfacing arthroplasty is intended for joint replacement in patients who are at risk of requiring future ipsilateral hip joint revision.

            While it's impossible to predict if a patient will require more than one joint replacement, several factors are known to increase risk of revision surgery, including age less than 55 years at index surgery and/or high physical activity level postoperative.

            These are the contraindications for the device, and I believe the sponsor has covered these, focus particularly on one bullet item, which was regarding the inadequate bone stock, which includes severe osteopenia, osteonecrosis or avascular necrosis, with greater than 50 percent involvement of the femoral head and multiple cysts on the femoral head greater than one centimeter.

            Also, females of childbearing age due to unknown effect of a fetus on the metal ion release.

            In addition to the factors described above, the sponsor also considered a review of 50 BHR femoral neck fractures reported by Schimmon and Beck (phonetic) in the development of the labeling.  In this publication, the authors reported a review of 3,497 BHR cases performed in Australia by 89 surgeons.  There were 50 femoral neck fractures in the series, or 1.46 percent, which the authors attribute to osteoporosis and difficulties in implantation of the head and the cup leading to notching of the superior femoral neck, varus placement of the device by more than five degrees, difficulty in interoperative alignment, impaction of the femoral component, and poor exposure.

            Based upon these findings, the sponsor added the following warnings and precautions to the labeling.  "Warning:  avoid notching the femoral neck as this may lead to femoral neck fracture.  Avoid placing the femoral component in varus.  Varus placement of the femoral component has been associated with femoral neck fracture," and the following precaution.  "Improper selection, placement, positioning, and fixation of the implant component may result in early implant failure."

            The objective of this PMA is to demonstrate the safety and effectiveness of the BHR system.  The safety assessments included data on revisions, adverse events, and a metal ion literature review.

            Effectiveness assessments included survivorship, radiographic data, pain and function data as evaluated by the Oswestry modified Harris Hip Score and patient satisfaction data.

            Before I go into the description of the results, I quickly wanted to describe for you and paraphrase what valid scientific evidence is.  This information was provided to the panel members as a part of their training, and I just wanted to paraphrase some particular parts of what the FDA considers to be valid scientific evidence.  This is just being added to my talk at this particular time.

            Again, you can read the entire statement of what valid scientific evidence is in your packet, but I just wanted to summarize these parts quickly before I present the clinical data.

            Valid scientific evidence is evidence from well controlled investigations, partially controlled studies, studies and objective trials without match controls, well documented case histories conducted by qualified experts, and reports of significant human experience with a marketed device.  That is what valid and scientific evidence is.

            Valid and scientific evidence is not isolated case reports, random experience, reports lacking sufficient details to permit scientific evaluation, and unsubstantiated opinions are not recorded as valid scientific evidence.

            Again, you have the entire definition in your packet, and it is a question for the panel today on whether or not you believe this to be valid scientific evidence.

            I will now go into a description of the study.  As the sponsor described, there were 2,385 BHR procedures, and they were divided into the following three main cohorts.  Again, an X-ray cohort which is the first 124 BHR cases implanted in 1997; the Oswestry cohort, which was the next 1,502 cases; and the McMinn cohort, which was the next 759 cases.

            Note that there were five cases in the McMinn cohort whose implantations were performed prior to April 2002.  These cases should have been part of the Oswestry cohort, but for unknown reasons were not.  Therefore, unlike the majority of the McMinn cohort, some of these five cases have longer term follow-up.

            Where there were common data elements collected in these three cohorts, the sponsor pooled this information into two combined cohorts, which included what we're calling the overall McMinn cohort or the combination of the X-ray, the Oswestry, and the McMinn cohorts, as well as the X-ray-Oswestry combined cohort.

            The overall McMinn cohort contributed to the assessment of safety, including adverse events and revisions.  The X-ray cohort contributed to the assessment of radiographic effectiveness.  Radiographic evaluations were not provided for the 1,502 procedures in the Oswestry cohort or the 759 procedures in the McMinn cohort.

            The X-ray and Oswestry combined cohort contributed to the assessment of survivorship and patient satisfaction, and the 1,111 unilateral procedures in this combined cohort contributed to the assessment of pain and function effectiveness data as evaluated by the Oswestry modified Harris Hip or OSHIP Score.

            Note that the pain and function data for the 759 procedures in the McMinn cohort were collected using the Oxford Hip Score evaluation method and not the OSHIP score.  The sponsor explained that because these data were not tracked by the Oswestry Outcomes Center, but by the National Health Services Center, the sponsor did not have access to the Oxford Hip Score data.

            The main data sources were just presented, but the sponsor also included additional, less complete data on 3,374 BHR cases performed by 140 surgeons worldwide other than Dr. McMinn.  The follow-up for these cases also was contracted with the Oswestry Outcomes Center and includes primarily the same data as that provided for the X-ray and Oswestry combined cohort.  The Oswestry Outcomes Center has provided Smith & Nephew  access to all available data for the BHR cases from its database.

            Although the sponsor considers the data from this additional cohort to be of some value, Smith & Nephew has no ability to independently verify any of the data provided to the Oswestry Outcomes Center by sites other than the McMinn Center and Dr. McMinn and has no ability to request additional follow-up or clarifications of any kind from non-McMinn patients or physicians.

            For these reasons, this data has some limitations and is not considered a primary data source for this PMA.

            Now I will summarize the way in which the safety data was collected in this study.  Please note that one of the FDA questions will ask for your comments on the reliability of these data collection methods.

            The safety data, including adverse events and revisions, were collected by the following three methods:  the Oswestry Outcomes Center using an annual patient completed mail-in questionnaire, the McMinn Center by recording the findings of postoperative patient visits, and recording information provided to Dr. McMinn by primary care physicians.

            Dr. McMinn's follow-up was described as follows.  Dr. McMinn performed regular evaluations which included history, physical examination, radiographs to assess implant status, and any necessary laboratory work in the preoperative and postoperative time periods according to standard practice, although the time points and evaluations were not according to standard protocol.

            All revision surgeries were performed by Dr. McMinn, except in one known case.  Therefore the revision status was directly known to Dr. McMinn.

            There were no predefined follow-up time windows, standardized clinical evaluations, adverse event report forms or standardized radiographic evaluations.

            The sponsor also provided the following information regarding the follow-up procedure or what I'm going to call the OOC.  The OOC collected safety data on revisions and adverse events, again using an annual patient completed mail-in questionnaire.  With the exception of eight cases who withdrew or did not agree to participate in this study, all other cases are not considered lost to follow-up since the OOC continues to make attempts to contact patients.

            Of the 180 cases missing, the last theoretical expected mail-in questionnaire follow-up, 84 are missing on at least two-year reevaluations, while 96 are only missing their last evaluation.  These 180 cases represent 11 percent of the Oswestry/X-ray combined cohort.

            The OOC identified several steps taken to regain contact if a patient does not respond to a request for information, including sending reminder letters, E-mail or phoning the patient, contacting the consulting surgeon by letter or telephone, using a national strategic tracing service database and Internet census information to determine the patient's whereabouts.

            If the patient is still not found, an additional request for information is sent to the patient's last known address via Royal Mail and E-mail until the tenth anniversary of the operation.

            Patients are not classified as lost to follow-up until all avenues have been exhausted.  The sponsor stated that they performed a 100 percent audit of all 2,385 procedures in the overall McMinn cohort and, therefore, believe that all reported adverse event information has been captured.

            In addition to the safety date collection methods outlined above, the sponsor provided a metal ion literature analysis.  Included in the sponsor's analysis was an unpublished report by Daniel Zee and McMinn.  The authors conducted four metal ion studies, and I believe the sponsor has adequately summarized those studies in their presentation.

            Now I will summarize the way in which the effectiveness data was collected.  Please note that one of the FDA questions will ask for your comments on the reliability of these data collection methods.

            The primary effectiveness measurement was survivorship for procedures with a minimum of two years postop.  Of these 1,626 procedures, data were available on 546 of the 601 BHR procedures eligible for five-year follow-up or 90.8 percent.

            The data for the survivorship study was collected using the same methods presented for the safety data that included the OOC and the McMinn Centers.  The PMA also contained the results of an independent radiographic review of the X-ray cohort, which was the first 124 procedures performed.  The sponsor adequately summarized the radiographic information and, again, comparison to baseline films were made for each of the 108 cases that were out to five years in the postoperative time period, and the baseline films were, again, usually within three months, but eight of those 108 procedures the baseline films were evaluated between 110 and 860 days postoperative.

            The radiographs were independently evaluated by Dr. Nick Evans.  A prospective protocol was used to assess the radiographs.  The five-year AP and lateral view radiographs were compared to baseline radiographs for migration, acetabular orientation, radiolucency, heterotopic ossification, and other radiographic findings.

            A radiographic success was defined as having all of the following:  absence of radiolucencies or radiolucency in any one or two zones, component migration less than or equal to two millimeters, and a change in acetabular angle less than five degrees.

            A radiographic failure was defined as the presence of incomplete or complete radiolucencies or radiolucency in all zones and migration of the component greater than two millimeters or a change in the acetabular orientation less than or equal to five degrees.

            Pain function and movement data were collected by the OOC using an annual patient completed, mail-in questionnaire.  The patient responses to the questions were used to generate the Oswestry modified Harris Hip or OSHIP score.  I believe that this is the first time the FDA has evaluated the OSHIP scoring system in a marketing application.  Therefore, the FDA has asked the sponsor on how the OSHA data were collected, how the OSHIP scoring system was developed, and asked for a justification for its use.

            The OSHIP questionnaire allows patient assessments without direct physician evaluation.  No other sources of pain and function information were used to support this PMA.

            The sponsor summarized the OOC standard operating procedure for data input and clarification of the patient administered OSHIP questionnaires.  Any questionnaire with missing, unclear, or conflicting information was returned to the patient with specific instructions for completing the form.  The preferred method was by mail.  However, E-mail and telephone were also used to complete these questionnaires.

            If the data were not collected, the score from any missing item was assumed to be the lowest possible, which was typically zero.

            Now I'll discuss how the OSHA patient questionnaire was developed.  It was developed by Professor James Richardson with the following premises.  He believed that a long-term evaluation following hip replacement is essential.  Follow-up must be regular and large samples are necessary.  Long-term and large sample follow-up is difficult to obtain when using a score that requires a surgeon or radiologist assessment.

            Physician administered surveys are susceptible to bias which may inflate the final score and may not truly represent the patient's own feelings, and questionnaires needed to be simple and relatively short to make long-term and large scale collection of data more efficient.

            Building on these premises, Professor Richardson developed the OSHIP scoring system by combining elements of both the Harris Hip Score and Merle Dobbinier (phonetic) scores.  As presented, the OSHIP produces an overall index score similar to that of the Harris Hip Score between zero and 100.  The OSHIP score is made up of three domains of pain, function, and hip movement.  The main difference between the OSHIP questionnaire and the Harris Hip Score is that the OSHIP allows patient assessments without direct physician evaluation.

            In addition, the OSHIP questionnaire does not include the three HHS questions regarding a physician assessment of range of motion, which is five points, absence of deformity, which is four points, and the patient's ability to put on shoes and put on socks and tie shoes, which is four points, but substitutes a movement question which is 13 points that is intended for the patient to estimate their ability to flex their hip.

            There are additional differences between the OSHIP and Harris Hip Scores in the phrasing of some of the questions and the point values that correspond to some of the answers.

            Again, FDA requested that the sponsor justify the use of the OSHIP scoring system, and also the validity of patient self-administered questionnaires, and the sponsor summarized several literature references which I'm now going to discuss in some detail.

            While a paper by Ragab and co-workers reported a lack of correlation between patient self-assessment of pain and function and physician assessment of pain and function with a correlation of .467, several other researchers have reported the opposite, a very close correlation between patient self-assessment and physician assessment. 

            Research by Mohammed and co-workers demonstrated that patients are able to accurately respond to Harris Hip Score questions regarding pain and function with little difficulty, and there is excellent correlation between overall HHS pain and function scores reported by the patient and the overall HHS pain and function scores reported by physicians with a correlation of .99.

            In this study, Mohammed also reported that the case statistic, which is a measure of the reproducibility between repeated assessments of the same categorical variable ranged between .79 and one for each item of the HHS and, according to the paper, indicated excellent reproducibility.

            Note that both the Ragab and Mohammed studies did not include patient or physician evaluations of range of motion or deformity.  These questions were eliminated from both the patient and the physician assessments.

            Furthermore, McGrory and co-workers found that a brief follow-up phone call similar to the OOC follow-up procedure was effective in capturing missing data or clarifying multiple or contradictory responses from male patient self-assessment questionnaires.

            In addition, the sponsor provided a literature article by Barnes and co-workers which evaluated the reliability and validity of the OSHIP score as documented in their research paper.  In Dr. Barnes' study, a group of 61 patients completed the OSHIP questionnaire.  They were then sent a second copy to be completed two weeks later and returned by mail.  The results of these two sets of surveys were compared to look for reproducibility.  When comparing the responses the total interclass correlation coefficient was .93.

            Dr. Barnes' study also included a separate group of 28 consecutive patients who were given both the patient administered OSHIP and a physiotherapist administered Harris Hip Score.  The correlation between the patient's overall OSHIP score and Harris Hip Score was .91.  The correlation between the individual corresponding domains ranged from .6 to .89, with the lowest correlation being for the patient's assessment of limp.

            FDA requested additional correlations be provided that were not included in Dr. Barnes' study.  In addition, FDA performed a linear regression analysis to predict HHS score from OSHIP score for the 28 subjects.  The linear regression analysis is summarized in your executive statistical summary, and the calculated R squared is approximately .83.

            A review of the raw data for the Barnes study also revealed the following.  The average OSHIP score was lower than the Harris Hip Score, 62 and 67, respectively.  Less subjects had an OSHIP score greater than 80 and more subjects had an OSHIP score less than 70 as compared to the Harris Hip Score.

            There were 14 pairs of data where the OSHIP and HHS scores differed by more than five points.  Of the 14 pairs, the HHS score was higher in 12 cases, while the OSHIP was higher in only two cases.

            Additional information regarding the correlation and regression analysis and the limitations of the Barnes study will be summarized by our statistician, Dr. Chang Lao.

            Like the Barnes study, Ragab also reported a relative lack of correlation between patient assessments of limp and the physician assessment of limp, which he believed was due to a physician's tendency not to report limps that occurred only after long walks or during weather change, while patients were likely to report such limps.

            However, unlike the Barnes study in which the OSHIP and HHS item regarding pain had a correlation of .83, Ragab found that when patients reported significant pain, they were often attributing the pain to their hips when the pain in most cases was not truly hip related.  The authors reported that the physician was better able to distinguish true hip pain from pain coming from other sources, for example, secondary to trochanteric bursitis, lumbar spondylosis, and arthrosis of the contralateral hip.

            An additional finding by McGrory and co-workers was that questions about whether patients could cut their toenails and put on shoes and socks correlated significantly with Harris Hip Score range of motion, with correlations of .57 and .53, respectively.  The authors concluded that responses to these two questions could, therefore, be an estimate of the weighted Harris Hip Score range of motion.

            Finally, Johnson and Schmidt also reported that there is a distinct relationship between hip flexion and the question about shoe tying.

            In the final comment of Dr. Barnes and co-workers' study, the authors stated that the Oswestry hip score is not intended to replace clinical evaluations at the critical phases following hip surgery, that is, at one year, five years, and ten years.  However, it can be a useful tool along with  X-rays to replace unnecessarily yearly follow-up following hip surgery.

            The sponsor used the reference studies by Mohammed, McGrory and Barnes to justify the use of patient self-administered questionnaires to adequately report pain and function data.

            Furthermore, the sponsor asserted that the close correlation of the overall OSHIP and Harris Hip Scores reported by Barnes and the tendency of the OSHIP scores to be somewhat lower relative to the Harris Hip Scores suggest that the OSHIP is a very close, although conservative, estimate of the Harris Hip Score.

            Finally, for the purpose of the BHR study, an additional question about patient satisfaction was appended to the end of the OSHIP assessment questionnaire.

            Now for the results.  Procedures in the overall McMinn cohort were 70 percent men, 29 percent women, ages ranging from 13 to 86 years, with an average of 53 years.  Ninety-one, point, nine percent of the procedures were less than or equal to 65 years of age.  The primary diagnosis was osteoarthritis in 75 percent, dysplasia in 15.8 percent, AVN in four percent, inflammatory arthritis in 2.4 percent, and other diagnostic indications in 2.7 percent.

            All femoral head sizes were used in the overall McMinn cohort and almost all patients received either the standard cup or the dysplasia cup styles.

            The follow-up rates for the X-ray/Oswestry combined cohort upon which most of the effectiveness analyses were performed are shown in this table.  The follow-up rate at baseline was 80.6 percent and 90.8 percent at five years.  There were 546 procedures evaluated of the 601 expected at five years postop.

            Now I'll present the safety data, and please note that one of the FDA questions will ask for your comments on whether or not the data contained in this PMA provide a reasonable assurance of safety.

            There were a total of 27 revisions, which included ten revisions due to femoral neck fracture, six femoral head collapse, one dislocation, two AVN, and eight were revised due to infection.

            Factors contributing to femoral neck fracture and head collapse included osteopenia, poor bone quality as evidenced by cysts in the femoral head and acetabulum, SLE, severe rheumatoid arthritis, infection leading to bone death, AVN, femoral cysts, and a malpositioned component.

            There were a total of 2,912 adverse events in 1,669 of the 2,385 procedures for a rate of 70 percent.  I believe that the sponsor has adequately summarized the adverse events in this study, except I want to comment on that it was reported that there were 589 procedures with a wound exudate for a rate of 25 percent.  The sponsor stated that this was probably due to a difference in reporting requirements.

            There were 20 patient deaths in 26 procedures which the sponsor indicated were unrelated to the BHR device, and again, narratives were provided in the PMA for each of the patient deaths.

            The sponsor also provided a metal ion literature analysis.  The reference is in the PMA reported at serum and urinary metal ion concentration in patients with total hip replacements, with metallic components in general, and metal-metal articulating implants, in particular, increase in the post operative period.  However, there does not appear to be conclusive evidence that elevated cobalt and chromium levels have detrimental effects in the total hip arthroplasty patients.

            Now I'll present the effectiveness data.  Please note that one of the FDA questions will ask for your comments on whether or not the data contained in this PMA provide a reasonable assurance of effectiveness.

            The 1,626 procedures in the X-ray/Oswestry cohort contributed to the assessment of survivorship.  The estimated percent of procedures remaining free from revision at five years after the BHR procedure was 98.4 percent with a 95 percent confidence interval, having a lower bound of 97.3 percent and an upper bound of 99.5 percent.

            The only marginally statistically significant difference in five-year survival probability was between patients with osteoarthritis at 98.8 percent and avascular necrosis at 92.1 percent as a primary diagnostic indication.  Again, we evaluated survivorship for the X-ray/Oswestry combined cohort, as well as the McMinn patients, which included the additional McMinn patients.

            And Dr. Chang Lao will present that information as well in FDA's analysis.

            Regarding the radiographic data, three of the 108 procedures in the X-ray cohort for whom radiographs were available were radiographic failures at five years or 2.8 percent.  One failure was due to a femoral radiolucency, one due to an astabular radiolucency, and one due to both an astabular radiolucency and a change in the astabular orientation  of greater than five degrees.

