* * *


* * *


* * *


NOVEMBER 20, 2003

                       * * *


            The Panel met in the Grand Ballroom of the Gaithersburg Hilton Washington D.C./North, 620 Perry Parkway, Gaithersburg, Maryland, at 9:00 a.m., Dr. Warren K. Laskey, Chairman, presiding.



WARREN K. LASKEY, M.D.  Chairman

SALIM AZIZ, M.D.        Voting Member

JOHN W. HIRSHFELD, M.D. Consultant

ALLEN A. HUGHES, Ph.D.  Consumer Representative


WILLIAM MAISEL, M.D., M.P.H. Voting Member

SHARON-LISE T. NORMAND, Ph.D.     Voting Member

DOUGLAS A. MORRISON, M.D.    Consultant





GERETTA WOOD            Executive Secretary
























                  C O N T E N T S


Conflict of Interest Statement ................. 5

Introductions .................................. 8

Open Public Session:

      Kevin McKim ............................. 11

P030025:  TAXUS Express2 Paclitaxel-eluting

Coronary Stent Systems:


Sponsor's Presentation:


      Dennis Ocwieja .......................... 15

      Mary E. Russell, M.D. ............... 20, 84

      Gregg W. Stone, M.D. .................... 45


FDA Presentation:


      Jennifer Goode ......................... 129

      Stephen Hilbert, M.D., Ph.D. ........... 140

      John Stuhlmuller, M.d. ................. 147

      Heng Li, Ph.D. ......................... 156


Committee Reviewers:


      Christopher J. White, M.D. ............. 172

      Judah Z. Weinberger, M.D., Ph.D. ....... 185


Questions to the Committee ................... 234


Open Public Session:


      Jim Gustafson .......................... 289



                                       (9:02 a.m.)

            CHAIRMAN LASKEY:  This being the magic hour, I'd like to call this meeting to order. 

            The topic today that the Circulatory System Device Panel will consider, the discussion of the pre-market application for the Boston Scientific Corporation, TAXUS Express2 Paclitaxel-eluting coronary stent systems, P030025.

            I'd like to ask the Executive Secretary to read the conflict of interest statement

            MS. WOOD:  The following announcement addresses conflict of interest issues associated with this meeting and is made part of the record to preclude even the appearance of an impropriety.

            To determine if any conflict existed, the agency reviewed the submitted agenda for this meeting and all financial interests reported by the committee participants.  The conflict of interest statutes prohibit special government employees from participating in matters that could affect their or their employer's financial interest.

            The agency has determined, however, that the participation of certain members and consultants, the need for whose services outweighs the potential conflict of interest involved, is in the best interest of the government.

            Therefore, waivers have been granted for Drs. William Maisel and Judah Weinberger for their interest in firms that could be affected by the panel's recommendations.  Dr. Maisel's waiver involves a grant to his institution for the sponsor's product study in which he had no involvement and for which funding to the institution was less than $100,001 per year.

            Dr. Weinberger's waiver involves stockholdings in competitors in which one value is between $5,001 and $25,000, and the other value is between $25,001 and $50,000.

            The waivers allow these individuals to participate fully in today's deliberations.  Copies of these waivers may be obtained from the agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building.

            We would like to note for the record that the agency took into consideration other matters regarding Drs. Weinberger, John Hirshfeld, Douglas Morrison and Clyde Yancy.  These panelists reported past or current interests involving firms at issue, but in matters that are not related to today's agenda.  The agency has determined, therefore that they may participate fully in all discussions.

            In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest, the participant should excuse him or herself from such involvement and the exclusion will be noted for the record.

            With respect to all other participants, we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement with any firm whose products they may wish to comment upon.

            CHAIRMAN LASKEY:  If I can have the panels introduce themselves, please, beginning on my right.

            DR. ZUCKERMAN:  Bram Zuckerman, Director, FDA Division of Cardiovascular Devices.

            DR. HIRSHFELD:  John Hirshfeld.  I'm at the University of Pennsylvania School of Medicine.

            DR. WHITE:  Chris White.  I'm from New Orleans, Louisiana.

            DR. SOMBERG:  Hi.  John Somberg, Rush Presbyterian, St. Luke's Medical Center, Chicago.

            DR. NORMAND:  I'm Sharon-Lise Normand, statistician, Harvard University.

            MS. WOOD:  Geretta Wood, Executive Secretary.