            For the pain and function data, the OSHIP score was used to evaluate the 1,111 unilateral procedures in the X-ray/Oswestry cohort.  Mean OSHIP scores improve from 60 to 94.8 at five years.  At first operative years two, three, four, and five the percentage of cases with good or excellent scores, that is, greater than 80 points, was 96.9 percent, 95.8  percent, 95.2 percent, and 92.8 percent, respectively.

            For the patient satisfaction data at five years, 99.5 percent of the procedures were pleased or very pleased with the operation.

            The sponsor submitted two literature controls.  The D'Antonio reference included data on 514 Homedica Ostionics, AVS, and Trident ceramic-on-ceramic total hip replacement procedures, and the Garino reference included data on 333 light medical ceramic transcend, ceramic-ceramic total hip replacement procedures.

            In our review of these references, they appear to have significant differences as compared to the data provided for the BHR device in this PMA, including different evaluations.  The OSHIP score was used for the DHR and the HHS score was used for the literature.  The length of follow-up, 18 to 36 months and two to four years for the controls with two to five years for the BHR study.

            Mean baseline payment function scores.  The mean baseline score was 60 in the OSHIP scoring system for the BHR study and 44 for the Harris Hip Score-Greenough study, and it was not reported for the D'Antonio Study.  There were also differences in the indications for use, including differences in the rate of dysplasia and AVN diagnostic indications.

            Additional information regarding the literature controls was summarized by the sponsor and was contained in the panel packs.

            Now, I'll discuss the applicability of the data collected outside the United States by a single investigator to the target U.S. population, practice of medicine, and U.S. orthopedic surgeon population.

            Please note that one of the FDA questions will ask for your comments on this information.  Again, the clinical data were derived from a foreign clinical study conducted by a single investigator, Dr. McMinn, at the Birmingham Neffield, but also six patients at the Little Aston Hospitals in the United Kingdom.

            There was no racial or ethnic data, origin data, for the patients presented in the PMA.  However, the sponsor provided the racial and ethnic distributions of the general U.S. and general U.K. populations and believes that they're similar.

            There were noted difference in the higher percentage of people with African descent and other races in the general U.S. population as compared to the general U.K. population.

            The sponsor also provided a comparison of the demographics and diagnostic indications for the BHR study and the literature reference by D'Antonio and co-workers, again, for the Homedica Ostionics ceramic-ceramic device.

            There are noted differences in the higher percentage of men, higher percentage of procedures with dysplasia, with 15.8 percent of the BHR study and none reported for the ceramic-ceramic study; a higher percentage of inflammatory diagnostic indications; 2.4 percent for the BHR study and none for the ceramic-ceramic study; a lower percentage of procedures with AVN; 4.1 percent for the BHR study and 16 percent for the ceramic-ceramic study and post traumatic arthritis, none for the BHR study, and four percent for the ceramic-ceramic study.

            The sponsor stated that the orthopedic practice of medicine utilized by Dr. McMinn is the same as the standard of orthopedic practice in the United States.  The sponsor described Dr. McMinn's practice of medicine as follows.  The operating room has laminar air flow and body exhaust suits.  Dr. McMinn used a posterior surgical approach. Antibiotic prophylaxis was used interoperatively and for 24 hours postoperatively.

            DVT prophylaxis using IV heparin interoperatively and compression stockings and low dose aspirin was used postoperatively for six weeks.  Interoperative venting of the femoral shaft was used to prevent fat embolar (phonetic).

            Early ambulation, including full weight bearing with a walker on postop day one, hospital discharge at postop day six.  After six weeks postop patients begin range of motion exercises.  Postop activity for the first year is nonimpact or low impact and avoidance of high impact exercises.

            Finally, the FDA advised the sponsor that the PMA may be subject to conditions of approval, including a post approval study to evaluate the long-term safety and effectiveness of the device.  In response to FDA's advisory, the sponsor included a post approval study protocol, which included a nonrandomized, prospective, longitudinal, unblinded, multi-center trial to evaluate the long-term safety and effectiveness of the device.  It included an enrollment of 150 patients at 15 sites.  Inclusion and exclusion criteria were defined.  Clinical and radiographic follow-up for five years; long-term follow-up assessment using a self-administered mail-in patient questionnaire for six to ten years, the questionnaire which would include three yes and no questions regarding patient satisfaction, whether they have had a revision or a replacement, and expectation upon removal in the near future.

            An analysis of the explanted device components and clinical and radiographic success and failure criteria were also defined.  Please note that one of the FDA questions will ask for your comments on the proposed post approval study,

            Now Dr. Chang Lao will briefly discuss the statistical information in the PMA.

            DR. LAO:  Good morning, panel members and audience here.  I'm Chang Lao, Division of Biostatistics.

            Today I'm going to present the first slide, summary of the patient accountability.    The patient accountability, the previous speaker has already summarized in greater detail, and this slide is by a Swiss study cohort and by unilateral and bilateral hip implant, total, 2,385 hips, and OSHIP score is only available for the X-ray and Oswestry cohort, not available in the McMinn cohort over a three-year time period. 

            My outline today basically constitutes different parts.  The first part are the basic statistical issues for the PMA.  The second part is the summarized PMA statistical analysis, and the PMA statistical analysis included, first of all, survivorship, revision free analysis, and Oswestry/modified Harris Hip Score, and the correlation between HHS  and OSHIP score, and it gives some summary.

            Basic statistical issue for the BHR device is the first issue is the unique investigator, Dr. McMinn.  No multi-center trial.  So the question is how to generalize to our doctors, to our other centers, how to carry out training.  That's a question based on the data from one doctor.

            The second issue is no control group, BHR only.  It is nonrandomized.  It is combined retrospective and prospective registry data.

            So the question is how to interpret results from this study to the general type of patient population, and sample size justification is neither prespecified hypothesis testing nor based on the confidence interval approach.

            And the sponsor's post hoc justification based on constant revision rate over five year.  Exponential distribution with some desired power, but probably if statistical justification cannot test that post hoc justification, usually in randomized trials samples should be prespecified beginning the study, in the product study design stage.

            Continued basic statistical issues is on nonrandom patient selection.  So the question is how generalized to a well defined group, subgroup of patients.

            Again, the sponsor did a post hoc justification, complete demographical in terms of age, gender, diagnosis.  Comparability of three study cohorts:  X-ray, Oswestry and McMinn.

            And another issue is unclear correlation between the OSHIP and the HHS scores, and no sample size justification between the subjects.

            Nonrandom sample, 28 paired data, and there's no masking or order randomization.  The masking and order randomization of the OSHIP HHS data are very important because if I'm the patient, when I rate my OSHIP score I should not be aware of the facility's rating.  If I'm  the physical therapist I should not be aware of the patient's self-evaluation.  Otherwise I would be very easy to introduce a bias.

            The second part is the PMA statistical analysis, a summary of the PMA data.  The first study here is the survivorship, five year, and separated by three study cohorts.  As you can see, all above 98 percent at five year for three study cohorts.

            At the bottom of slide and the last line is the number of patients at the beginning of the study and the number of patients sine then and the number of admissions by year.  So at the end of five years, actually this is based on real data, not based on accountability from the number expect due or theoretically due or number expect.

            If based on previous speakers, the number of patient accountability, the percentage of patient accountability, five years above 90 percent.  But if based on real data, only about 21 percent complete five-year study of the total, 2,385 hips at the beginning of study.

            So that's the difference between the real data and the hypothetical data.

            Figure 2 is the number of revisions, 27 total revisions, of the X-ray cohort is beyond five years.  So about 25 divisions by cohort and the reason.  And as you can see, total is 25 division out of 2,385 hips, about a 1.1 percent is relatively low.  So sample size, you know, is justification to test the comparability of the Swiss study cohort.  It really has not enough power to detect the difference because not enough for the divisions, 25 or 27.

            So one other question was how can you combine statistically justify the pooling of Swiss study cohort, the X-ray, Oswestry, and the McMinn cohorts.  There's basically three different tests.  One is a log rank test, which the description is a log rank, is actually the compared of  observed, expect number of revisions over time period, and you have optimum power if the Swiss survival curve, a parallel issue added.

            But you can see in the Figure 1 some crossover over time, not complete parallel. 

            Wilcoxon rank test is some more weight to earlier follow-up time when more patients at the risk of revision.

            And lastly, the Cox proportion hazard model, which also requires parallelism of survival curves.  The hazard rate, so-called hazard is a rate which is a lot of plausibility, not a proportion, an which is average the number of event, p as a revision per time interval.  This says per year they have 25 total revisions divided by five.  The average of five revisions per year, that's a hazard.

            But the Cox proportional hazard, you need a parallelism of the survivor curve, and in the Figure 1 and the three curves are not completely parallel with the statistics over the years.

            The results by statistical tests, the premise of doing this way statistically justifies pool of the study cohort into a overall conclusion, but at the bottom footnote there, as a statistical justification, there's only one of the requirements and which is not sufficient to justify putting some other clinical technique required here to adjust the pooling of data because those we study, the cohorts come from a different time period and maybe different data correction scheme.  So you need some kind of other clinical input.

            Anyway, the statistics here, the p value now is significant.  Even if the parallelism is assumption of violative, but because consistency of three different tests, so I would say that there's no statistical difference unless we start survivor curve of the three study cohorts.

            This is the total number of complications by combined cohort and combined unilateral and bilateral hips over time, and you can see that year one the AVN, avascular necrosis has a very large number, don't automatically decrease.  So the year one, 35, the total AVN 35 complications, and because not every complication will result in revision.  So the total number completing is much larger than total number of revisions, 25 revisions. 

            There's a second effective endpoint by year based on available data, combined cohort, and on unilateral hip only.  We exclude the bilateral hip is hard to evaluate OSHIP score.

            As you can see at the baseline, 1,111 hips, total capital N, and the first column, second line, the small n, total number observed, 892.  So the ratio of the small n and the capital N is 80 percent, and if you look at the common ration, small n over capital N, you can see that year one, year, two you have about 25 percent of the hips not observed.  They're missing here.

            Then the missing data decreased at the year five, 91 percent of hips observed.  What that means here is that the person missing one year, two, they reappear in the later year in the missed co-state (phonetic) and not a dropout.  We call intermittent missing, intermittent missing, the missing to come back again.

            So we should be very careful here to interpret the OSHIP score over time, and the assumption here is that you look at the name here, main OSHIP score 60 at baseline, dramatic improvement in the year five, right?  About 95 percent.    Standard deviation, standard error and a 95  confidence interval for the main OSHIP score. 

            Ninety-five confidence interval very, very narrow because sample size is quite large.  That's good, and the subject here is we assume OSHIP account between the patient date missed during the year, one year and other year to approve similar as those patients who have complete data.

            This assumption, unfortunately we can miss at the random or miss completely random, cannot test statistically, and so this assumption we should be very careful.  Hopefully the missing data, year, one year, two, other year and another due to device or not due to complication but to something else.  That's why some people have to be very careful.

            So to put that PBS table into this slide there, which is quite a medical improvement after year one, two to five year, but again, the missing data should be very careful.

            The final issue is the correlation between the gold standard HHS OSHIP score.  The assessment in general probably is a little  randomized, only 28 paired data, no synthesized justification which is not based on hyper test or confidence interval approach, and the OSHIP of a patient's safety evaluation, for instance, at the HHS by fissile therapist from the same clinic.

            So the question here is go masking the order randomization and no track.  Timing I'm not sure what's the time distance between three different measurements.  If you measure in the relevant time period, maybe more relevant than those times, they fly apart because OSHIP score can change.

            So OSHIP has three major areas, four major areas.  Those that score zero are the worst two, 100 the best, and OSHIP pair function hip movement.

            There's a summary of the correlation between HHS and OSHIP for each individual component, and the plant function and the total, plant function, the blue color here because they give them all weight and are more important than the other components.

            The total score, as John Goode said, is about 90.91 correlation, 95 confidence interval.  It's inside parenthesis there.  So movement is a correlation motivator and otherwise the other I would say correlation about .8 is a moderating pretty good.

            But again, this correlation is subject to potential bias because of the masking and randomization.

            How can predict HHS from OSHIP?  So we did the linear regression analysis, and you can see here the straight line is the linear equation, the intercept and slope at higher air range (phonetic) because the fitting is better, and better than the intermediate range.

            And R-square, .83 NMEs, 83 percent of reliability of the data above the mean were explained by this question mark, .83.  The square root of .83 is .9.  That's the correlation for total score, as you can see from the previous table.

            Now, we look at many ways to compare the OSHIP and HHS because the correlation is just one necessary condition, but a lot of sufficient condition for the prediction.  So second analysis of complete mean score, as your mean score in six or seven versus OSHIP of 62, a difference of five or eight points, and that's the main difference in 28 pair.   Still it is a  95 confidence interval, 1.36 to 8.8, which doesn't include  zero.  So it would means HHS statistically higher than OSHIP mean score.  P value is  .01.  That completes the mean difference on the total score.

            Another comparison that can be done is to compare the binary  account to be sure that the code of total OSHIP score based on excellent plus good.  It means scores 88 and above.  As you can see on this two-by-two table, a total agreement 80 or larger is eight plus 15, 23 pairs, to eight pairs, which is about 82 percent agreement.

            And a probability, HHS larger or equal to 80 is 12 over 28 versus OSHIP, which is nine over 28.  So OSHIP-HHS is higher proportion, 33 percent versus OSHIP, 32 percent, and you will compare, say this proportion differs a significant difference, which unfortunately cannot be tested because it has enough problems.  In order to test the difference you will need a large number of prespecified study designs.  You have enough number for these codes and payroll, and here is only four and one, five discordant pair, which are informative pair and 23 agreement pair that they don['t give you much information in testing the hypothesis up to proportions.

            So summary.  Basic statistical issues, the one investigator cannot statistically justify generalization of result to other physicians' centers.  Post hoc justification for sample size, patient selection, inclusion/exclusion criteria cannot statistically write (phonetic) post hoc justification.  No prespecified masking, no order randomization, no sample size determination for correlation analysis.

            Some incomplete/missing OSHIP data.    That 25 percent missing on the year one, two and nine percent year five.  Assumption is similar clinical results between complete and incomplete patients, as I say, missing at random, but again, we cannot statistically test that assumption.

            And the life study is statistical conclusion based on available data.  The mean age means about 53 years old.  The range is 30 to 86, 92 percent above equal to 65.  Twenty-seven revision two beyond five years for the X-ray cohort.

            Survival analysis of the Figure 1 of everyone is above 98 percent, five year, and the mean OSHIP score at five-year, 94.8, which equates to 100, and the percent excellent plus good, 92.8 percent at five-year, and the correlation, .91 total score.

            So thank you.  This is the end of my talk.  Thank you.

            MR. GOODE:  This concludes FDA's presentation in the morning, and  I'd like to acknowledge our team who worked on this, including Dr. Chang Lao, Patty Jahnes, Tracy Bourke, Maryann Wollerton, Mike Courtney, and from OSB Ronald Kaczmarek.

            Thank you very much.

            PANEL CHAIRPERSON NAIDU:  Thank you. 

            I'd like to thank the FDA speakers for their presentations.

            Does anybody on the panel have any questions for the FDA now?  You may also ask the FDA questions this afternoon. 

            (No response.)

            PANEL CHAIRPERSON NAIDU:  Seeing none, let's break for lunch.  We'll reconvene at 1:00 p.m.

            Thank you.

            (Whereupon, at 12:06 p.m., the meeting was recessed for lunch, to reconvene at 1:00 p.m., the same day.)



















                 AFTERNOON SESSION

                                       (1:03 p.m.)

            PANEL CHAIRPERSON NAIDU:  Dr. Jan Mabrey from Dallas will be opening this part of the meeting.  He will help us focus our deliberations.

            The panel will then deliberate on the information in the PMA and on the information the sponsor and FDA has presented this morning.

            The panel can ask the sponsor and FDA questions at any time.  After a general discussion, the panel will address the FDA questions.  Then there will be a second open public hearing and FDA and sponsor summations.

            Then the panel will conclude the deliberations by voting on the recommendation to the  FDA concerning this PMA.

            Dr. Mabrey.

            DR. MABREY:  Thank you.

            Just to remind the panel, this is the disease that we're looking at today that's confined osteoarthritis of the femoral head.  This particular image comes from Peter Bullough's excellent Atlas of Orthopedic Pathology, who was also one of my professors at Cornell.

            The purpose of my presentation this afternoon is to acquaint the panel with the clinical aspects of the device under consideration and to provide a perspective on its place in the orthopedic armamentarium.

            Total hip versus hip replacement arthroplasty basically concentrates on two factors.  Number one is that the femoral neck is replaced and that there is no femoral stem.

            As you can see here, there's a substantial amount of bone left behind with the femoral replacement or the head replacement arthroplasty.

            The proposed advantages, therefore, are that this bone is conserved, that it reproduces anatomic hip mechanics.  There is greater stability as opposed to total hip replacement, and as you've heard earlier, easier revision to total hip arthroplasty.

            The concepts of hip resurfacing include conservation of bone, sparing of femoral neck, optimizing stress transferred to the neck, and enabling future revision.  It also includes the placement of a large femoral head with a relatively thin acetabular component that usually relies on press fit fixation.  It has a stable range of motion, and again, the purpose is to preserve normal hip biomechanics.

            Contraindications to use, again, as you've heard earlier today, absolute contraindication should include the elderly with osteoporotic bone, metal hypersensitivity, and impaired renal function.  Relative contraindications include inflammatory arthropathy, severe acetabular dysplasia, grossly abnormal geometry, and large areas of avascular necrosis.

            The evolution of intelligent design of the hip resurfacing arthroplasty began with the Smith-Petersen Mold, which actually started out as a glass device implanted by Smith-Petersen in 1928, and I'll go into the details of that in a moment.

            It then followed two paths of evolution.  First was the path to hip resurfacing in which polyethylene was the bearing surface, and here is a list of some of those individuals who participated in that path.

            The second path was the path to metal-on-metal articulation.  Again, a list of those participating in that path, with the result being a device similar to the one being presented today, resurfacing metal on metal.

            The Smith-Petersen Mold, the first Vitallium prosthesis was implanted in 1938.  As I said, he implanted a glass one in 1928.  The particular case you see here was implanted in 1948, and that radiograph has a 46-year follow-up.  The patient was functioning quite well and was symptom free, except for a slight limp on the left side.

            The McKee-Farrar came along and now we're following the path of metal-on-metal technology and implanted several of these devices.  This radiograph demonstrates 23-year follow-up after implantation and the authors suggest that there is approximately two millimeters of linear wear.

            At this point in time, sphericity, clearance, and surface roughness were not necessarily appreciated, and the success and longevity of individual components were probably more a result of chance than of design.

            Muller introduced the metal-on-metal total hip arthroplasty in 1987.  He implanted 18 surface replacements, 35 stem replacements, six of which were revised after functioning for up to 25 years.

            Getting back to hip resurfacing, Paltrinieri and Trentani came out with this device in 1971, which is a thin walled, all polyethylene acetabular cup, and as we're all aware from the literature, it was prone to wear and developing significant osteolysis.  The metal component was composed of stainless steel.

            Wagner introduced his device in 1974.  It was widely used in Europe, but again, the acetabular fitness here was only four millimeters.

            The Tharies hip, the total hip articular replacement using internally centric shells was introduced in 1975 and had a variable thickness of 3.5 to 5.5 millimeters, but again, remember that the bearing surface is all polyethylene.

            Dr. Harlan Amstutz designed the porous surface replacement, introduced it in 1983.  The femoral head was a titanium alloy with mesh.  The acetabulum was a titanium shell with a polyethylene liner, again, polyethylene in between.

            Finally, in 1988, the Metasul metal-on-metal total hip arthroplasty was introduced.   Larry Dorr was one of the early proponents of that device in this country and reported on 70 patients in the year 2000 using the cemented web or cup.  He demonstrated a 94 percent survival rate at seven years and demonstrated no osteolysis with that particular device.

            This is probably one of the earlier examples of metal-on-metal hip resurfacing introduced by Grigoris and Roberts, utilized hybrid fixation with an uncemented cup and a cemented head.  They also introduced improved instrumentation for preparation of the femoral head and for sizing.

            Right now this represents the world market for hip resurfacing with the Conserve Plus being introduced by Wright Medical Technologies in 1996, all the way up to the Icon hip resurfacing introduced in 2004.

            Four areas of interest that I'll discussion.  Number one is one area that's associated specifically with hip resurfacing, and that is femoral neck fracture.  Two and three are associated with metal-on-metal hip arthroplasty in general, and that includes acetabular fixation, as well as wear and metal ion concentration, and the fourth area of interest results from published U.S. studies of metal on metal hips resurfacing.

            Femoral neck fractures and hip resurfacing arthroplasty are a result of a demanding surgical technique.  They usually result from some type of femoral head defect or from an error in implantation.

            This is the result of one of those that developed several months after implantation of the device.

            The surgical technique itself is rather demanding.  One must maintain careful angle of implantation of the femoral head.  One must avoid notching the femoral neck and avoid impingement.

            With regards to implanting the acetabulum, it's not quite analogous to total hip arthroplasty as the surgical exposure is somewhat more challenging when the femoral neck and portions of the head remain in place.

            This demonstrates the close proximity of the femoral neck with the edge of the acetabular component, and if there are errors made an implantation impingement can occur. 

            In those cases of osteoarthritis resulting from femoral acetabular impingement, these areas may be compromised and may be prone to fracture.

            Femoral head cyst formation is a problem in both osteoarthritis, as well as osteonecrosis.  As this radiograph, again, from Dr. Bullough's Atlas demonstrates.

            Femoral head defects are often encountered in implantation of femoral head resurfacing devices.  This one demonstrates the appearance of the head after reaming for placement of the resurfacing implant.  At that point the defects are filled with bone from the acetabular reamings.

            One of our key concerns with regard to hip resurfacing arthroplasty studies which indicate that there is a difference in the stress distribution in hip resurfacing as opposed to total hips is you'll notice in the middle the total hip arthroplasty concentrates its load at the distal tip along the shaft and then distributes the rest of the load along the stem.

            However, with his resurfacing arthroplasty as this study from Kuhl and Balle demonstrates from this year, the stresses tend to concentrate right at the femoral neck, and this has several implications.

            Number one is stress shielding.  On the left is a radiograph from Lilikakis' report in Orthopedic Clinics of North America this year, demonstrating a hip resurfacing one month out.   If you look on the right, the same patient, the same resurfacing two years out, and there's a significant amount of thinning of that medial femoral cortex from stress shielding.

            The other problem in this area is that of acetabular fixation.  As with the other hip resurfacing arthroplasties, the fixation of the acetabular cup is necessarily dependent upon press fit fixation.  Most authors recommend under reaming by one millimeter and then protecting or at least controlling patient weight bearing for some time until there's osteointegration.

            So the type of osteointegration and the type of technique used to insure stability is very important in a consideration of hips resurfacing arthroplasty as a device in the United States.

            Another area of interest is that of metal on metal ion levels.  As you can see from this study, from Clarke, et al., published in the Journal of Bone and Joint Surgery in 2003, they matched 22 patients with metal-on-metal hip resurfacing arthroplasty, with 22 patients with 28 millimeter metal-on-metal total hip arthroplasties, both with age, weight, and length after surgery.  At a median of 16 months postop they found that the cobalt and chromium ion concentrations, the average concentrations for hip resurfacing arthroplasty were 38 and 53, respectively.  The average ion concentrations for total hip were 22 and 18, respectively.  Those differences were significant.

            The black bars up there represent the maximum values seen, and if you can see it, down at the bottom is a small, green bar.  That represents the upper limit of normal.

            Manufacturing of these devices is key and is probably responsible for their renewed popularity and their increased longevity at this point.  As I said before, earlier devices did not appreciate sphericity and clearance, and it was by chance that some of them lasted up to 25 years.

            In this study from Rieker, et al., they looked at radioclearance versus run-in wear.  The red star is outlining a 38 millimeter head with a 100 micron clearance between the ball and the cup, and you can see that the run-in wear is approximately ten.

            When you increase that  clearance now to almost 300, you basically increase the amount of run-in wear fivefold.  This is appropriate for larger devices.  Here's a 50 millimeter cup with approximately 140 microns of clearance, and again, if you increase the clearance in that same device, you also increase the amount of run-in wear, thus increasing the amount of metal debris.

            The U.S. experience with metal-on-metal hip replacement arthroplasty as a result of studies conducted by authors using IDE classified devices.  This is probably the most well known by Harlan Amstutz that included 400 hips in 355 patients, IRB approved.  Seventy-three percent of those patients were male.  Note that the average age is only 48 years, but the range was from 15 to 70. 

            This was a hybrid metal-on-metal hip resurfacing arthroplasty.  The Conserve Plus, there was no HA coating on the back side of the acetabulum, and this represents three and a half years' average follow-up.

            Also, 43 patients in that series had dysplasia, representing 11 percent of the cases.  Three-fourths of those were Crowe Type 1.

            Survivorship.  Dr. Amstutz is always highly critical and analytical of his own results, and he divided survivorship into two areas.  The overall survivorship of this device at four years was 94.4 percent.  However, if you divide the group up by what he termed surface arthroplasty risk index, those with a high risk index for failure only had an 88.8 percent survival rate, four years, versus those with a low surface arthroplasty risk index with a 97 percent survival rate.

            The Harris Hip Score in spite of all of this was 93.5 on average at four years.

            Just as a bit of explanation, the surface risk index consists of a maximum of six points.  Two points are for femoral head cysts of greater than one centimeter.  Two points are for weight less than 82 kilos.   One point for previous surgery, and one point for a high activity level, and a risk index of greater than three was associated with a much higher revision rate.

            So if one had one femoral head cyst and had one prior surgery or had a high activity level, one could be considered at high risk for further revision.

            As I said, Dr. Amstatz is painfully honest about the results in his studies.  This is a study presented in the Journal of Bone and Joint Surgery in 2004 looking at 600 metal-on-metal hip replacement arthroplasties in which he reported five femoral neck fractures.

            In this particular case, you see that there's a collapse of the cyst underneath the head of the prosthesis, and in this one, this was probably a procedural error.  It's a little difficult to appreciate.  However, the head was not fully seated because the pressurization of the cement did not allow the cap to come all the way down.  This allowed a reamed area of the femoral neck to be exposed, and this patient suffered from femoral neck fracture.

            In summary, metal-on-metal hip resurfacing arthroplasty is prone to high cobalt and chrome ion concentrations comparable to that of metal-on-metal total hip arthroplasty.

            And, number two, femoral neck fracture is unique to this family of devices and deserves careful scrutiny with regards to appropriate patient selection and surgical technique.

            We will note that devices similar to that being considered for this PMA are currently being tested within the United States.

            I'd also like to point out that femoral head resurfacing arthroplasty is not a standard procedure taught in U.S. orthopedic residency programs, and one has to look at whether widespread implementation of this technique would or would not reflect the results seen in countries where the procedure is more commonplace and may be part of their usual training program.

            Thank you very much.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Anybody on the panel have any questions for Dr. Mabrey?

            (No response.)

            PANEL CHAIRPERSON NAIDU:  Seeing none, we should start a general panel discussion.  We are asked to consider an unusual PMA based on a retrospective study designed by a single surgeon based on British data set.  It is a challenging PMA.  Nevertheless, I'd like to get input from all of the panel members, and Dr. Mayor if you do not mind starting off commenting.

            DR. MAYOR:  I'd be happy to if I'm not held to too high a standard.

            I'd like to go through the PMA as we received it, particularly with regard to the FDA review memo because there were several things in that review memo that might not have suggested that the writer had English as a second language, but there were some things that I was concerned about in terms of how that came out.

            For instance, on page 4, the second paragraph describes a metal-on-metal resurfacing component with a high carbon content, but the high carbon content is assessed at 25 to 35 weight percent, which seems a little extreme.

            I'm assuming that what that really meant to say was a .25 to .35 weight percent of carbon, and that a surface roughness was identified as greater than .05, which I think should have read less than. 

And if there's anything about that that I should be reinformed about, I'd be happy to hear about it.

            The manufacturers may be better equipped to address the question on page 6 which had to do with the screws locking into the cortical bone, but I understand that they also lock into the threaded lug during the final phases of their insertion.  And when you're actually tightening the screw down, it becomes snugger and snugger because of the dimension between the two.

            PANEL CHAIRPERSON NAIDU:  Yes, would somebody from the sponsor like to respond to that?

            MR. VALEZ-DURAN:  Yes, I would like to.  There's two questions.  One is on page 4, a reference to high carbon content and surface roughness, and the other is about the screw locking and lug related to technique.

            For the first question if I can get the attention of my colleague, I would like him to talk about the high carbon content and surface roughness.

            MR. BAND:  Tim Band, again, an employee of Smith & Nephew.

            You're quite right.  There were two typographical errors in the text.  The carbon content is between .25 and .35 percent weight for carbon.  It was actually on my slide during the earlier presentation, and the surface references are maximum of .05 microns RA.  So it should be less than rather than greater than.

            PANEL CHAIRPERSON NAIDU:  Does anybody else want to address the technique part?

            MR. VALEZ-DURAN:  I'm sorry.  You may have to repeat the question on the technique.

            MR. BAND:  Tim Band again.

            I think the question was about the use of the dysplasia screw in the lug on the acetabular dysplasia cup component, and the fact that the screw engages in the cortical bone, but then finally as it finally drives home in the acetabular component, the thread is also timed to be a full engagement.

            The purpose for this is so that there's no leverage of the dysplasia cup component as the screw is driven further into the bone.  It would have a potential levering of the cup over because it's as a timed position for the thread.  It advances the cup in a perpendicular manner as opposed to any leverage.

            Does that clarify?

            DR. MAYOR:  I think so.

            Page 7 identifies the cobalt chrome beaded surface as a coating.  I think if you are to be semantically rigorous, because it's a cast shape it's not a true coating.

            MR. VALEZ-DURAN:  The cast surface actually is a fully integral cast surface.  So it is produced as a component part fully attached to the substrate and is, in fact, not a coating.

            DR. MAYOR:  It was on the basis of that perspective that abrasion testing for integrity was not done.  Is that fair?

            MR. BAND:  In fact, we did do that, and that was submitted within the PMA package.  So we have done testing of the porous surface.

            DR. MAYOR:  There was one specimen in the test protocol.

            MR. GOODE:  Dr. Mayor, if I could just clarify.

            DR. MAYOR:  I'm sorry.  Yes, please.

            MR. GOODE:  My understanding is we did ask about abrasion.  The sponsor, I believe, did provide a justification for not doing it exactly like you just said, that the strength of that interface would be along the lines of the substrate of the metal because it is entered in cast.  Therefore, that would not be required.

            DR. MAYOR:  Right.

            MR. GOODE:  So you're exactly correct.

            DR. MAYOR:  Yeah.  Page 11-12, there was a discussion of the similar testing done on specimens of the implant system, and one component that was reported as having produced a wear rig (phonetic) which was considerably higher than the other four, but there was no convincing discussion of the assessment of why that might have occurred.

            I'm concerned because we've had some retrievals at the laboratory up in Dartmouth which would suggest that that may be an occurrence with clinical significance.

            MR. VALEZ-DURAN:  Yes, I'd like to introduce Professor Unsworth, who is going to respond to that question.

            DR. UNSWORTH:  Thank you very much.

            I'm Professor Tony Unsworth, Director of the Center for Biomedical Engineering at Stone University in the United Kingdom.

            The university does receive research funding from Smith & Nephew, and Smith and Nephew covered my traveling expenses to come here today to be with you.  Other than this, I have no financial interest in Smith & Nephew nor any of its products.

            Yes, to try and answer the question you raised, sir, it is quite common in experiments of this sort.  There are a number of reported incidents that the odd specimen, even though they're produced at the same specification as the rest do produce a high wear rate, and it has to do with the running in process.  I'm afraid I don't know the exact mechanism, but it does happen from time to time, but normally they do restore themselves as this one did.

            In fact, after about  between three and five million cycles, the wear rate dropped very considerably, and that can be seen in terms of the surface asparities (phonetic) that were -- or the surface topography when we got to that stage. 

            I do have a slide if you're interested.  I could show you the slide.  I think it's about number six, something of that sort.

            DR. MAYOR:  Well, while I have you there, there is an issue related to surface topography that also came up on page 13 where the topography was measured, but could all be measured at the polar area because the lens of the instrument wouldn't fit into the cup.

            How close to the edge of the lip could you get before you had to --

            DR. UNSWORTH:  You really only have the polar region in the cup because it's a noncontacting method of measuring the surface so that you don't damage the surface by putting a stylus across it, and so we have to get the lens into the cup, and it's fine on the head because, of course, it scans on the outside, but in the cup it was difficult to get it, other than in the polar region where the contact actually took place.

            DR. MAYOR:  Well, I'm wondering if you could get down perhaps to the Tropic of Cancer on the ‑-

            DR. UNSWORTH:  Sorry.  If you could get down to?

            DR. MAYOR:  To the Tropic of Cancer on the cup if you couldn't get to the equator.

            DR. UNSWORTH:  I just don't know because I didn't do the experiment myself.  It was one of my research associates.  So I couldn't tell you how far down it could get.  I apologize for that.

            We've got the slide if you'd like to see what happened.

            DR. MAYOR:  If you've got them and could throw them up, that would be fine.

            DR. UNSWORTH:  Yes, thank you.

            This was the first joint that didn't wear very rapidly, that wore at a nice, steady rate.  So this is joint one, and when you friction test it this was the start of it.

            And then after one million cycles, you can see that the surfaces were then becoming smoother.  After two millions cycles, it looks like three million cycles is the next one please.

            That's three million cycles, and most of them did that, except for that one that did not give the low pictures.  So I'll just show you what happened with that one.

            Could I have the next slide, please?

            Again, started off very like the one.  After four million cycles, it started to get smoother, but not as quickly as the others.  Then again at three million cycles it was considerably rougher, but then at five million, by the time we got to five million cycles, we continued that on.

            Yes, please, can we press it?  There.  It has become very like the original, which is now showing up there.

            So it is smoothed down eventually, but it just took longer and not really in process.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            DR. UNSWORTH:  Thank you, sir.

            DR. MAYOR:  And finally, I wonder.  It has occurred to me to ask the clinicians in the group supporting this PMA.  Why not do a resurfacing BHR on a 70 year old person, male or female, if bone stock and habitus are favorable.  Even with low activity on the list of intentions that the patient identifies is that because it lacks a track record or is that because there are some concerns about its use in those individuals who might be 70 or older, to expand on what it is that has inspired me to raise this issue.

            My clinical experience in following patients who had their arthroplasties done at age 70 reminds me that I get very uncomfortable when I see them back at age 85 and they're starting to show lucent lines around their cement metal, and I'm beginning to think that I'm going to have to do a very difficult and stressful operation, which will be stressful not just for me but for them, and now that they're 15 years senior to the time at which I did their index total hip, and I guess the answer to that question might inform me a little bit better as to what would be the circumstances once this becomes widely available.  Are there likely to be surgeons that would take the same approach in terms of for whom this approach might be beneficial since it's revisable, and revisability is not an insignificant factor.

            MR. VALEZ-DURAN:  This is Marcos Valez with Smith & Nephew.

            The first point of clarification I wanted to make is in the labeling that we discussed earlier.  The activity level of the patient was a consideration in the selection of the patient.  Age was one, but also the activity level of the patient was also as important.

            But you asked for a clinician experience and comment.  So I'd like Dr. Cecil Rorabeck to come up.

            DR. RORABECK:  Well, good afternoon, ladies and gentlemen.  I'm Cecil Rorabeck, and I'm a conflict of interest as a consultant for Smith & Nephew and have been for many years.  I consult with them on their hip systems.

            I should also tell you that I've had conflicts with J&J and DePuy on the knee side and Zimmer as well on the knee side, but that is not relevant I don't think to what we're talking about today.

            I'm a Professor and Orthopedic Surgeon at the University of Western Ontario in London, Canada, a member of the Hip Society in the United States, and the International Hip Society, American Academy of Orthopedic Surgery, and I restrict my practice to total hip and knee.

            I have also been the past president of the Canadian Orthopedic Association, the Canadian Orthopedic Research Society, and currently am the Vice President of the College of Physicians and Surgeons of Canada.

            So with all of that out of the way, let's try to deal with the question at hand, which is a good question, and I also am one who travels the world a lot like you all do, I'm sure, and looking at the probable indications of this procedure, it seems to me at least when you're starting out that you want to choose people with good bone stock and bone stock that's reproducibly good with time.

            So what does that mean?  Well, in my hands that means a male under the age of 65, probably, and it means a woman with a normal DEXA scan, probably under the age of 60.  Now, that's purely empiric.  There's no scientific data to back it up, but if we accept the fact that with time people are more likely to become osteoporotic and if we accept the fact that the major shortcoming, if there is a major shortcoming, it's a potential for neck fracture.  In my view at least when we are starting out, we should restrict the cases to patients under 65 or patients with normal bone stock, normal bone density.

            Does that answer your question?

            DR. MAYOR:  In some respects, yes.  I still wonder whether there's going to be a temptation because of revisability that this implant system may become increasingly attractive as a possible solution for a wider and wider array of patients.

            DR. RORABECK:  Well, I mean, you're quite right.  It might happen that way, but you know something?  We have such good, good things for patients, as you know, at age 75 to low demand.  To me I don't think this is really what this implant is trying to address.

            Now, if you have a 75 year old Swedish farmer from Minnesota who has got fantastic bone stock and is very aggressive, perhaps there is a place for it, but I think it would have to be individualized in older patients.

            DR. MAYOR:  And the my final question is based at anyone who feels willing to approach it, and that is that we've dealt a lot with the issue of metal ion concentration in serum and urine, but I was concerned at the last academy meeting by my reading of a poster exhibit by Josh Jacobs who looked at a series of patients with bilateral, metal-on-metal total hips, and on the basis of the serum levels that he could identify in those patients during the subsequent measurement of their serum levels, he raised the concern that the levels for the bilaterals was more than twice the levels for the unilaterals, and that as he pursued the question in discussing the issues began to suggest to him that there might be a saturation problem that you could encounter in regard to clearance mechanisms that the body can use to deal with these ions, and that the load from a bilateral implant might actually exceed the capacity of those clearance systems to respond to it.