            CHAIRMAN LASKEY:  Warren Laskey, Uniformed Services University and National Naval Medical Center.

            DR. MORRISON:  Doug Morrison, Tucson VA, University of Arizona.

            DR. YANCY:  Clyde Yancy, UT- Southwestern, Dallas.

            DR. WEINBERGER:  Judah Weinberger, Columbia University, New York.

            DR. MAISEL:  William Maisel, Brigham & Women's Hospital in Boston.

            DR. HUGHES:  Allen Hughes, George Mason University, consumer representative.

            MR. MORTON:  Michael Morton.  I'm the industry representative.  I'm employed by CarboMedics.

            CHAIRMAN LASKEY:  And, Ms. Wood, if you could read the voting status statement, please.

            MS. WOOD:  Pursuant to the authority granted under the Medical Devices Advisory Committee charter, dated October the 27th, 1990, and as amended August the 18th, 1999, I appointed the following individuals as voting members of the Circulatory System Devices Panel for this meeting on November 20, 2003:

            John W. Hirshfeld, M.D.

            Douglas A. Morrison, M.C.

            John C. Somberg, M.D.

            Judah Z. Weinberger, M.D., Ph.D.

            Clyde Yancy, M.D.

            For the record, these individuals are special government employees and are consultants to this 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.

            This is signed by David W. Feigal, Jr., M.D., M.P.H, Director, Center for Devices and Radiological Health, and it's dated November 17th, 2003.

            CHAIRMAN LASKEY:  Prior to the open public hearing portion this morning, I'd like to read the following statement.

            Both the Food and Drug Administration 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 that it is important to understand the context of an individual's presentation.

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

            For example, this financial information may include the sponsor's payment of your travel, lodging or other expenses in connection with your attendance at this 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 this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

            I'd like to now open the open public hearing portion this morning and invite Kevin McKim to speak before us.

            MR. McKIM:  Good morning.  Thank you.

            First of all, in regards to your first question, I have no financial reimbursement from either Boston Scientific nor their competition.

            I am here this morning to introduce the Metricath system, which is a new and novel product on the market, which offers quick, accurate, and cost effective measurement in arterial stent sizing.

            There's been a great deal of research that has demonstrated that guidance stent implantation offers better clinical outcomes for patients.  There's less risk or possibility of complications, while there is less possibility or reduced rates of restenosis and at the same time there is actually an increase in the average diameter or cross-sectional area of the stents that are implanted

            As well, there's a number of studies, such as the stress study, that demonstrated that angiography alone is oftentimes not sufficient to be able to assess stent apposition and to know whether or not stent are accurately placed within the arteries.

            And this, we believe, is particularly true with the drug eluting stents, that accurate and proper sizing of the arteries will allow the physicians to select appropriately sized stents.

            Additionally, it will allow them to insure that the stents are, in fact, fully expanded and apposed against the artery wall, and this has been mentioned in reports from both the manufacturers of the stents, as well as the FDA.

            So very quickly, what is the Metricath system?  The Metricath is a catheter based technology, and very simply the way it works is a low pressure balloon that conforms to the shape and the size of the artery that is being measured or to the inside diameter of the stent after it has been deployed, and from the information that's gathered from the balloon, it's very easy for us to quickly and accurately determine both the cross-sectional area and also the average diameter of the stent.

            So this allows us a method of (a) insuring that appropriate size stents are being chosen, and also afterwards to insure that the stents are, in fact, fully expanded and apposed against the artery wall.

            And a few of the benefits of the system are, one, that it is very accurate.  We have evidence to show that we are equivalent to IVUS, much more accurate than angiography alone, and more importantly, it doesn't require image interpretation, and it is given simple, objective measurement results.

            Second, it is very simple.  It offers very significant both technical and operational advantages over other systems that are currently available.

            Thirdly, it's very cost effective.  So we're offering some significant cost savings compared to competition.

            And finally, it is a very safe product that operates at only one-third of an atmosphere and at the pressures that it's operating at, we do have some research to demonstrate that there is no damage or change to the artery wall whatsoever.

            So in conclusion, we believe that sizing, in fact, is critical for drug eluting stents; that you offer better clinical outcomes.  At the same time, there is a much lower risk of complications, lower risk of perforations, lower risk of restenosis, and this is especially true with the drug eluting stents, where proper sizing and full apposition is absolutely critical for them to work effectively.  And these are all benefits that can be achieved with the use of the MetriCath system.