            And since there is some evidence provided just now by Dr. Mabrey that the serum levels may be higher for the resurfacing than they are for the smaller head metal-on-metals, does this raise a serious concern for those patients with bilateral implants and have we done any measurements which would either confirm or assuage the concerns that Dr. Jacobs is raising?

            MR. VALEZ-DURAN:  Marcos Valez from Smith & Nephew.

            There's data on the PMA about the metal ion.  They are all related to the DHR.  However, to respond to your question to go into more details to response to your question, I would like to invite Mr. Joseph Daniel to the podium. 

            MR. DANIEL:  Hi.  I'm Joseph Daniel from Birmingham, England.  I'm an orthopedic surgeon.  I don't have any financial interest with Smith & Nephew, but my travel and stay here are being paid for by them.

            Regarding the question, shall I take the second part of the question first, which is that Birmingham hip resurfacings produce more hip resurfacings as a group, produce more metal ions than smaller diameter, 28 millimeter total hip replacement.

            Now, if we look back at the fluid film lubrication theory, it would suggest that a larger diameter bearing has the potential to generate a full fluid film lubrication, and therefore is likely to wear less. 

            The article by Clarke which was shown earlier seemed to suggest that the resurfacings produce higher levels of metal ions.

            Now, there was a confounding factor in that article and in that two types of resurfacings were combined, and the metallurgy and microstructure in that variable group is different in the two types of resurfacings, and in fact, that point has been highlighted by Dr. Josh Jacobs himself in his recent article in Journal of Arthroplasty in December 2004, that this confounding factor has been found.

            And the work of Dr. Josh Jacobs himself in that article he presents, that he does not find a difference between a 28 millimeter replacement and a larger diameter resurfacing.

            Our own data also show that either in the 24-hour cobalt output or in the whole blood metal ion levels there is no significant difference either at the two-year period or at the five-year period between resurfacing and 28 millimeter metal-on-metal total hip replacements.

            On the other issue about bilateral resurfacings and the question of renal threshold, the question of renal threshold can be looked at.  When you look at the paper done some time ago in patients with renal failure and it's found that the metal ion levels in patients with renal failures tends to go up 100 times what it went up in regular people with no renal failure.

            So the metal ion generation is not in terms of one or two times the metal levels in the blood, but much higher, and kidneys seem to have a large renal threshold to get rid of this excess metal ions.

            We are, in fact, in the process of doing a study of patients who had one resurfacing some time earlier and then come back for a quadrilateral hip resurfacing later on.  We found in these patients that compared to the metal ion levels output in urine, daily output in urine, before the second of the quadrilateral hip operation, after the operation the levels go up more than three times in daily urine output.

            The whole blood levels also go up, but they do not go up three times.  They go up around twice the level before the second operation.

            Now, there are differences between Josh Jacobs' technique and specimen and the technique and specimen that we have used.  We have used whole blood rather than serum, and we have used high resolution and inducted a couple plasma mass spectrometry (phonetic) rather than rapid burning atomic absorption spectrometry.

            Now, this point is also relevant.  The reason why we chose whole blood rather than serum is because it has been shown in 1995 by Merritt and Brown that chromium especially tends to get sequestrated in blood cells, and so she has recommended that the ratios of chromium levels in red blood cells to plasma is so different that both blood compartments need to be evaluated, and that was a very valid statement.  In fact, it is intracellular chromium that tends to be more of a health hazard than the extra cellular chromium.

            Can I have the next slide, please?

            So we looked at a group of nearly 260 patients, and we did simultaneously seat them and hold the balances, and when we look at their results, we find that this is normalized scatter in which the whole blood level has been brought to one by adding a correction factor, and the same correction factor was applied to the serum levels are allowed to scatter as they correct the whole level.

            So this scatter shows that at the lower concentrations, the variability between whole blood and serum is much more than at the higher concentrations.  So they are starting with, say, a little lower concentration and then going on to a patient bilateral that's much higher.  We're not exactly sure whether the serum level if representing everything that is contained in whole blood as a whole or not.

            So there is a larger element of variability at lower levels.

            Can I have the next slide, please?

            And the same thing is also seen by the Pearson correlation corruption (phonetic).  In the first graph, the whole blood levels for below one microgram per liter, and you find that the correlation coefficient is very low, not .14 as compared to over two where the correlation seems to be very good.

            Can I have the next slide, please?

            We also tested it with another statistical method that was just called a Bland Altman plot, and we find that the limits of agreement between whole blood and serum are too far apart to give us confidence that serum is able to predict lively what is contained in whole blood, and we believe whole blood is a measure of the systemic metal ions exposure or the body burden of metal ions as a whole.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            I do have a question for you, but nevertheless, you still have not answered the question.  Dr. Mayor stated that it went up more than double for bilateral, and you did admit that, didn't you?

            DR. DANIEL:  Yes.

            PANEL CHAIRPERSON NAIDU:  So it did go up no matter how you measure it, whether it be in the serum or whole blood.  After the bilateral arthroplasty, correct me if I'm wrong, but it did more than double, at times triple, correct?

            MR. DANIEL:  No, it didn't go up to triple.  The daily output went up to triple or the urinary excretion of metal ions went up three times, but the whole blood levels did not go up to three times, and the daily output -- the fact that whole blood does not increase as much in terms of factors as urine seems to suggest that the kidneys still have a threshold to take in more chromium and get rid of it.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            DR. MAYOR:  One final question and then I'll yield.  For the company is there a formal design for a revision system that would assure that the conversion from a surface replacement to a total hip stemmed implant would produce a good match between the head and the cup.

            MR. BAND:  Tim Band, Smith & Nephew.

            There was actually a revision system for this component for the femoral side, which is a modular head.  It has produced the same exacting standards and specifications material both in terms of microstructure, sphericity and so on and has a taper which is compatible with all of the Smith & Nephew 12-14 tapered stems.  So there's a full modular system.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            We'll come back to Dr. Blumenstein when we deal with the statistics in more detail.  I'd like to get the clinical reviewer's comments over with at this point.  I'd like to go to Dr. Skinner.

            DR. SKINNER:  Me first?  Okay.  This is Dr. Skinner.

            I'd like to ask a couple of questions.  I'd like to follow up on the renal function thing.  the thing that worried me from the data that I saw was that for a patient who would walk two million cycles a year -- and we're talking active, young patients -- looking at the wear data for the steady state, it looked to me like walking two million cycles per year would give you about a steady state disposal of the cobalt and chromium that you get into your blood by the kidneys.

            But if you walked more than that, you would have to boost up the excretion in the kidneys to keep the mass balance from shifting to a higher level that would mean you'd be accumulating cobalt chromium.

            You didn't address cobalt.  It sounds like chromium would take care of itself, but it also raises a question that perhaps this should be contraindicated in people who are likely to have renal failure, for instance, in diabetes.  Could the company or one of the physicians comment on that?

            MR. DANIEL:  This is Joseph Daniel again from Birmingham in England.

            The first question about cobalt and activity, cobalt levels and activity, we did a study looking at output versus activity as measured by step activity monitor, and we did not find any correlation between activity and the cobalt or the chromium levels.  We did the same thing against age and against body weight, and a product of activity and body weight, and we did not find any correlation.

            In fact, Dr. Josh Jacobs, again, has done another study in which he subjected patients to rigorous activity like treadmill walk.  You did the serum levels before the activity.

            If I can have a minute, I'll just show you.

            I'm sorry.  I'm unable to find that, but he concluded in that study that activity and serum metal ion levels do not show any correlation.  In fact, he did not find any increase at all when he measured the serum metal ion levels before the activity, during the activity, after the activity, and a time period a few days after the activity as well.

            I might be able to show that slide if I can have a few minutes.

            DR. SKINNER:  I think that would be okay.  I was concerned because, of course, an accumulation of cobalt might be a problem based on those cardiomyopathies that were associated with high cobalt doses years and years ago in the literature.  We certainly don't want to get into a situation like that.

            MR. VALEZ-DURAN:  Marcos Valez from Smith & Nephew again.

            In the presentation of labeling, we are planning to contraindicate this product for patients with borderline renal failures.

            DR. SKINNER:  Well, there's a very high percentage of people with diabetes or will have diabetes, which means they'll have probably hypertension, and then they'll have renal problems.  I'm asking is diabetes something that ought to be considered in there.

            MR. VALEZ-DURAN:  Actually we did not consider that, but we'll take it under advice, and we'll discuss that with FDA as well.

            DR. SKINNER:  Going to another question, Dr. McMinn, Mr. McMinn did this procedure according to a protocol that was given with a rather sizable incision, and it looks like to get the exposure you need to do the cup and keep the femoral head on, but it's going to be difficult to do with something that's very popular in the States now, the mini incision.

            Should the labeling perhaps mention something about relatively contraindicated for a mini incision?

            DR. McMINN:  Derek McMinn, again.

            I'm not sure that would be required.  I've been doing mini incision resurfacing for some years.  So, for example, we've looked at all the cases that I've done through 2004, and the mean incision length was 11.6, I believe, centimeters.

            So it can perfectly well be done through a small incision if the surgeon wants to do that.  We have looked at a load of objective measures, flood loss, length of time in hospital, and we find no correlation between incision length and the objective data that we have recorded.

            So from a surgeon point of view, there really is not a very pressing reason to go to mini incision resurfacing.

            We have got a group of over 40 patients, however, who have had a standard incision resurfacing in the past, and more recently have had a smaller incision, and we took the opportunity to ask those patients what they felt, and unlike our objectively collected data or at least what we thought was objective data, blood loss, et cetera, the patients almost uniformly preferred the short incision.

            So there will be surgeons who want to go to a short incision, but it is technically more difficult, and when I take surgeons through their early procedures, I strongly advise them not to go with a small incision surgery and make life even more awkward for themselves.

            So we have started surgeons in 23 countries.  So we know quite a lot about how to get surgeons going safely with the Birmingham hip resurfacing, and the one thing I would urge them not to do is to try mini incision surgery. 

            When they're really proficient  at the operation, if they do desire and their patients so desire, then they can reduce the length of the incision.  But since it's not a big deal in terms of outcome, objectively assessed outcome, not a good reason to start there.

            Does that answer you questions?

            DR. SKINNER:  It certainly does.  I didn't know you had done them through the mini incision.

            No questions.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Mr. Whittington.

            MR. VALEZ-DURAN:  We have somebody else that can offer a different point of view on the incision size from the U.S.

            DR. ROGERSON:  Dr. John Rogerson from Madison, Wisconsin.

            Again, I've not put in a lot of these, but I've sent 30 patients to Europe, and the thing that has been impressive to me is that these patients have come back from Dr. DeSmith with large incisions, and yet have qualitative differences in terms of less pain with those bigger incisions than the ones that I'm doing with mini invasive approach.

            And I think that that speaks to pain generation in total joint arthroplasty, and I don't know that the pain generation is all related to the size of the exposure.  I think that my experience with the Coffield shoulder resurfacing arthroplasty is that those patients have much less pain than a traditional hemiarthroplasty or one that's stem, and the people that I see that have had resurfacing arthroplasty in the hip really don't have much pain with the procedure even though the incisions and the exposure are considerably greater.

            So I have been impressed that the pain related to the -- and I guess the final thing I would say is the patients that come in, and you know, I will tell them you're going to have a longer incision if you have this operation -- they have all done their homework.  They've been on the Internet.  They know exactly what the incisions are involved.  What they're interested in is not minimally invasive skin incisions.  They're interested in minimally invasive surgery and what their long-term prospects are for a conservative resurfacing replacement, but more particularly for revision later than just the fact that it's a small incision.

            They're not at all interested or even really care about the size of the size of the incision.  They'll tell you that right up front.  They really want to continue with an active life style and then when it comes time for revision later to have that be a more conservative operation.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            MS. WHITTINGTON:  My questions related to patient safety specifically as the consumer representative on this panel.  I'm concerned about the patient labeling information that would indicate to the community reader, to the typical patient population.  Renal insufficiency I don't think means a whole lot to them, and I think the previous question on long-term effect on the potentially chronic kidney disease patient, what kind of methods do you have in place to screen patients or is there a proposal to screen patients for the level of kidney disease and have you identified those patients in which this would be contraindicated because of the level of kidney disease?

            Similarly, the same type of question for the patient with avascular necrosis.  In my practice, my experience those patients surf the net dramatically, and they're very aware of the procedures that are available, and yet it simply says avascular necrosis, and yet there is a specific level of tissue death or damage to the femoral head that you've identified that you wouldn't progress to.

            And yet that, indeed, is where as previously asked about the femoral neck fractures, and it's a higher incidence in this patient population, and the inability to differentiate or really decide between avascular necrosis and femoral head collapse.

            So I'm concerned that these younger, more active people are not going to understand why they can't have this.  I think it needs to be clearly identified.  I don't see that identification or that description in your patient labeling information that you submitted with the materials.

            MR. VALEZ-DURAN:  These are good points that you have brought up, and our proposal would be to work with FDA to consider those concerns and reflect them in our labeling.  And if there's anybody from Smith & Nephew that would like to add something, they're welcome.

            DR. McMINN:  Derek McMinn again.

            Please have in mind that my main patient age at operation is 53, and so it's rare in a fit individual of 53 to find significant renal problems.  I think the issue of renal insufficiency with metal-metal bearings is going to be much more important in the metal-metal total hips, which will get done on a much older group of patients, and they will be done because of the large head to reduce the chances of dislocation, but the resurfacing group, they're generally fitter and they don't have renal problems, in general.

            Does that address some of your concerns at least?

            MS. WHITTINGTON:  It doesn't address just a screening, you know, a GFR or a creatinine on patients.  In my experience I see more people than I'd like to see with chronic kidney disease that are not aware that they have it. 

            So asking a patient history is not going to give you the information, especially given the incidence of increased metal ions and the potential problems with that.

            DR. McMINN:  All of our patients have creatinine and urea preop.  If they're abnormal, they don't have a metal-metal bearing.

            MR. VALEZ-DURAN:  Once again, the comments are well taken, and we will work with FDA on the specifics of the labeling.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Ms. Adams.

            MS. ADAMS:  Just as a follow-up to that question regarding labeling, I want to clarify the labeling that we're looking at is the labeling that the sponsor has submitted which would be part of the prescription labeling that the physician would see.  Is it intended that this is the labeling that would be used for consumers if someone were to go to the Smith & Nephew Website, for instance, to read about this?

            MR. VALEZ-DURAN:  There will be specific patient labeling that would be developed.

            MS. ADAMS:  So when you say you would work with the FDA to consider these concerns, it would be in the context of the patient labeling?

            MR. VALEZ-DURAN:  Correct.

            MS. ADAMS:  Okay.  Thank you.

            MR. VALEZ-DURAN:  Thank you for the clarification.

            MS. ADAMS:  I have another question, which is really I'd like to ask the sponsor to return to some of the questions that were asked earlier.  There were two questions that were asked regarding deaths and wanting a little bit more information about patient deaths in the study, and then there was also a question by Ms. Whittington having to do with whether there was a correlation of infections to revisions.

            MR. VALEZ-DURAN:  Thank you for bringing those.  We did promise we would look into that particular information and bring it back so the opportunities for us to respond.

            Ms. Marlow has put together a summary of that information based on the data that was provided on the PMA.

            MS. MARLOW:  Marie Marlow.

            Pam, thank you for giving us the opportunity to provide these responses.  During the break we were able to do these additional analyses for you.  As far as the deaths go, we're going to get a table up here for you.  We stratified them according to age.  It turns out that the average age of the patients who died was 63 at the time of death.

            For example, the youngest patient, I believe that age is 53, but we'll have that for you up here in just a moment, died of suicide.  There are several cancer deaths in this patient population.

            All right.  So here's the summary of the causes of death, and again, the range in age there, of course, is the time at death.

            Is the table available?  Great.  Thank you.

            All right, and then there's the breakdown for you.  Age at the time of death and cause of death, and we've stratified this table for you by age at the time of death, and if I can provide any more information than this just ask, and hopefully I'll be able to do so for you.

            MS. ADAMS:  If I could ask a follow-up question, how you determined that there have been any deaths associated with this device?

            MS. MARLOW:  No, there's been no deaths whatsoever associated with the device or the procedure.

            MS. ADAMS:  Thank you.

            MS. MARLOW:  All right, and then I'll turn the second question over to George DeMuth who took some of the data on the infections and the wound exudates for you.

            MR. DeMUTH:  Yes, I may have to refer to somebody else as well. 

            The first cut was just to look at whether the revision rate at all was hiring patients that had wound accident.  Just in a percentage basis, it was four out of the 589, you know, .7 percent essentially, and out of the remaining patients were the 23 other revisions, 1.3 percent.  So more revisions in the patients that didn't have wound exudate.

            Then when you go look at specific due to infection, only one of the four in the wound exudate patients was related.  So it's 0.2 percent of the patients.  These aren't survival rates.  They're just cut rates.  The other seven were actually in the non-wound exudate patients and so, again, 0.4 percent.  So it didn't appear to be at that kind of gross look associated with it.

            MS. ADAMS:  Thank you.

            I'd also like to ask some questions and make a comment regarding financial disclosure and conflict of interest.  The panel received three letters that are, to my knowledge, intended to be part of the record, but there was no information regarding conflict of interest associated with the letters.  I just want to clarify whether or not we have received any of that information.

            MS. SCUDIERO:  The letter are just since we received them.  There was nothing -- the people who sent the letters received that same little statement that I provided to all the other speakers about conflict of interest, and it says that you don't need or  you don't have to state your position or if you have any affiliation you can still submit or still speak, and I think it was probably just overlooked.

            MS. ADAMS:  Thank you.

            Also I'd like to come back to the comments that were made during the public session.  The distinguished Dr. Maloney made some comments regarding financial interest.  I know we as a panel because we received training on this are not supposed to consider those, and that that's not the subject of our deliberations today, and given that, I would like to also just acknowledge that it is, of course -- there's probably no panel that ever meets that doesn't have a significant amount of financial impacts that will come from our deliberations, and even though we're not going to talk about them, I don't think there's anything unusual associated with that sort of thing, especially with devices that are presented to our panel which represent new technology such as this.

            Nevertheless I do want to ask the sponsor whether or not they have met their regulatory obligations regarding disclosure in the PMA having to do with finances and conflict of interest.

            MR. VALEZ-DURAN:  The appropriate financial disclosure materials have been submitted to PMA as required and FDA has at least part of our PMA.  If FDA wants to comment on that they're welcome to.

            MS. ADAMS:  Thank you.

            PANEL CHAIRPERSON NAIDU:  No comment?

            MR. MELKERSON:  The financial requirements are actually part of filing.  If that information is not there, the PMA would not be.  It's the one reason for not filing a PMA.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Melkerson.

            MS. ADAMS:  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Let's continue with the clinical discussion.  Dr. Kim.

            DR. KIM:  Hi.  Joel Kim from UCSD.

            I have a couple of questions.  Let me just start out with one at a time.

            There's been some mention that the surgery can be challenging technically, but we didn't hear much about any analysis of a learning curve.  Do you have any information or data that would give us a better idea of this learning curve and how many cases one would expect to have to do to be at a level where we can do the surgery reliably?

            MR. RORABECK:  All right.  Cecil Rorabeck again. 

            That's a key question, and I'm sure to some extent it's going to vary from surgeon to surgeon, but let me just say this.  As someone who is doing the technique and somebody who has spent my life basically doing hip and knee surgery, first of all, I don't think people should be turned loose with a technique like this until they've had appropriate training, and in our country we have workshops where they come and actually do a cadaveric lab to learn the technique, and I think in my opinion that's a good thing to do.

            Having said that, I think to get up to speed with this, one first has to see somebody do the technique and probably see two or three cases, follow a video, do the sawbones, and so on, and at the end of that, assuming the person is going, say, 50 hips a year, I would think that the learning curve is going to be about ten to get up to a level of comfort.

            And why do I say that?  Well, I think that I'm talking about picking and choosing patients carefully, and if we pick men under the age of 60 that are not particularly obese or women with normal DEXA scans, again, 60 or less, I think we're going to learn pretty quickly. 

            On the other hand, if we're picking people that are 68 that are obese and various things, it's going to be more complex.

            DR. KIM:  I take it from your answer though that there is no data that we can evaluate to get a handle on this.

            DR. RORABECK:  Well, you're quite right.  There is no data.

            DR. KIM:  The second question has to do with bone stock.  Is there a quantitative value that was identified to help us figure out which patient has adequate bone stock and which one does not?

            DR. RORABECK:  Well, as I've said, I think that a DEXA scan is important in women in particular.  In men we evaluate them the same way as evaluate a patient for a cementless total hip replacement.  If they've got appropriate bone stock in terms of a Type A or B bone, then they're probably going to be a candidate for resurfacing.

            So in that sense we use a similar indication.

            DR. KIM:  One final question has to do with the metal ion discussion.  I haven't had a chance to review this data and unfortunately the PMA does not do a very good job of explaining to us at what value having metal ions is detrimental to our health.  It doesn't even talk about animal data.  So is there -- do you know of any data or anyone in the sponsor group have some information that would reassure us that the metal ion concentrations that we're seeing are detrimental, safe?  How far from detrimental are they?  What orders of magnitude below those thresholds are they?  Questions like that.

            DR. RORABECK:  I think one of the things that has been most interesting to our group, and we did a prospective randomized trial looking at metal-metal versus metal on polyethylene in a group of patients blinded as to which hip they had, and that data is now out five years.  Our metal ion levels, and these are intracellular, cobalt and chromium ion levels and the metal-metal group are clearly higher.

            And while they're higher in year one and two, there is a tendency for them to fall off toward year five, but what's the significance of these data and what does that really mean?

            And our data is not dissimilar from the data that's in your handout. 

            I think that we've looked at it as carefully as we can, and what we've done is we've gone to the various governments around the world like the British government, Canadian government, the German Ministry of Labor to find out when somebody would be taken off an assembly line with a serum cobalt of X, and that's the  yard stick or the meter stick that we've been using in our country to try and define what's acceptable and what isn't because you're quite right.  We do not know the answer to this.

            And I can assure the panel that with no exceptions in our data and I think also in this data, the serum ion levels fall within acceptable levels of the Entero (phonetic) Ministry of Labor or the German Ministry of Labor or the U.K. Ministry of Labor.

            Now, I could not get data for the U.S. on that, but that would be my answer to it.  We don't know the answer, but I think we're acceptable.

            DR. KIM:  Thank you.

            MR. VALEZ-DURAN:  I just wanted to follow up to the response.  You're correct that we do not have a specific analysis of learning curve.  However, we do have significant numbers of series of patients that have been performed over the years.  The survivorship percentages are very high, not just for Mr. McMinn, but published series in the British Journal of Bone and Joint, as well as the experience in the Oswestry Center for a total of about 140 other surgeons. 

            So still the survivorship is very, very good at five years for those patients.  So just to give you a level of confidence on that.

            And there may be some additional information on ion levels.

            DR. DANIEL:  I'm Joseph Daniel from Birmingham again.

            Regarding data about the possible adverse effects on health due to metal ions, there are two main concerns which seem to come up from time to time.  One is what does it do in terms of cancer rates, and two is what happens in children and women of child bearing age whether they crossed the placenta or not.

            So in holding on to these two questions, there is one reference by Visuri (phonetic) from Finland which was submitted along with the PMA, which talks about cancer rates in patients with metal-metal total hip replacements.

            Can I have the slide, please?

            Vissudi (phonetic) followed up a series of 579 historic metal-metal total hip replacements done from 1967 and followed them up for a 30-year period.  The 28-year results are shown here.  He has shown that when compared to the general population there is no difference either in the all site cancer rate -- next one, please -- or in the site specific cancer rate, and the only difference is in lung cancer in which if you see the 95 percent confidence intervals both the limits are less than one.

            So lung cancer is lower in incidence as compared to the general population, and in the rest of the site specific cancer rates, all of the 95 percent confidence intervals cross one, which means there is no significant difference between this group and the general population.  The total number of person-years considered in this is 9,700 and on, and this compared very well with the metal-on-polyethylene total hip replacements as well, which was studied in other studies, a series of analysis of 70,000 patients also shows similar results and there is no difference.

            On the question of whether it affects the unborn child -- can I have the next slide, please? -- there is only one publication at present, and this seems to suggest -- the next one, please -- that the blood serum metal ion levels are not recordable in umbilical cord blood in patients with metal-metal hips.

            DR. KIM:  Thank you.

            Are we going on to discussion of the statistics?

            PANEL CHAIRPERSON NAIDU:  Let's just finish with Dr. Mabrey, and then we'll go into statistics.

            DR. MABREY:  I had three questions.  Two of them are primarily for clarification.

            Number one, I'd like to just clarify that 100 percent of all of the acetabular cups evaluated with the Oswestry study were HA coated throughout from the very beginning of the study.

            MR. BAND:  Tim Band, Smith & Nephew.

            All of the components were HA coated.  There's been no design changes to the system.

            DR. MABREY:  And no change in the technique of application of HA.

            MR. BAND:  No.

            DR. MABREY:  Second, what percentage of cases does the sponsor anticipate would require use of the dysplastic cups?

            This application is slightly different in that we're not just looking at one acetabular system, but actually three, the acetabular cup and the bridging cup as well as the standard cup.

            I do remember some numbers from Mr. McMinn's study, but his population may not necessarily reflect the number of acetabular or dysplastic cases out there.

            DR. McMINN:  Derke McMinn again from Birmingham.

            In my series of Birminghams, I had 176 dysplasia cups out of, I believe, 2,300.  So somebody mentioned a figure of seven percent, if anybody is any good at math.

            However, that very much depends on what you are prepared to take on as a surgeon, and a surgeon who is starting the Birmingham resurfacing would be well advised not to start with a high CDH.  That would not be a good place to start, and so the level of usage of the dysplasia system will very much reflect the experience and confidence of the surgeon.

            And many surgeons throughout the world using the Birmingham system have a usage of zero.  Mine is rather higher than that, but I find it an extremely useful device.

            Incidentally, in the 176 I had no implant related failures in the 176 dysplasias.

            DR. MABREY:  This gets back to Ms. Adams' comment about the contraindications on the package insert.  Now, does the sponsor anticipate adding restriction as to the crow (phonetic) type of dysplasia that one would recommend this for or will that be left to the discretion of the surgeon?

            DR. McMINN:  No, I think you have to leave that to the discretion of the surgeon.  Of course, it's very rare that you would ever dream of doing a resurfacing in a crow Rate 4.  So you have a tiny acetabulum with a head well up. 

            (A) you don't need a dysplasia cup because that socket is complete even though tiny, and because it's tiny, there's very little prospect of it being able to resurface and get a head that's small enough onto the head leaving any bone.

            So fours, I think, for very obvious practical reasons will not be done, but Crowe Rate 3 and Crowe Rate 2 are absolutely prime targets for the dysplasia cup.  In my practice Crowe Rate 1 often you don't need anything other than a regular spherical cup.

            DR. MABREY:  Thank you.

            DR. McMINN:  Does that answer your question?

            DR. MABREY:  Yes.

            MR. VALEZ-DURAN:  In addition, as you mentioned, it's something that we could consider in labeling, but most importantly we could consider as a factor or something to include in our surgeon training.

            DR. MABREY:  My third question relates to a topic brought up with Dr. Skinner earlier on, and this is the comparison of populations and the percentage of diabetes that one would anticipate.  Prior to moving to Dallas, I was situated in San Antonio for 13 years where 50 percent of the population was of Mexican American heritage. 

            In that particular population of patients and only within the United States they have a much hither incidence, in some cases up to one third of Mexican American Hispanics over the age of 65 have some type of diabetes mellitus.

            Now, one does not see this in the population in Mexico City.  So it happens to be a cultural effect.  I bring this up because one of the questions that has been addressed to the panel is are the populations similar, and when I look at the population statistics, I didn't see much comment on the use of this device or follow-up of this device in Hispanics or Mexican Americans.

            MS. MARLOW:  Marie Marlow, again.

            That's a very good point that any study in the U.S. would have to address is patient demographics, although I wonder with the percentage of Spanish Americans, Mexican Americans in the United States what size of study would have to be undertaken in order, one, to recruit a sufficient population of Mexican Americans, and two, to recruit a sufficient population of those with diabetes, although this is an excellent consideration.

            This may be one of the things that we could address in the post approval study by making sure that we select cities where these kinds of representative populations live.  That's a very good idea for us.

            DR. MABREY:  That's all.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Let's go on to the statistical discussion.  Dr. Blumenstein, if you could start off on that, that would be treat.

            DR. BLUMENSTEIN:  Okay.  I just want to go over a couple of things.  First of all, from what I can see intervention appears to be effective in some sense and also appears to be reasonably safe given the effectiveness.

            The issue to me is whether this is well controlled, and by that I mean the effectiveness relative to predicate effective interventions has not been established, and I'm going to go on and explain that and then make some comments on the fact that this is all limited by the study design.

            First I want to cover some minor issues that are -- just to get them out of the way.  First of all, the estimation of the survivorships are simply wrong, and the reason they're wrong is that the competing risks are either lost to follow-up, but in this case, in particular, death, are censored rather than using a methodology that takes into account the difference between competing risks and limited follow-up, and I can give you references on this if you need to have them.

            This is a common problem, and it's particularly a problem in things like presentation of Kaplan Meier curves with cause specific deaths censored, deaths not relative to the primary cause.

            There are substandard data collection practices in this study.  There's an unplanned criterion for success.  We've already heard these  things, and then, of course, there's a single clinical site in another country and a single surgeon which means that you cannot assess the homogeneity of results across surgeons or site.  We just simply do not have that data.

            What I wanted to address here is the control and cohort studies.  No matter what we do when we evaluate an intervention, we're comparing it to something, and there's a specific situation that's important to think about, and that's what's called the unmet medical need, and this is a case where there's no intervention, and you expect either no improvement or worsening relative to the primary efficacy measure if there is no intervention, and there is no existing effective intervention, and this is a situation called unmet medical need, and in that case a cohort study, that is, you call it a case series or if we want to be insulting as statisticians we call it a convenience sample, and so forth.  These are appropriate in this unmet medical needs setting because what you're doing is you're trying to see if you can get your measure of effectiveness to show that you do have effectiveness and you have no expectation.

            But we don't have that situation here, and I wanted to point out that this first in class which we've heard about here does not refer to the first intervention for this disease.  In other words, we do not have in the strictest sense the unmet medical need definition that I gave on the previous slide, and specifically we have effective predicate intervention, and the safety and effectiveness cannot be adequately characterized without comparison to the existing predicate effective interventions.

            The exception to this is where the predicate would be minimally effective, where you feel that you can, based on the historical data that your result is going to be so overwhelmingly positive compared to some predicate that's only minimally effective, but in general you cannot establish effectiveness when an effective predicate exists.

            Now, when we compare interventions, the highest standard, of course, is the randomized clinical trial, and that's what I think should have been done here, but what we have here is a single cohort study with a comparison to historical data, and this suffers because the differences in effect, that is, the outcome, are confounded with the cohort differences, that is, the data that you can use to compare to might have been in a different population or different methods are used to collect the data.  Different endpoints are used.

            We have already seen the score.  The differences between the scores have been discussed extensively, and this is further exacerbated when the data that you're comparing to come from literature because you don't have control over the data.  You have to accept whatever measures are there.

            Some people use meta analysis, methodology for comparing interventions, and I just wanted to put up this sort of descending order of validity, and the best kind of meta analysis is when you get the data on all randomized clinical trials, and you put all of the data  in a computer and you crunch it.

            The next level down is where you extract estimate from one or more randomized clinical trials from the literature, that is, you don't have the complete data sets on all of the randomized clinical trials.

            The next one down is where you get data on all cohorts and you put it together and crunch the data in a computer.

            And the last and the least of the methods is extract estimates from one or more cohorts from the literature, and this is exactly the situation that we have here, is that we're comparing a cohort of data from a single institution  with a single surgeon, and we're comparing it to literature.

            And the whole point here is that a meta analysis of randomized clinical trials, there's no confounding of populations to the differences in effect, but just as it would be in any other setting, but the metal analysis of cohorts is still confounded.  So there's just no way around that.

            So for me the bottom line is that the phrase "well controlled" doesn't apply.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Anybody have any questions for Dr. Blumenstein?  Dr. Kim, you had a previous statistical question.

            DR. KIM:  I think you have answered it, but I just want to make sure I understand.  In several of the five-year groups the loss to follow-up percentage is about ten percent.  Yet the survivorship is well over 95 percent.  If we assume that those ten percent had a high percentage of failures, wouldn't that significantly affect the survivorship analysis, for example, if eight percent out of the ten percent were lost to follow-up because they had a terrible result?  Would that affect the final survivorship and if so, by how much?

            DR. BLUMENSTEIN:  Well, the lack of follow-up, if you're estimating survivorship at a particular point in time, say, five years, and not all patients are followed for five years, we have actuarial methods of estimating survivorship based on that, I mentioned that the estimate that's being used is not valid.  There are estimates that are valid.  The reason the estimate to me is not valid is because deaths which are a terminating event, a competing event, were censored, which is what you do when you have inadequate follow-up on patients.

            But the lack of complete follow-up on all patients in and of itself is not a problem.  If the lack of complete follow-up is biased in some way, then that's a problem and that has to be dealt with in some fashion.

            The scores that were analyzed over time, there was some mention about missingness, whether missingness was random or not, and that's going to affect the validity of the statistical analyses, but inadequate follow-up or incomplete follow-up, rather, is in and of itself not necessarily a problem.

            DR. KIM:  Sorry to be so nitpicky.  If we were to assume that that group of patients did poorly, in other words during the worth case scenario, what would the survivorship value be roughly, in the worst case scenario?

            Maybe the sponsor can answer that since they may have access to that data.

            MR. VALEZ-DURAN:  Thank you for the opportunity to respond to the questions.  I wanted to mention that in the selection of the literature comparison, we actually did a full review of the literature to select the two comparative histories or literature controls that we selected.

            It's also worth noting that there are not a lot of randomized clinical studies published on orthopedic devices.  So that presents a limitation on any meta analysis that you could perform.

            I also want to mention that we need to just keep in mind that we only have 27 revisions.  So regardless of the method you use to calculate the survivorship, there's only 27 out of a very large cohort.

            But I would invite our statistician to further comment on the comments by the statistician on the panel.

            MR. DeMUTH:  Well, part of that I'd like to defer, but I think we do know we have 20 patients that died.  Without going back and fitting a survivor -- you know, treating those as all failures, for instance to try and maybe to get a bound on death plus revision, we just note, you know, that number of patients or 27 out of the 20 -- I don't know how many of the 1,626.  So it seems there's a bound on some, a couple of percent perhaps.  Maybe it's a little more, and so that leads to the possibility that we'll have to defer to what's an appropriate adjustment on that because I think we'd have to go back and estimate to get an accurate or the exact correct estimate if that's the case.

            If there are other adjustments, I think we kind of have need to know because if that move to survival to three or four percent or adding two percent or one percent, I think it's worthy to know or at least me to know whether that -- it becomes an acceptable rate.

            There are a couple other comments that I think just in terms of -- well, I'll let it go in terms of potentially.

            DR. RORABECK:  Yeah, this is Cecil Rorabeck, and I just want to talk about your comment about randomized clinical trials because wearing my surgeon scientist hat -- and this is something we've been involved in for many, many years during randomized clinical trials of new surgical technology at our university, and we started initially doing cementless hips and with prospective blinded trials.

            They probably have here, and I'm not apologizing for the data.  The data is what the data is, but from a practical standpoint, and we have the same problem today doing this particular thing in our country, is that people are coming asking specifically for this procedure, and it's very hard to enter them.

            We currently have a randomized clinical trial going now comparing Birmingham to synergy Big Head Birmingham come, randomized, prospective, blinded trial, but the vast majority of our patients come because they want to have a Birmingham hip replacement.

            So, you know, you get into ethical issues and other issues, and I don't disagree with your comment.  That certainly is the gold standard.  Whether it's generalizable to all devices is another issue entirely, but it's difficult for us to be able to do this even though we're trying.

            PANEL CHAIRPERSON NAIDU:  Yes, Dr. Skinner.

            DR. SKINNER:  I'd like to ask Dr. Blumenstein or the statistician for Smith & Nephew.  It would seem to me that if a patient dies, he's no longer at risk for a revision.  So that would tend to make the survivorship better.  Is that wrong?

            DR. BLUMENSTEIN:  Exactly.

            MR. DeMUTH:  I think what it is is you've lost the potential that you're not going to catch an event.

            DR. SKINNER:  So that would only make the data look better.  Well, I mean, I think the point is that if you want to say we've lost potentially 27 potential events because a patient has died, then we may have some bias going the wrong direction in favoring the device as opposed to, say,  the worst case, which maybe you treat them all as failures.

            I feel like the truth is probably somewhere in between

            DR. BLUMENSTEIN:  Yeah.  I mean, the issue is that it's just the wrong methodology to estimate survivorship.  It is not going to make a lot of difference in these estimates if you use cumulative incidence, which is the proper methodology in the face of competing risk versus an inverted Kaplan-Meier or in this case I guess they use just a Kaplan-Meier, but it's just wrong, and you know, maybe I'm tilting at windmills, but it pervades the literature and so forth, and I've just got to say it's wrong every time I see it.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            You've heard that.  I think we should proceed by focusing our discussions on the FDA question at this point.  Copies of the questions are at the tables outside the room.  At this point I would like to ask Mr. Goode to come up and read the first question because there are fairly extensive questions, and we should -- we'll try to go around the room and discuss each question.

            MR. GOODE:  Dr. Naidu, would you like for me to present all the questions or just a single question and then proceed one by one?

            PANEL CHAIRPERSON NAIDU:  I think let's do one by one because these are fairly complex questions.

            MR. GOODE:  Question No. 1, please discuss the evaluation methods used to collect the safety data.

            So these are the collection methods used for the safety data, that is, how the data on revisions was collected, adverse events, deaths, and metal ion literature analysis, and whether or not these methods of data collection are reliable to assess the safety of the device.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Goode.

            Why don't we start with Dr. Mayor.

            DR. MAYOR:  I think we've covered the issue fairly thoroughly in the discussions we've already had, and my response to that would be to suggest that given the FDA's definition of valid evidence, it's valid but certainly far from impeccable.