            And for more information, we do have a Web site, www.angiometrix.com.

            Thank you very much.

            CHAIRMAN LASKEY:  Thank you, sir. 

            Is there anyone else in the audience who wishes to address the panel on today's topic or any other topic?

            (No response.)

            CHAIRMAN LASKEY:  If not, I'd like to close this open public hearing portion and move on to the highlight of the day, the sponsor's presentation.

            MS. WOOD:  I would just remind the speakers to introduce yourselves and state your relationship to the company or the competitors.

            MR. OCWIEJA:  Mr. Chairman, members of the panel, consultants, at first I'd like to thank you for the opportunity to come before you today and present data and information relative to the TAXUS Express2 Paclitaxel-eluting coronary stent system.

            I am Dennis Ocwieja.  I am Senior Vice President, Regulatory Affairs and Quality with Boston Scientific Corporation.  There will be two other presenters today:  Mary Russell, M.D., Medical Director, Vice President, Cardiovascular Clinical Sciences, Boston Scientific Corporation; and Gregg Stone, M.D., Vice Chairman, Director of Cardiovascular Research and Education, Cardiac Research Foundation, Lenox Hill Heart and Vascular Institute, and a principal investigator for our TAXUS-4 clinical study.

            We also have with us a number of expert consultants.  Let me read their names to you and what area of expertise we expect them to be able to help us in.

            Dr. Black, Hugh Black and Associates, Toxicology.

            Dr. Chiaccierini, RPC Consulting, statistical.

            Dr. Stephen Ellis, Cleveland Clinic, co-PI on the TAXUS-4 study.

            Dr. William Kramer, Kramer Consulting, pharmacologist.

            Dr. Jeffrey Popma, Harvard Clinical Research Institute, Angiographic Core Lab.

            Dr. Eric Rowinsky, Institute for Drug Development and Cancer Research, oncologist.

            Dr. Robert Schwartz, Minnesota Cardiovascular Research Institute, preclinical.

            Dr. Neil Weissman, MedStar Research Institute, our IVUS Core Lab.

            And Dr. Greg Wilson, Hospital for Sick Children, Toronto, pathologist.

            Our agenda for today is as follows.  After this brief introduction that I will be presenting to you, I'll also discuss our proposed indication for the TAXUS stent.

            I'll then turn the floor over to Dr. Russell, who will talk to you about our TAXUS technology, the TAXUS preclinical program, and clinical development program.

            Dr. Stone will then come and give you information relative to the TAXUS-4 pivotal U.S. trial.

            Dr. Russell will come back and present conclusions, and then the three of us will be available to take questions from the panel and direct them to our expert consultants as necessary.

            Why are we here today?  The rationale for drug eluting stents.

            Since restenosis is a major limitation, we're saying 800,000, and we have indications that maybe as much as a million coronary stenting procedures are performed annually, and 20 to 30 percent of the patients receiving a bare metal stent develop angiographic restenosis with half or approximately 100,000 of the patients requiring repeat intervention, either PCI or CABG. 

            The TAXUS stent technology was developed to reduce restenosis and the need for repeat reintervention.

            We're looking at a proposed indication for the TAXUS stent in our product labeling that reads as follows:  "TAXUS Express2 Paclitaxel-Eluting Coronary Stent Systems," both "(Monorail and Over-the-Wire) are indicated for improving luminal diameter and reducing restenosis for the treatment of de novo lesions less than or equal to 28 millimeter in length in native coronary arteries, greater than or equal to 2.5 to less than or equal to 3.75 millimeter in diameter."

            In addition, while we're talking about labeling, there was an oversight that we made perhaps, and certainly a question that was addressed to the panel member relative to antiplatelet labeling recommendations, and we're proposing recommendations for the DFU that Clopidogrel be recommended at a 300 milligram loading dose prior to implant, a 75 milligram dose per day for six months, and that aspirin also be provided to the patient prior to implant and used indefinitely thereafter.

            We are looking for a proposed stent matrix for approval of the product that includes stents from 2.5 to 3.5 millimeter in diameter and eight millimeters to 32 millimeters in length.  The shortest stents, the eight and 12 millimeter length stents are intended to be used as in the clinical study for multiple stent procedures to optimize angiographic outcomes.

            At this point I'd like to turn it over to Dr. Russell to get into the technology overview.

            DR. RUSSELL:  Thank you very much.