            With that observation in mind, I think I would conclude that the methods for collecting safety data are  adequate.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Dr. Blumenstein?

            DR. BLUMENSTEIN:  Well, I can't feel good about the way the data have been collected, and in terms of the methods used to extract the data from the records, the lack of prospective design in data collection and so forth.  This falls far short of what a study should be.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mabrey.

            DR. MABREY:  I have to echo the sentiments of Dr. Blumenstein that while the data presented is a testament to Mr. McMinn's surgical skills and his clinical practice, much of this represents his data as opposed to that of a much larger group of surgeons.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Kim.

            DR. KIM:  I would just like to echo the concerns of Dr. Blumenstein and Dr. Mabrey, and I agree with their comments.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Skinner.

            DR. SKINNER:  Well, I have to agree with Dr. Mayor.  I think that it's an imperfect world, and it's imperfect data, but I think it's adequate.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Ms. Whittington.

            MS. WHITTINGTON:  I would concur with Dr. Blumenstein.  I don't think it's adequate, especially in the adverse events and the prediction of potential disasters for patients who might receive this implant.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Ms. Adams.

            MS. ADAMS:  Well, I feel a little differently.  Being on the side of -- I'm speaking as an industry rep.  I'm well aware, even though I don't work with these types of devices, that typically what happens is that 200 patients are studied for a period of one to two years, and based on that, these types of devices are approved.

            So I have a fair amount of comfort in the fact that there are thousands of cases here, and there's quite a significant amount of safety data, and although, no, it doesn't follow what we would call the perfect way of doing a clinical trial, which is randomized, controlled, blinded, all those kinds of things,it certainly fits to me within the purview of what the Congress has asked the FDA to do, which is to find the least burdensome way to bring devices to market.

            The concept in 1997, and I won't bore everybody with it, was that the act was amended so that we could have a least burdensome way.  So I will leave it at that and just say that to my way of thinking there is a large amount of data here for us to be able to get some sense of comfort about how it's collected.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Mr. Melkerson, in regards to Question 1, the panel is, as you heard, generally split.  There appears to be equal number of inadequate data positions, and there appears to be equal number that say that there is valid scientific evidence.

            And is that adequate?

            MR. MELKERSON:  Being that it is split, as far as adequacy, what would your comments be on this question?

            PANEL CHAIRPERSON NAIDU:  When you look at the valid scientific evidence as defined by the FDA, reports of significant human experience with a marketed device from which it can fairly and responsibly be concluded by qualified experts that there is reasonable assurance of the safety and effectiveness of the device under its conditions of use, in my opinion, there is a large series.  There's a large body of data in this study presented.  Even though it is retrospective, there has been a few other PMAs that have come up with retrospective data, and we have approved it.

            And I think there is enough here.  There is enough valid scientific evidence.  That's my opinion.

            MR. MELKERSON:  And then one follow-up question.  We were talking here not necessarily the quality of the data, but the question of the methods used to collect that data.

            And I heard one or two comments regarding that it wasn't prospectively designed, and I believe the -- and I ask the sponsor this -- I thought you had proposed ahead of time what that data collection was prior to sending the monitors out.

            MR. VALEZ-DURAN:  Just a couple of clarifications.  The first thing is that the Oswestry Center had to get this registry prospectively.  So the only thing retrospective was us going back and collecting that information, but prospectively the registry and the method to collect that data were set out prospectively in 1997.

            The safety data, there was a protocol that was set out for an independent group to collect all of the safety, any adverse event, or any indication in the patient records that there may have been an adverse event.  None of those were edited by the person that collected the information.

            In addition to that, the X-ray evaluation was done independently, and there was a protocol developed with success and failure criteria prespecified.

            MR. MELKERSON:  I think that's my comments.  Thank you.

            PANEL CHAIRPERSON NAIDU:  Have we adequately addressed Question No. 1?

            MR. MELKERSON:  I believe so.

            PANEL CHAIRPERSON NAIDU:  Mr. Goode, could you please pose the second question?

            MR. GOODE:  Question No. 2, please discuss, again, the evaluation methods or the way in which the data is collected for the effectiveness data, that is, the data on survivorship, and if you really again the survivorship in the same way that the safety data was collected was by the Oswestry Outcomes Center and by the McMinn Center.

            The OSHIP score, which again was a patient self-evaluation of pain and function information, the radiographic information was the zero and five-year evaluation of radiographs on a subset of the patients, as was the patient satisfaction information which, again, was through the mail-in questionnaire.

            That's what we're wanting you to comment on, those data collection methods.  Comment on whether or not those methods are reliable to assess the effectiveness of this device.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Goode.

            Why don't we start with Ms. Adams.

            MS. ADAMS:  Well, just one small comment.  I just wanted to comment that I liked the patient questionnaire even though I know that there were some felt that maybe that wasn't the best way to go.  I think some of the surgeons that implant these devices are very charming, and that when they ask questions of their patients, their patients want to make them happy, and so I think there's a potential for bias when it's a physician administered questionnaire.

            So just that is a comment.  I thought that that was an interesting approach.

            PANEL CHAIRPERSON NAIDU:  Ms. Whittington.

            MS. WHITTINGTON:  I concur with Ms. Adams.  In my experience of over 30 years of dealing with orthopedic patients, the best person to ask the outcome is the patient.  So I think that your choice of this method of evaluating the effectiveness needs to be from the patient's perspective because they're the recipient of that.

            PANEL CHAIRPERSON NAIDU:  Dr. Skinner.

            DR. SKINNER:  I also think the data is appropriately collected.  My Harris Hip Scores are collected by the MA by asking the patients the same questions, and I think that's the way it's done in most studies.  At best you'd have a physical therapist.

            So I think they were at least equivalent in this particular collection of data to the way it's collected in, say, either of those two ceramic hip studies.  So yes.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Dr. Kim.

            DR. KIM:  I would agree with the previous panel members that collecting the data from the patient is as important as collecting the data from other objective measures, including physical exam and radiographic findings.  So their method of data collection biasing the patients satisfies that criteria, but I would question the validity of the statement that the data was collected prospectively and retrospectively analyzed since 1997 because if that were the case, I would assume that the data would be collected in a uniform fashion.  You wouldn't have three cohorts.

            So I question the validity of that statement, and that's where I have some significant dilemma in answering this question of the methodology, not the actual contents of the data, but the method by which it was collected I do not agree was appropriate.

            And in addition, there's some issue -- I'm not a hip surgeon.  I'm a spine surgeon -- but there's some issue with the utilization of a relatively uncommon method by which to evaluate patient outcomes, essentially the OSHIP questionnaire, when it seems to me that the Harris Hip Score is the standard in this industry.  Correct me if I'm wrong.

            So there is significant flaws in the methodology in my mind, and we can talk about the actual data in Question 4, but that's my view.

            MR. VALEZ-DURAN:  If I could just make a few clarifications, and maybe I did not explain the cohorts well.

            The cohorts were identified simply to identify what group of patients had a specific set of data.  They're all part of the 2,385 patients.  So in a way, from my perspective, even though we comply with the request by dividing those patients into three cohorts, it's almost an artificial way of looking at these patients.  However, that's my opinion.

            What I would say is that the cohorts, for example, the X-ray cohort and the Oswestry cohort both receive the same questionnaire.  So there is no difference in there.  The only difference is in the what we call McMinn cohort, which is the most recent cases, and just to explain that a little bit, at the time of introduction of the product, the Oswestry Outcomes Center was commissioned to follow 5,000 cases, and all the funding for that was done up front, not based on the result of the registry.

            At the end, when they collected 5,000 cases, which is what we have provided in this PMA, there was no additional data collected by the Oswestry Center.  That's why there is a small group of -- there's a group of cases, the most recent, that do not have an OSHIP questionnaire.  I just wanted to clarify how that happened.

            DR. KIM:  Well, then I ask the question:  if you knew that this was a limited study to test out a new questionnaire system, why not have also, if this was something that you thought up ahead of time, to also include  Harris Hip Score and continue that throughout the entire study from 1997 through 2004?

            MR. VALEZ-DURAN:  Yeah, of course, I was not around in 1997 to make that decision, but the Oswestry Center evaluated the Harris Hip Score, evaluated the SF-36, the Womack and other questionnaires, and they decided to develop one that was a patient self-assessment because based on numerous publications the patient self-assessment perhaps is a better assessment of pain and function.

            And if we want a concise explanation of that at that time, we have Professor Richardson, which is the director of the Outcomes Center; he can provide an explanation for that.

            But certainly at the time the idea was not to come in to present and to compare a Harris Hip Score.  In fact, the scoring systems are very similar, and from the perspective of the Oswestry Center what was developed was better because it can be assessed by the patient, and it's also validated.

            If I could bring Professor Richardson to talk about that, Dr. Naidu, or is that sufficient?

            PANEL CHAIRPERSON NAIDU:  I think we'll just continue our deliberations at this point.  Thank you.

            MR. VALEZ-DURAN:  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Mabrey.

            DR. MABREY:  My answer has two parts to it.  Number one, I do think that ask the patient for a self-evaluation is appropriate, and that is the way to gather that data.  So to specifically answer that question, yes.  I think that was appropriate.

            However, I would like to add after seeing the presentations both in the FDA and from the sponsor today, it appeared as though a tremendous effort was brought forth in showing that the Oswestry was equivalent to the Harris Hip Score.  What I would have liked to have seen or I think most investigators would like to see would have been other scores by which to judge that, and those are readily available, the SF-12, the SF-36, the Womack from Canada, the Miltore Venier (phonetic), which is popular in Europe.

            It would have been nice to see some type of parallel score to go with that.  If there are plans for further data collection, then I would certainly put my suggestion in that some of those be included.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Blumenstein.

            DR. BLUMENSTEIN:  I would answer that there's adequate in the setting of an unmet medical need, but in the setting of comparing to other cohorts, there wasn't enough compatibility of endpoints or schedule of assessments or other things of that nature to allow even a cohort comparison.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mayor.

            DR. MAYOR:  I come from a hotbed of patient centered evaluation instruments at the Center for Clinical Studies at Dartmouth.  So I have no argument with the patient centered assessment instruments that were used here, but I would suggest that this Question 2 is very different from Question 1, and my response would cover not just Question 2 but Question 3 as well.

            While I have no argument that the methods of evaluation used to collect effectiveness data in the hands of Dr. McMinn and his operating skills demonstrates the effectiveness of the device in that context, I'm afraid I have no reassurance regarding the applicability of that result to the experience that we might  generate when the approach is generalized to the United States.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Mr. Melkerson, I'll summarize this question as follows.  With regards to Question No. 2, the panel in general believes that it is good to have the outcome scores directly from the patient, but nevertheless, the instruments of outcome used in this study are not very adequate.  It's not standard.  There have been references to SF-12, SF-36, Womack scores referred to us by Dr. Mabrey, and in general, it's good to have patient information rather than surgeon evaluation, but nevertheless, the instruments are lacking.

            MR. MELKERSON:  That's an adequate response, but I would again ask:  do you have any comments for the same question?

            PANEL CHAIRPERSON NAIDU:  I would concur with the panel with that.

            Mr. Goode, would you mind posing Question No. 3?

            MR. GOODE:  Question 3:  please discuss whether or not the foreign data from a single investigator in U.K. practice of medicine is applicable to the target U.S. population and practice of medicine.

            PANEL CHAIRPERSON NAIDU:  Let's start with Dr. Mayor.

            DR. MAYOR:  Again, as I just finished suggesting, it's not the issue of this being foreign data or the issue of the question of whether the citizens of the U.K. are a unique, strange or alien population, but it is the work of a single unit and an individual commission, and so I can't see within that data any reassurance that it's applicable to the U.S. population at risk.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Dr. Blumenstein.

            DR. BLUMENSTEIN:  My concern here is that we have no estimate whatsoever of the variability across surgeons or clinical sites.  All we have is the experience of the inventor.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mabrey.

            DR. MABREY:  I concur with Drs. Mayor and  Blumenstein.

            PANEL CHAIRPERSON NAIDU:  Dr. Kim.

            DR. KIM:  I also concur with the last three panel members.

            PANEL CHAIRPERSON NAIDU:  Dr. Skinner.

            DR. SKINNER:  I'm afraid I'm going to sound like an industry rep.


            DR. SKINNER:  If this study were going to be done in  a randomized controlled trial situation in the United States, the company would go out and find five excellent hip surgeons, and all of those hip surgeons would be completely different from the routine orthopedic surgeon who would do hip replacements after it was approved.

            The only difference I see between having one surgeon do this and having the five or six or eight other surgeons in the United States do it is you only have one learning curve instead of six of them.

            And I think that the population in the U.K. is very similar to the population we would have in the United States going to those six orthopedic surgeons.  It would be a referral practice of people going to get a particular procedure done by a particular surgeon.

            So I think that this data is perfectly applicable to a target U.S. population.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Dr. Whittington.

            MS. WHITTINGTON:  I'm going to have to agree with Dr. Skinner this time.  I would agree that it would be a referral practice, probably very, very much like the population that Dr. McMinn's practice is.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Ms. Adams.

            MS. ADAMS:  Well, it's clearly a weakness that there's one investigator involved.  I think that's obvious.  I'm not concerned about foreign data, and then when you get to the single investigator question I think the things that give me more comfort is the data that the sponsor presented indicating there were 140 surgeons who had conducted 3,374 cases.  I think that's good information.

            I think the extensive training plan that they've presented is very comforting, and the last piece that I'm comforted by is even though it may not necessarily be required, the sponsor has already offered to do an extensive post approval study, which I think would be a source of additional information that would be useful.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Mr. Melkerson, with regards to Question No. 3, again, the panel appears to be split, although there is a majority opinion stating that experience from a single surgeon from a large referral practice would not be applicable to the U.S. population in general.

            And did you want my opinion?


            MR. MELKERSON:  Of course.  See the pattern?

            PANEL CHAIRPERSON NAIDU:  Yeah.  I do have to concur with Dr. Mayor, Dr. Blumenstein, Dr. Mabrey, and Dr. Kim.  This would not be applicable to the general practice of orthopedic surgery  in the U.S. population.  This is a single surgeon study.  That is my opinion.

            MR. MELKERSON:  Thank you.  That's an adequate response.

            PANEL CHAIRPERSON NAIDU:  Let's move on to Question No. 4.  Mr. Goode, would you please post the question?

            MR. GOODE:  Question 4:  based upon the safety data in the 2,385 patients in the overall McMinn cohort, that is the data on revisions, adverse events, and deaths, and the analysis of the metal ion literature.  Please discuss whether or not you believe that the data contained in this PMA provide a reasonable assurance of safety.

            PANEL CHAIRPERSON NAIDU:  Why don't we start off with Dr. Mayor again?

            DR. MAYOR:  Given that no reassurance of safety is going to be perfect, my response would be that the data presented in this PMA provides a reasonable assurance of safety.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Dr. Blumenstein.

            DR. BLUMENSTEIN:  I concur that the data shown on safety is adequate when referred to the amount of efficacy demonstrated.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mabrey.

            DR. MABREY:  I concur.

            PANEL CHAIRPERSON NAIDU:  Dr. Kim.

            DR. KIM:  I concur as well.  Despite the flaws in the methodology of obtaining the data, the data is robust in number and significantly -- provides a significant amount of data quantity to assure that this device is likely safe.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Skinner.

            DR. SKINNER:  I agree with Dr. Kim.  I think that the large number of patients, the lengthier follow-up than would be done in a 100 total hip in one group, 100 total hip in the other group study for two years that you see in this group is definitely enough to show that it's reasonably safe.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Ms. Whittington.

            MS. WHITTINGTON:  I concur with Dr. Kim and Dr. Skinner.  I think the robustness of the data that was presented is good.

            PANEL CHAIRPERSON NAIDU:  Ms. Adams.

            MS. ADAMS:  I agree.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Mr. Melkerson, this is one of the few questions where we have reached a unanimous okay.  The panel, in general, believes that this PMA does provide a reasonable assurance of safety.

            Did we address your concern adequately?

            MR. MELKERSON:  Yes, it did.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            MR. MELKERSON:  But I'll give you the opportunity if you want to put your two cents in.

            PANEL CHAIRPERSON NAIDU:  I do.  I will echo the opinion of the panel.  Thank you.

            Mr. Goode, could we move on to Question No. 5, please?

            MR. GOODE:  Question 5:  based upon the five-year survivorship analysis of the 1,626 procedures in the combined X-ray/Oswestry cohort, the five-year radiographic data on the 124 procedures in the X-ray cohort, the five-year pain and function OSHIP data of the 1,111 unilateral procedures in the X-ray/Oswestry combined cohort, and the five-year patient satisfaction analysis of the 1,626 procedures in the X-ray/Oswestry combined cohort., please discuss whether or not you believe the data contained in this PMA provide reasonable assurance of effectiveness.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Goode.

            Why don't we start with Ms. Adams?

            MS. ADAMS:  Well, I think the numbers are pretty good.  We're seeing things that are comparable to -- they did a good job of comparing to other devices, radiographic success, survivorship, adverse events even.  I think the numbers look very good.  So I have no additional comments.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Ms. Whittington.

            MS. WHITTINGTON:  I would concur with Ms. Adams.  I think the data supports it.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Dr. Skinner.

            DR. SKINNER:  I think the data provides reasonable assurance of effectiveness.

            PANEL CHAIRPERSON NAIDU:  Dr. Kim.

            DR. KIM:  I still have difficulty with this question, although the robustness of the data, the sheer numbers, makes me feel comfortable as to the safety of this device.  The methodology is significantly flawed such that I cannot confidently determine if this device is more or less effective than the existing treatment that's available.

            So I would say that it lacks enough information to make that determination.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Mabrey.

            DR. MABREY:  If I could just make a comment first.  I also served as a reviewer for the Journal of Arthroplasty, and I have to say that having reviewed all of the data today and having seen all of the presentations that being here today is like reading one of the very best written papers for the Journal of Arthroplasty.  The only difference is that the author is sitting right there and you can put him on the spot at any time.

            I think the data here is testament to the skills, surgical skills and design skills, and I think that it provides a reasonable assurance of the effectiveness of this device, and that I feel that it's similar to other devices that have already been used widely throughout the rest of the world.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Blumenstein.

            DR. BLUMENSTEIN:  Well, as I said in my presentation, I think that there's adequate evidence of efficacy against an unmet medical need, but where I have a problem is whether this study fits the definition of a well controlled study in light of the fact that there are predicate interventions in this disease, and I don't think it serves the public well when we can't compare the device under consideration to the predicate devices.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mayor.

            DR. MAYOR:  Well, my response may seem at odds with what I've said earlier.  I think bullet points one, two, three, and four do provide reasonable assurance of effectiveness even in the hands of a well qualified and expert orthopedic surgeons in this country and referral practices as is likely to be the case with the application of this technique to those individuals.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Mr. Melkerson, with regards to your Question No. 5, it appears that the panel is in general consensus that the device is effective, although there are concerns raised again.  The theme has been repetitive throughout all of the discussions.  The effectiveness apparently cannot be judged adequately, namely, because of the single surgeon experience and the retrospective analysis.

            Have we addressed the question adequately?

            MR. MELKERSON:  Yes, you have, and also, again, I'd like to ask your opinion as well.

            PANEL CHAIRPERSON NAIDU:  My opinion is that the device is effective, but the data quality is deficient.

            Thank you.

            Mr. Goode, would you mind going to Question No. 6?

            MR. GOODE:  Question No. 6:  do the patients' selection methods and the data presented on the BHR device support the proposed labeling indication?

            And also, please comment on any other aspects of the product labeling, such as the contraindications, warnings, precautions, potential adverse effects on health.

            PANEL CHAIRPERSON NAIDU:  Again, Dr. Mayor, if you don't mind starting off the discussions of this question for us.

            DR. MAYOR:  I think the contraindications have been covered adequately as written.  I think the warnings related to patient selection and potential adverse effects on health deserve extra emphasis in regard to the expectation of future difficulties handling metal ion release into the patient system with particular emphasis on the possibility of anticipated decline in kidney function.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Dr. Blumenstein?

            DR. BLUMENSTEIN:  No comment.

            PANEL CHAIRPERSON NAIDU:  Dr. Mabrey?

            DR. MABREY:  I agree with Dr. Mayor.  I think that the contraindications have been well covered.  I don't think there's anything else you can do about warning people against the fact that their femoral neck may fracture other than to advise and to pick a very good surgeon, and I think the precautions are appropriate.

            Most of the precautions and warnings tend to apply to that of metal ions, and this is a more general discussion.  I would defer to Ms. Adams in terms of where the warnings and contraindications have gone with respect to other metal-on-metal devices within the FDA.

            But at this point I would agree that they've been well identified.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Kim.

            DR. KIM:  I would agree with Drs. Mayor and Mabrey, and make a special emphasis that if it's not already in the labeling insert, indicate that we do not know what the relationship between metal ions and adverse health is at the moment.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Skinner.

            DR. SKINNER:  I agree with the previous panel members except Dr. Blumenstein.


            DR. BLUMENSTEIN:  I said no comment.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Ms. Whittington.

            MS. WHITTINGTON:  I would agree with the other panel members, I guess, except Dr. Blumenstein as well, and ask that because you have a distinct and separate section for patient labeling that that section really be screened and effective for the patient to read and not for a physician to read to the patient because that never happens.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Ms. Adams.

            MS. ADAMS:  I agree with Dr. Kim's comments, as well as Ms. Whittington's.

            I have two other comments.  One is I'd like to say that I think the sponsor did a pretty good job handling the fact that they had to parse their way through the patient selection methods by one investigator and come up with a proposed labeling indications.  I think they did a pretty good job.

            The other thing that I would just like to remind the sponsor is that we did make comments previously about some other considerations that should be made with FDA having to do with other items, such as GFR, creatinine and that sort of thing, that should be considered in the labeling.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Mr. Melkerson, with regard to Question No. 6, the panel in general believes that the data presented on the BHR device does support the proposed labeling indication.  There are some concerns, however, with regard to metal ion release and renal insufficiency  in patient populations such as pre-diabetes, hypertensive patients with renal failure.  Those things have to be obviously refined.

            Did we adequately address your concerns?

            MR. MELKERSON:  Yes, you did.  And again, I would like your endeavor to give us your opinion as well.

            PANEL CHAIRPERSON NAIDU:  My opinion is similar to the panel's conclusion.  Thank you.

            Mr. Goode, let's move on to Question No. 7.

            MR. GOODE:  Question 7:  a reasonable assurance of safety and effectiveness as defined in Questions No. 4 and 5 above must be demonstrated for device approval.  If you believe the data in this PMA demonstrate a reasonable assurance of safety and effectiveness but think that there are remaining specific questions regarding this device that should be addressed in a post approval study, please identify those questions.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Goode.

            This is a loaded question.  I'd like Dr. Mayor to start off again, please.

            DR. MAYOR:  I would urge a five and ten-year interval of post market approval supervision if this device is approved for release in the United States, with specific attention to femoral side complications, including substance of the femoral component on the femoral head, and surveillance to detect the effect of particulates and bioreactivity in the region of the hip joint itself.

            I would also urge the establishment of a five-year follow up for the cadre of patients treated by the instructor group described by Dr. Rogerson.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Dr. Blumenstein.

            DR. BLUMENSTEIN:  I think the study as proposed would be just another example of a poorly controlled study, and it would pollute the literature, and I would not be recommending such a thing be done as a post approval study.

            I think that if a post approval study is done that it should be well controlled and specifically a randomized clinical trial providing adequate data with respect to other predicate devices.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mabrey.

            DR. MABREY:  This particular device does have the potential of failure at a much later time, as patients age, become more osteoporotic.  So I concur with Dr. Mayor that a longer period of time directed follow-up be applied.

            I'd be interested to see what happens to these people at ten years as they age to see if there's any stress shielding around the stem, but I do appreciate the sponsor's offering of a post market approval study.

            I do think that there's an opportunity here, especially at the champion surgeons' sites to initiate some controlled studies if that's possible, and I want to encourage that.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Kim.

            DR. KIM:  If this device is approved, I would strongly recommend some conditions.  I have two concerns.  The first is to identify if there are any major learning curve problems.  In other words, if we have 100 surgeons during the first five procedures, that's 500 procedures.  I think that's enough to identify a significant learning curve problem.

            Therefore, I would recommend that one of the conditions be that it's released to a limited number of people/centers, and that the surveillance is on the order of a clinical trial.  Of course, a randomized controlled clinical trial would be optimal, but if it is a longitudinal prospective study only, that protocol should be agreed upon with the FDA to include all of the inclusion/exclusion criteria, the criteria for revision, a standardized follow-up protocol along with agreed upon measures of success.

            I can just go on and on, but just suffice it to say a well designed study for a period of some time.

            And one final thing.  On Slide 38, the conditions of approval, the sponsor recommended 150 patients at up to 15 sites.  That doesn't seem like a lot of patients, but one cannot create that number until you do a power analysis.  So that number should be revisited as well.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Skinner.

            DR. SKINNER:  Well, I think Dr. Kim's suggestion of looking into the learning curve is interesting,b ut I think we can almost guarantee there will be a learning curve.  It's just a matter of how bad it's going to be.

            But I think Dr. Blumenstein really has the right answer here.  I think a post marketing study of any significance, of any size that would be significant to get a randomized controlled trial would simply be another PMA type of thing, and I don't think that's the information we want.

            That information won't come down the line until five years, ten years later.  I think if we were going to request the company to spend that money, we should ask them to take Dr. McMinn's first 200 patients and follow them for another five years and see what happens to those patients as more information sooner about the long-term effects of this prosthesis.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Ms. Whittington.

            MS. WHITTINGTON:  I think that there definitely needs to be a post market approval process and it definitely needs to be well controlled, but I also agree that taking the initial patients from Dr. McMinn's patient population and tracking them along with those same criteria that are defined in this well controlled post market approval survey that's done, with special attention to addressing those adverse events and identifying the criteria so we have a better handle on them, I think that's what I have the least faith in what we've seen today, is how that was collected and how it was defined and identified.  So that goes along with a well controlled study.

            I also think a limited number of practitioners may be a prudent thing to do, given the difference in the technique with this procedure versus some of the techniques that we currently are using.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Ms. Adams.

            MS. ADAMS:  I do believe that the data in the PMA demonstrated a reasonable assurance of safety  and effectiveness.

            I agree with Dr. Skinner that there would be real value in following some of the earlier patients out to later dates.  I disagree with Dr. Kim that 150 patients in a randomized controlled trial will provide significantly more or different information than we have today.  I think that's unlikely.

            I also think, even though I don't typically design clinical trials myself, that having this device on the market and people being interested in coming and asking for it provides some real problems to the sponsor in terms of randomizing and how the clinical trial is handled.

            And finally, I'd like to suggest that the sponsor and FDA take into account our own comments here today and work up a post approval study that would address them in the most effective way.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Mr. Melkerson, the answer to this question is all over as you have heard.  There is a general consensus that the panel in general wants a fairly extensive post market surveillance study.  It appears that we're back to randomized controlled trials, and it appears that the panel does not believe that there is a reasonable appearance that based on this submission, this data alone, we can answer this question adequately.

            Does that generally answer?

            MR. MELKERSON:  Just for clarification, can you go over again the specific questions?  I heard long term.

            PANEL CHAIRPERSON NAIDU:  Yes.  Dr. Mayor suggested five to ten-year post market approval study including surveillance, all femoral components, five-year follow-up of metal ion release.

            Dr. Kim suggested additional randomized controlled trials of 150 patients.

            Dr. Mabrey thought that post market approval would be appropriate.

            Dr. Blumenstein plain ol' said, "No need for that.  Just do a randomized controlled trial."

            Dr. Skinner went on to say that it would be nice to follow 200 initial patients of Dr. McMinn.

            DR. SKINNER:  But not do a randomized controlled trial.

            PANEL CHAIRPERSON NAIDU:  Not do.  Okay.

            DR. SKINNER:  Because we'd be getting the data from five years out already.  That would be interesting data.  If we do a randomized controlled trial, we don't have that data for five more years, even the five-year data.

            So I think that randomized controlled trial is a waste of money, to be honest.

            PANEL CHAIRPERSON NAIDU:  Okay.  Ms. Whittington wished to have a more controlled PMA.

            And the only one with full support is Ms. Adams.

            Yes, Dr. Kim.

            DR. KIM:  Point of clarification.  The number for 150 that was mentioned, that's the number that the sponsor had proposed.  My point was that that 150 number is a number that has no basis and that we need to identify what that proper number is if we're going to do a post market surveillance study.

            It could be more than 150.  It could be less or it could be 150, but that number was derived without any basis.  I just wanted to make that clarification.

            MS. ADAMS:  May I make a comment?


            MS. ADAMS:  There is a basis, and as we talked about before, these types of devices are typically studied in 150 to 200 patients  with two-year follow-up.  So typically that is what you would see.

            What we have here is something far in advance of that already, and we're asking the sponsor to go on and do that in addition.  So I think we ought to keep that in mind.

            MR. MELKERSON:  I think that answers our question, but again I would like to ask your opinion on this.

            PANEL CHAIRPERSON NAIDU:  Yes.  My opinion is that I think there is a reasonable assurance for safety and effectiveness in this PMA that's presented based on the single surgeon experience retrospective of you, but as the other panel members have already cited, post market study is important with regards to metal ion release, failure of devices, and Dr. Skinner's suggestion of following 200 patients from the initial quote of Dr. McMinn would be a valuable addition to the study.

            Have we addressed all of the panel questions adequately, Mr. Melkerson?

            MR. MELKERSON:  I believe so.

            PANEL CHAIRPERSON NAIDU:  At this time we'll adjourn the meeting -- well, not quite.  We'll take a break.


            PANEL CHAIRPERSON NAIDU:  Let's take a short break and we'll come back.

            (Whereupon, the foregoing matter went off the record at 3:26 p.m. and went back on the record at 3:40 p.m.)

            PANEL CHAIRPERSON NAIDU:  Now that the panel has responded to the questions, we will open our second open public session.  Does anyone here wish to address the panel now?  If so, please come forward to the podium and state your name, affiliation and indicate your financial interest, if any, in the device being discussed today or any other device.

            MR. THOMAS:  Hello.  My name is Craig Thomas.  I'm an orthopedic surgeon in Washington, D.C.

            I have no financial interest in this device or Smith & Nephew.  I'll just give you a brief history of my background and why I have an interest in this hearing.

            I did my residence in orthopedic surgery and did a fellowship with Michael Mont (phonetic) at Sinai Hospital in Baltimore, Maryland.  For that reason I have about one year and six months experience clinically with metal-metal resurfacing, not specifically this device.

            I just have some answers to some of the questions that  panel had that were not provided or were not as clear and so I just want to address them.  The first one was learning curve, surgical technique; what is the learning curve; are there any studies?    Is there any data that we have to offer?

            Well, at this year's academy meeting for orthopedic surgery, I believe it was Poster 410.  Michael Mont and myself presented a paper that addressed the learning curve of resurfacing arthroplasty. 

            One of the papers, there was a three-year follow-up, and just to summarize it briefly, with our first 50 patients, we had 11 thermal neck fractures, which was significant.  The second 50, we had one.  Then after that 100, from one to 200 we had a zero fracture rate.

            So there is a learning curve, but we learned from our learning curve, and we've learned how to avoid that, and I think that as different companies educate different surgeons how to do the procedure that you won't have what we had in the first 50.  And this is in the United States in Baltimore.

            So to address that I can't predict what it's going to be, but there is a learning curve, and I think we can make it better.

            The other thing that some folks had a concern about was a limp.  That's hard to contribute to this actual implant device.  Most total hip patients will have a limp, both preop and postop, and it does go away, and sometimes it doesn't, and that's dependent on several factors.

            You had a limp before surgery.  The surgical approach that was used, the type of repair that was used, the patient's motivation, the physical therapist.  It has a number of factors, and also at this year's academy we presented a paper that showed that a -- a gait analysis paper with resurfacing patients compared to standard total hip patients -- that showed that the resurfacing patients had a near normal gait pattern, and this was evaluated in a gait lab at the Rubin Institute for Advanced Orthopedics.

            Again, I believe it's poster 410.  I'm not sure of the other one, but you can look up Mont, et al.  We had a total of 18 presentations at the academy, and it would be in one of those on the academy Website.

            Let's see.  There was a question, I believe, in reference to why not patients who are older than 70, and I think that's a good question, and it's going to be surgeon-patient dependent based on the integrity of your bone structure whether or not it would be offered.  And I think it would be a mistake to put an age limit on that number.

            Let's see.  The next point.  Someone asked about incision size because there is sort of I would say patient and industry driven desire to have smaller incisions.

            Well, also at this academy we presented a paper on resurfacing in general where we took the size of the incision from 11 centimeters to six centimeters which is two and a half inches.  So if you pull out your car keep -- now the car keys are bigger -- but the standard house keep and make that one and a half, that's about two and a half inches.  That's a fairly small incision, and this is for a hip resurfacing. 

            I do agree that this is not something that we start out training folks how to do minimally evasive resurfacing arthroplasty, but it can be done.

            Patient labeling was a question that was addressed about renal patients.  Preoperative lab work is obtained, and that identifies that.  I think that was appropriately addressed.

            The other question was about avascular necrosis and size, size of the lesion, and contraindications.  I think that that should be a discussion that's left between the patient and doctor as we review radiographs and MRIs and talk about whether or not you're a candidate for it.

            And actually what we did at the Rubin Institute is a lot of times that's an intraoperative decision.  The patient understood that if we interoperatively could not or did not feel confident putting a femoral head resurfacing on that they would be converted to a total hip replacement, and I think that that should be preserved.

            Let's see.  The one thing that is sort of bothersome to me, but I understand, is the analysis of some of the data here.  In my opinion we're being very critical of some of the data.  I could take any one of the members of this panel tomorrow if you meet the preoperative criteria and give you both components separately.

            So the acetabular component, I can use that part as a total hip replacement, as an FDA approved device.  The femoral component, I can cap your femoral head as an FDA approved device.  I just can't put the two together.

            And when you put the two together, you have the metal-on-metal articulation, which I do that already, but with a stem.  So now I'm taking the femoral part, putting a stem on it.  You still have the metal-on-metal component, basically the came component that you guys are looking at, with a stem.

            So when you're talking about metal ions and everything, everything that you're really focusing on in a negative mode is the articular interaction, but I could do that tomorrow FDA approved.

            So if I can do that and if this does not get approved, then you may want to strongly look at what's already on the market and pulling it.

            So I just wanted to address some of those concerns.  I would take any questions from anyone.  I was not prepared to even talk to you guys.  I didn't even know this was happening until this week.  I happened to have a dental appointment that I canceled once I found out that this was here.  So thank you for that.


            DR. THOMAS:  And I stayed longer and longer, I just had more things that I thought should be addressed or at least explained to you.

            My patient population cannot afford to go overseas and spend the 15, 20, $35,000 that's required to do this.  I have a list of 55 patients that have been waiting for this to be approved.  About five of them I've had to recently go ahead and give them a hip replacement because it's affecting their life.  The pain that they're going through and the wait and the wait and the wait.

            So I just give you this information.  Does anybody have any questions for me?  I will take any.

            That's it.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Thomas.

            Any questions from the panel?

            DR. MAYOR:  It's not exactly a question.  You may have missed the CBS Sunday Morning broadcast last weekend.  They had a feature from India, and there was an Indian surgeon who has been trained to do this procedure, and the patient went from Vero Beach, Florida to India and back, spending about $6,000 total for the surgery, entire hospital stay, and three weeks in a beach side resort about an hour from the hospital.

            DR. THOMAS:  Maybe I'll go get my hip done.


            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Thomas.

            Is there any further comment or clarification from the FDA?  mr. Goode?

            MR. GOODE:  No, sir.

            PANEL CHAIRPERSON NAIDU:  Thank you. 

            Is there any further comment or clarification from the sponsor?  Mr. Duran?

            MR. VALEZ-DURAN:  No, there's no additional clarifications.  We want to thank the panel for your time and effort.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            We're now ready to vote on the panel's recommendation to the FDA for this PMA.  Ms. Scudiero will now read the panel recommendation options for the PMA.

            Ms. Scudiero.

            MS. SCUDIERO:  These are the panel recommendation options for premarket approval applications.

            The Medical Device Amendments to the Federal Food, Drug and Cosmetic Act, as amended by the Safe Medical Devices Act of 1990, allows the Food and Drug Administration to obtain a recommendation from an expert advisory panel on designated medical device premarket approvable applications that are filed with the agency.  The PMA must stand on its own merits, and your recommendation must be supported by the safety and effectiveness data in the application or by applicable publicly available information.

            Safety is defined in the act as reasonable assurance based on valid scientific evidence that the probable benefits to health under conditions of intended use outweigh any probable risks.

            Effectiveness is defined as reasonable assurance that in a significant portion of the population the use of the device for its intended uses and conditions of use when labeled will provide clinically significant results.

            Your recommendation options for the vote are as follows:

            One, approval if there are no conditions attached.

            Two, approvable with conditions.  The panel may recommend that the PMA be found approvable subject to specified conditions, such as physician or patient labeling, education, labeling changes, or further analysis of existing data.  Prior to voting all of the conditions should be discussed by the panel.

            Three, nonapprovable.  The panel may recommend that the PMA is not approvable if the data do not provide a reasonable assurance that the device is safe or that the data do not provide a reasonable assurance that the device is effective under the conditions prescribed, recommended, or suggested in the proposed labeling.

            Following the voting the Chair will ask each panel member to present a brief statement outlining the reasons for their vote.

            PANEL CHAIRPERSON NAIDU:  Are there any questions from anyone on the panel about these voting options before I ask for a main motion on the approvability of this PMA?  Dr. Mayor.

            DR. MAYOR:  A minor question related to the last sentence in the preamble.  Effectiveness is defined as reasonable assurance that in a significant proportion of the population use of this device for its intended uses and conditions of use when labeled will provide clinically significant result.

            And I suggest that we really are looking for more than just that, but a clinically significant beneficial result.

            MS. SCUDIERO:  I believe that is what is intended even though that is not stated.  That's a good comment.

            DR. MAYOR:  Okay.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Is there a motion for either approvability, approval with conditions or disapproval from the panel?