            Let's begin with the TAXUS technology overview.  There's three components:  the drug, the polymer, and the stent.  The drug we've selected is Paclitaxel.  It's located on the stent at one micrograms per millimeter squared per stent surface area.

            Paclitaxel is the active agent used for chemotherapeutic purposes.  Paclitaxel is the active agent in Taxol used for chemotherapeutic purposes.  It binds to beta tubulin shown here in yellow, disrupts the dimerization of tubulin, and hence interrupts microtubulin dependent activities, including, but not limited to, signal transduction, cell activation and migration, inflammatory cell chemotaxis, cell proliferation, growth factor release, and matrix organization.

            There are contrasting use of Paclitaxel.  There's the anti-neoplastic use contrasted against the anti-restenotic use.  For anti-neoplastic purposes, Taxol incorporates Paclitaxel in a Cremophor to solubilize it for intravenous administration for systemic therapy.  This systemic therapy is targeted at malignant, actively dividing cells.  A single dose would involve about 300 milligrams for an average human being and a cumulative dose cycle involving four cycles would be about 1,200 milligrams.

            In contrast, for anti-restenotic purposes, Paclitaxel is coated onto the TAXUS stent.  This is done through a polymer, provides local vascular dosing, targeted at  activated vascular cells in the post implant period.

            A single 32 millimeter stent has 209 microgram loaded on it.  Four stents, should they be required, would have a total of 0.8 milligrams.  If all of the dose were released from the four stents, which is improbable as I'll show you on subsequent slides, there would be a 1,400-fold lower Paclitaxel exposure with the anti-restenotic use than anti-neoplastic systemic exposure.

            The second critical component of our technology is the polymer.  It's termed Translute.  Polymer based delivery was specifically selected because it provides local, targeted, short-term drug delivery, controls release of the drug from the polymer, provides homogeneous drug coverage along the length of the stent, retains the drug on the stent during handling, and hence prevents drug loss during the implant procedure, nonuniform drug distribution, and then uncontrolled release.

            Translute is a proprietary polymer.  It's a hydrocarbon based elastomer specifically selected because of its processing capability and the performance of uniform coding over the stent struts.  It's highly biostable and standardized testing and as I'll show you, it's vascular compatible.

            Procine stent implant studies evaluating mortality, morphologic and quantitative morphometric parameters were performed to establish that the Translute only coated stents are comparable to bare control stents at Day 28, 90 and 180 days.

            Additional studies evaluating polymer stability testing were performed on canine graphs out to two years.  Polymer coating integrity was maintained long term without evidence of flaking, cracking or delamination.  There was no particulate generation, and there was no polymer degradation in vitro or in vivo.

            Molecular weight profiles are stable by sizing analysis using a polydispersity index, and these findings are true for TAXUS stents implanted in the porcine model and then explanted for evaluation at up to 180 days.  This allows us to conclude that the Translute polymer is vascular and biostable, and durable in vitro as well as in vivo.

            I'd like to now cover the TAXUS release kinetics, first touching on the characteristics of a polymer matrix delivery system, and then reviewing in vitro, in vivo, and systemic release characteristics.

            The Translute Paclitaxel polymer system follows the model of a matrix drug delivery system.  There's low drug loading in the polymer.  It provides for uniform or homogeneous drug distribution throughout the polymer matrix by phasic release, as I'll show you, and the release of very low concentrations of Paclitaxel over a short period of time, depending on the formulation.

            The characteristics of our low Paclitaxel dose loading on the TAXUS step are depicted in this cartoon.  There's a biphasic release mediated by an early burst phase in the zero to two-day period.  This is due to dissolution mediated release of Paclitaxel from the surface of the coated stent.

            It's followed by a sustained slow release phase, two to ten days, where the drug diffuses through channels in the subsurface shown here in yellow to the surface, and then finally there's a plateau phase after 30 days where drug is sequestered within the polymer matrix without pathways to the external environment in the so-called bulk phase shown here in pale green.

            This depicts in vitro release from the TAXUS stents, Paclitaxel release shown here.  You see the early burst phase, zero to two days, followed by the sustained slow release phase, two to ten days.

            We have two dose formulations under active investigation, the slow release formulation the subject of this PMA and the moderate release formulation that produces three times higher initial release.

            They both have a dose density of one microgram per millimeter squared, but they're distinguished because there's a lower Paclitaxel polymer ration in the slow release formulation compared to the higher moderate release formulation.  This results in sparser drug distribution depicted here in the cartoon for the slower compared to the denser for the moderate release.