            DR. MABREY:  Mr. Chairman, I'd like to make a motion.  I move that the panel approve with conditions, option number two.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Is there a second for the motion?

            DR. MAYOR:  Second.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Now, Dr. Mabrey, since you did state that this is approvable with conditions, would you like to introduce the first condition?

            DR. MABREY:  Yes.  First, that there be a post approval study.   I realize that that's already presented in the material that we have, but I'd like to just for the record be assured that there is a post approval study, and I'm comfortable with the way it's presented on page 20 of our handout, although I would make one modification as suggested by Dr. Mayor that we have some radiographic follow-up at ten years as well.

            I don't think we need radiographic follow-up at years six, seven, eight, and nine.  I try not to bring all of my patients back to the office that frequently either, but because this has a particular propensity for failure and we don't know what the long-term failure results would be, I would add radiographic follow-up at ten years.

            PANEL CHAIRPERSON NAIDU:  Is there a second for this modification?

            DR. MAYOR:  I would concur.

            PANEL CHAIRPERSON NAIDU:  The motion on the floor right now is to approve with condition, the first condition being radiographic follow-up at ten years as one of the prerequisites.

            Is there any discussion on this condition?  Dr. Skinner?  Oh, Dr. Blumenstein.

            DR. BLUMENSTEIN:  I don't know whether I need to make this as another motion or ask for modification, but the size of the study should be based on statistical principles and criteria for success.

            PANEL CHAIRPERSON NAIDU:  Dr. Mayor, is that acceptable?

            DR. BLUMENSTEIN:  I would concur.

            Mr. Melkerson?

            MR. MELKERSON:  Just a point of clarification.  You said vote for approval.  You're only voting on this condition.

            PANEL CHAIRPERSON NAIDU:  That's correct.  We're just going on the conditions so far.

            MS. ADAMS:  And can you repeat the condition that we're considering?

            PANEL CHAIRPERSON NAIDU:  The condition is to have a post market approval study with a radiographic follow-up of ten years.  That is the condition on the floor right now.

            Dr. Blumenstein has added another condition that the size of the study be statistically significant.

            Is there a second for that motion?

            MS. SCUDIERO:  Wait.  Is Dr. Blumenstein's comment, is that meant to be added as part of the description of the post approval study?

            DR. BLUMENSTEIN:  If it isn't added, then I'll make it as a separate motion.

            MS. SCUDIERO:  Okay.

            DR. BLUMENSTEIN:  So I don't care.

            MS. SCUDIERO:  I believe that we can sort of like friendly amendments to a condition if it's agreeable with the person who made the motion and he who seconded it.

            DR. MABREY:  I'm very agreeable.

            MS. SCUDIERO:  That's good.

            DR. MAYOR:  As am I.

            DR. MABREY:  Just ask my wife.

            PANEL CHAIRPERSON NAIDU:  Great.  So the motion, again, the condition for motion, the modification condition is post market approval study with the radiographic follow-up at ten years, with the size of the study to be determined.

            DR. SKINNER:  Statistically.

            PANEL CHAIRPERSON NAIDU:  Statistically significant.

            Is there discussion on this motion?  Dr. Kim?  Dr. Skinner?

            DR. SKINNER:  I thought I was agreeing with Dr. Blumenstein earlier when he said that he thought it was unlikely that a study would be very valuable unless it was a very large study.

            I think that any information we get from such a study is going to come out after we've either abandoned the procedure or we've already decided it's a great procedure.

            If we wait ten years we're going to be past.  We're not going to have any information to really derive from that.  If we want ten-year data, we should get the data from Dr. McMinn's study group that already has five years in.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            DR. KIM:  Can I make a comment to that effect?


            DR. KIM:  On this subject.  I would agree there are actually two issues here.  The one is it's long-term efficacy, and I would agree to have to do a randomized controlled trial to look at that specific question would be too burdensome.

            But there's a main question that's still open, and that is its short-term safety and efficacy in the hands of multiple surgeons at multiple sites.  That has not been clearly shown, and I don't think it is fair to just look at Dr. McMinn's first 200 since he may have done this type of surgery before.

            So I think the utility of a randomized controlled trial at least the way we do it at the FDA is it only lasts for two years, and it's really to look at big, egregious, early problems that we can identify and address prior to releasing a device out into the general public.

            So I guess what I'm trying to say is I would agree with Dr. Blumenstein that a post market study needs to be done.  It needs to be a good study based on sound study principles and statistics.

            PANEL CHAIRPERSON NAIDU:  Dr. Mabrey, do you have something to add?

            DR. MABREY:  Yes.  If I could just comment on Dr. Skinner's analysis.  I was under the assumption that the sponsor had already agreed to perform this post approval study and would be collecting at least questionnaires at ten years; is that correct?

            And I see nodding heads out there.  So I'm just suggesting that while you're collecting the questionnaires at ten years that I think it would be useful to see what the femoral neck at least looks like at that tome.

            MR. VALEZ-DURAN:  Understood.

            PANEL CHAIRPERSON NAIDU:  Thank you.

            Dr. Blumenstein, you had something to add?

            DR. BLUMENSTEIN:  Well, I was just going to mention that you could put another condition on the approval for the long-term radiographic follow-up of patients already in the study, as I intend to do for a randomized clinical trial.

            DR. SKINNER:  I'll raise you three aces.


            PANEL CHAIRPERSON NAIDU:  Yes.  Do you want to vote on the condition?

            MS. SCUDIERO:  Dr. Mabrey, would you like to restate your condition?  Maybe we need to go back to square one so that we don't get confused?

            DR. MABREY:  My condition is that in addition to the post approval study proposed by the sponsor, as outlined on page 20, that in addition to collecting clinical data at ten years, that we collect radiographic data at ten years.

            I don't mean to imply that we should wait those ten years before we finally bring this device to market.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Ms. Adams.

            MS. ADAMS:  Well, I would just like to make a comment to Dr. Kim's point.  He's said a couple of times and I've even been convinced of it that we don't have long-term data, and in the break I spent some time revisiting it.

            I think we should keep very much in mind that we've heard that there have been 33,000 implants in 23 countries.  There were 140 surgeons that were included for 3,300 cases with five-year follow-up.  That's pretty significant.

            So I want to be very cautious about implying that we only have a short period of time of information here and pressing upon the sponsor that there should be some significant amount of study added to them.

            I think they've been gracious in indicating that they would be willing to do additional study, but I think that that would be a very high bar compared to what we typically see and maybe not necessary.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Adams.

            Any other comment?  Is Whittington?  Dr. Skinner?  Dr. Kim?  Dr. Mabrey?  Dr. Blumenstein?  Dr. Mayor?

            Okay.  So the first condition for the post market study is in addition to the post market study proposed by the sponsor that there be long-term clinical data at ten years and also X-ray data at ten years.  Shall we vote on this condition?

            DR. MABREY:  And statistical principles used to determine the study size.

            PANEL CHAIRPERSON NAIDU:  And also statistical principles used to determine study size.

            MS. ADAMS:  That's a lot of conditions.  As a point of procedure -- and I'm new.  We just trained yesterday -- we were told that the idea was to have a condition and vote on the condition, and I think we've got three, or at least two with a statistical issue.

            PANEL CHAIRPERSON NAIDU:  Right.  There are two ways to address this long-term data.  One is to follow Dr. McMinn's original cohort all the way to ten years and report on that data or take Dr. Blumenstein's condition that there be a new cohort established, statistical significant sample size.

            Mr. Melkerson?

            MR. MELKERSON:  I would actually vote on your proposal from Dr. Mabrey, which is basically study as proposed, ten-year radiographic, and the statistical sample size.  The issue of can you address that by other comments from Dr. Skinner may be another issue for a motion and you're going to vote it up or vote it down.

            PANEL CHAIRPERSON NAIDU:  Okay.  Why don't we vote on this condition then?  Post market approval study as indicated by the sponsor plus radiographic follow-up and clinical follow-up for ten years which is statistically significant.  Let's have a vote on that motion.

            Dr. Mayor?

            DR. MAYOR:  I would vote affirmative.

            PANEL CHAIRPERSON NAIDU:  Dr. Blumenstein?

            DR. BLUMENSTEIN:  Yes.

            PANEL CHAIRPERSON NAIDU:  Dr. Mabrey?

            DR. MABREY:  Yes.

            PANEL CHAIRPERSON NAIDU:  Dr. Kim?

            DR. KIM:  Yes.

            PANEL CHAIRPERSON NAIDU:  Dr. Skinner?

            DR. SKINNER:  Yes.

            PANEL CHAIRPERSON NAIDU:  Ms. Whittington?

            MS. WHITTINGTON:  We're not voting.

            PANEL CHAIRPERSON NAIDU:  Okay.  I'm sorry.

            There is a unanimous consensus from the voting panel that the post market study be performed as drafted in the original PMA, plus clinical data and X-ray data from ten years be reported based on sound statistical principles.  Is that acceptable?

            MR. MELKERSON:  Yes.

            PANEL CHAIRPERSON NAIDU:  Is there a second condition that anybody else would like to introduce?  Dr. Blumenstein.

            DR. BLUMENSTEIN:  I would like to make as a condition of approval the conduct of a randomized clinical trial that would help establish the relative efficacy of this device with respect to the other predicate devices based on sound statistical principles.

            PANEL CHAIRPERSON NAIDU:  Mr. Melkerson?

            MR. MELKERSON:  Point of clarification.  A new study would not be supported by an approval.  If you need new clinical data to support efficacy, that is actually a reason for not approving a product.

            PANEL CHAIRPERSON NAIDU:  Thank you. 

            So is anybody to second this motion made by Dr. Blumenstein?  He wants a randomized controlled trial study based on a comparison to a predicate device.  Is anybody to second that motion?

            (No response.)

            PANEL CHAIRPERSON NAIDU:  Since I see no second for this condition, do we have any other outstanding conditions that we would like to make?

            DR. KIM:  Can I ask a question about a condition?  Can you make a condition where the sites are limited to that group of centers that would be doing the post market surveillance and not be considered a new study?

            So my motion would be to limit the release of this product to a select number of sites to be determined based on the statistical need of the numbers of the patients to look at this post market approval study.

            I don't know if I even understand that.


            DR. SKINNER:  Could I comment on that?

            DR. KIM:  Yes.

            DR. SKINNER:  If this device is as popular as it has been implied, you're basically giving a license to print money to five or six or whatever sites.  I don't think you want to go there.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            DR. KIM:  Well, Mark, can you elaborate on whether or not you would consider that a condition that's equivalent to a nonapprovable recommendation because it's a study?

            MR. MELKERSON:  This is Mark Melkerson.

            I wouldn't answer the question that way.  If we are approving a product, we are approving it for distribution.  You can put limitations on that, but in terms of a limited distribution, most restrictions on conditions of approval are related to potentially training or other methods, but in terms of approval, we are approving it for marketing in the U.S.

            DR. KIM:  Then I retract that condition.

            PANEL CHAIRPERSON NAIDU:  Dr. Mabrey?

            DR. MABREY:  Oh, before Dr. Kim retracted his condition I was going to suggest that the sponsor clarify their roll-out plan.  It sounded like they were going to restrict it to 15 champion surgeons at restricted sites for at least -- it sounded like at least the first 150 cases if you're going to do ten cases per surgeon, and that sounded almost like what Dr. Kim was proposing anyway.

            If I can ask the sponsor to comment, do I have that correct on your initial roll-out?

            MR. VALEZ-DURAN:  Yes, you're correct.

            DR. MABREY:  Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            MR. VALEZ-DURAN:  I'm sorry.  I misspoke.

            DR. THOMAS:  This is Marc Thomas.

            If I did leave that impression, that wasn't correct.  We weren't going to limit it to 15 surgeons.  We want to get a geographical representation of America's surgeons regionally to train them in the U.K. so they can come back and be the faculty to train surgeons here as is the world wide template training for the Birmingham that's being done in countries such as the U.K. and Australia.

            There was no limitation of numbers as such with that application, but it was going to be restricted to a number of surgeons that we have thought of between 30 and 50, but once again have not restricted ourselves to a number.

            PANEL CHAIRPERSON NAIDU:  Thank you

            If we do not have any additional conditions, we should go back to voting on the main motion.  The motion on the table is to approve with conditions, the only condition being clinical and radiographic follow-up at ten years along sound statistical principles for the post approval study as outlined by the sponsor in the submitted PMA.

            All those in favor for the motion, please raise your hands.

            (Show of hands.)

            DR. KIM:  Can I clarify?  Are we voting on just the condition or are we voting on the approvability with this condition?

            PANEL CHAIRPERSON NAIDU:  We're voting on the main motion, approvability with this condition.

            Dr. Mayor?

            DR. MAYOR:  Are we submitting votes at this point?


            DR. MAYOR:  Affirmative.

            PANEL CHAIRPERSON NAIDU:  Dr. Blumenstein?

            DR. BLUMENSTEIN:  No.

            PANEL CHAIRPERSON NAIDU:  Dr. Mabrey?

            DR. MABREY:  Affirmative.

            PANEL CHAIRPERSON NAIDU:  Dr. Kim?

            DR. KIM:  I vote no.

            DR. SKINNER:  Affirmative.

            PANEL CHAIRPERSON NAIDU:  The motion passes.  The motion passes to approve with conditions, the condition being a post market study with radiographic and clinical follow-up at ten years based on sound statistical principles.

            Now, I'd like to go back to each panel member and ask  for the reason as to why they approved yes versus no.  Why don't we start off with Dr. Skinner?

            DR. SKINNER:  Well, I think that this device has shown that the data that has been presented in favor of this device has shown that it's reasonable.  It's safe.  It's efficacious, and I think that even though the post market study that has been suggested I think is unreasonable, I think it's the lesser of two evils.  So I think this is a better motion than trying to get a different motion.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Skinner.

            Dr. Kim.

            DR. KIM:  This technology is promising, and there is clearly a place for some type of technology in the younger, more active patients that will likely outlive a standard total hip replacement, but unfortunately, the information that was provided by the sponsor on this particular implant is insufficient to make several important conclusions.

            First, we do not know of any safety issues as it relates to early widespread use of this implant.  Again, this speaks to the issue of the learning curve.  If there are or will be significant issues with the learning curve, these should be identified and addressed prior to the release to the general public.  This is only possible in a well controlled study throughout multiple sites and surgeons.

            I do not think that evaluating Dr. McMinn's first 200 cases will be sufficient to address this important question as it does not take into account differences in clinical practice even among experienced or referral based surgeons.

            Number two, although I'm confident that Dr. McMinn can use this device safely and effectively, there's lack of sufficient evidence that this will be the case in the U.S. when a wide variety of surgeons will implant this device.  I see no compelling reason why this device does not need to satisfy some basic study criteria for its approval.

            I do not believe or encourage sponsors to present the FDA this type of study, and I would encourage the future Dr. McMinns of this world to take the extra effort to collect data that will be more meaningful and more compelling and more applicable to the U.S. population.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Kim.

            Dr. Mabrey, you vote yes, and your reasons?

            DR. MABREY:  Well, I think there is substantial amount of data out there to support the clinical efficacy of this device, and as far as its applicability to surgeons within the United States, like I say, I appreciate Dr. Thomas coming forward and sharing his experiences with it.  I just reviewed their abstract on line and they're very honest about pointing out the problems they saw at the beginning of their use of this device.

            I see quite an analogous situation with another procedure that's currently, if I may be so bold to say, running rampant throughout the U.S., and that's the use of two incision, mini incision total hip, which did not require any type of approval, and has a steep learning curve, and yet many surgeons have continued to adopt it and it looks like the initial adopters learned from their mistakes, conveyed those findings on to subsequent surgeons and thus steep learning curves were avoided.

            For those surgeons coming later, I would suggest -- well, I'm assuming that Dr. Thomas doesn't have to go all the way back to England to learn how to put this device in, but if his experience is anything like everyone else's, I feel comfortable that the learning curve will be there.

            I would encourage the sponsor to be honest about the learning curve, but I don't see it as being as steep as suggested in Dr. Mont's presentation at the academy this year.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mabrey.

            Dr. Blumenstein, you voted not.

            DR. BLUMENSTEIN:  First I'd like to echo Dr. Kim's comments and especially the lack of an estimate on the variability across surgeons, and then my second reason is that in my opinion there wasn't an adequate control on this trial, and I gave those reasons earlier.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Blumenstein.

            Dr. Mayor.

            DR. MAYOR:  I voted yes based on the reassurances that I gained regarding safety and efficacy for this device, without implying any satisfaction with the design of the study.

            Further, I would suggest that future applicants should not assume that significant savings can be achieved by following its example.

            PANEL CHAIRPERSON NAIDU:  Thank you, Dr. Mayor.

            Ms. Adams, do you have any comments to add?

            MS. ADAMS:  I do want to return to the comments that I made earlier about the congressional mandate for least burdensome, and I'm not going to sound like a broken record, but I want to remind everyone that Congress indicated that they wanted to insure the timely availability of safe and effective new products that would benefit the public and insure that our nation continues to lead the world in new product innovation and development.

            They indicated their goal was to streamline the regulatory process, reduce the burden and improve patient access to breakthrough technologies.

            In FDA's own guidance as a response to that, they've indicated that it is their goal to consider alternatives to randomized controlled trials; that there should be an effort to look at valid non-U.S. data, paper PMAs, literature controls, and that sort of thing.

            I respect every one of my colleagues at the panel here, but I would say that I'm not so certain that we're still, even though this was issued in 1997, considering the mandate of Congress and that that is something that we should be trying to consider as we deliberate in future panels.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Ms. Whittington?

            MS. WHITTINGTON:  I would echo that the public certainly is always interested in a better mousetrap and hopefully this is a better total hip.  It's the FDA's responsibility to make sure that it's a safe and effective device.  So I think the caution to safety is the post market study, and that that does insure that findings will be reported across the sites.

            The emphasis on education of the practitioners I think cannot be understated as Dr. Mabrey had indicated earlier, as well as good education to both the surgeon and the public as to the fact that this is a new device and it needs to be used in the right person by the right surgeon at the right time.

            PANEL CHAIRPERSON NAIDU:  Thank you, Ms. Whittington.

            Anymore comments from the panel?

            (No response.)

            PANEL CHAIRPERSON NAIDU:  Mr. Melkerson, have we addressed all of the issues adequately?

            MR. MELKERSON:  I believe you have, but I do want to again ask your opinion regarding this device.

            PANEL CHAIRPERSON NAIDU:  Yes, I am with a yes mainly because I think there is enough valid scientific data, albeit there are issues with the study.  It is a retrospective design based on a single surgeon's experience.  We have approved such PMAs before.  It is an innovative device.  I think we need the device.

            I think with the post market approval study that's stipulated here as a condition of approval, I think this device will be a good addition to the surgical armamentarium.

            Thank you.

            MR. MELKERSON:  Thank you.

            I would also like to thank you for standing in on short notice as Acting Chair.  We regretfully had to identify Dr. Kirpatrick's father passed away, and that's the reason why he's not there, and that's why he shows up as being the Acting Chair on our list of panel attendees.

            Thank you.

            PANEL CHAIRPERSON NAIDU:  Thank you, Mr. Melkerson.  It was a pleasure.

            The meeting is now adjourned.

            (Whereupon, at 4:22 p.m., the meeting in the above-entitled matter was concluded.)