            Hence, to achieve the one microgram per millimeter squared, the coating thickness is larger in the slow release than it is in the moderate release.

            This results in differential in vivo release kinetics, with 7.5 percent of the total loaded drug released in the first 30 days for the slow release formula contrasted with 21.9 percent release for the moderate release formulation.  This leaves a residual 92.5 percent of the total loaded drug retained on the stent after 30 days for the slow release formulation compared to 78.1 percent for the moderate release formulation.

            This depicts cumulative in vivo release in the rabbit iliac model as shown here.  The slow release in yellow has less release than the moderate release in green.  There's a plateau after 30 days, and the moderate release formulation has been our focus for preclinical studies because of its higher release.

            The slow release has been the focus of our clinical studies focused on efficacy.

            We've examined the stability of the retained Paclitaxel in this bulk phase, performing studies aimed at identifying the maximum release conditions.  Physiologic conditions shown here, including phosphate buffered ceiling with various additives, such as surfactants, cyclodextrins, proteins including albumen, lipids, and whole blood, do not alter the release level.

            However, non-physiologic conditions, including 100 percent solvents, low pH and high temperature, are required to reach the FDA benchmark of greater than 80 percent release.

            Under these extreme conditions, the polymer's physical characteristics are destroyed as measured by increased tack force and increased swelling, and these non-physiologic conditions are impossible to reach in living human beings.

            We've also examined drug release after intentional coating disruptions.  For example, we've used scalpels to abrade 50 percent of the abluminal surface of the slow release stent, and you can see an abraded strike here with the elastomeric polymer adherent.

            And then we've gone on to measure in vitro the amount released contrasting the abraded with the controlled stents.  You can see there's a small increase in release to 1.8 percent for the abraded compared to 1.2 percent for the controlled stents, indicative of stable, low level release even with artificial abrasion.

            Systemic Paclitaxel measures were evaluated in studies simulating higher Paclitaxel release, specifically serum levels were measured in porcine stent implant models after exposure to maximum doses likely to be present in humans.

            We used the moderate release formulation, which releases more than the slow release formulation; studied overlapping stents and higher dose densities up to four micrograms.  We developed a high sensitivity liquid chromatography tandem mass spec assay that has a lower limit of quantitation of 0.03 nanograms per mL and a lower limit of detection of 0.006 nanograms per mL.

            In these studies, one of 247 samples, a pig that received two moderate release overlapping stents had a level of 0.048 nanograms per mL, just over the detection limit, immediately post implant in one of the specimens that was measured.

            The other 246 out of 247 samples have no Paclitaxel detectable.  Hence, to conclude with respect to the TAXUS technology, we have a vascular compatible elastomeric polymer.  It produces biphasic, low level release in the targeted 30-day period.  After mechanical challenge, release remains low and stable.  Retained drug is sequestered in the bulk and has no measurable release under physiologic conditions.  And we have no detectable systemic Paclitaxel release with a high sensitivity assay.

            With respect to our preclinical safety program, it is an extensive program that has involved more than 12 studies, more than 250 swine, 800 stents, and gone out to 360 days.  It is a program that has evolved from polymer only studies to a focus on both moderate release and slow release studies.

            The majority of our studies have focused on the moderate release formulation because of the higher amount released in the early ten-day period.

            We perform both overlapping stent and dose density studies to establish the safety of the moderate release formulation and, as important, to provide a safety margin for the slow release formulation.

            For the slow release formulation, we've looked at overlapping stents, and then we've looked at the histologic profiles over time with an alternative antiplatelet regimen, ticlopidine.

            For preclinical safety parameters, there's been consistency across the moderate release, as well as the slow release studies.  Out of 127 pigs, there were two that were categorized as cardiac morality.  However, one was a super renal abdominal laceration, and the other was an air embolism, during implant.

            So there were effectively no TAXUS related cardiac mortality.  There were no miocardial infarctions as graded histologically, and there were not stent thromboses in all 340 stents evaluated.

            Hence, the pre-clinical safety pattern correlates to the clinical safety pattern.  We focused on early histologic safety parameters, have been able to demonstrate surface passivation by  day 28 for moderate release stents using endothelial cell coverage.  It is complete by day 28 for all single layer, moderate release stents, for all slow release stents. 

            It's complete by 90 days for the moderate release overlapping stent layers.  There's resolution of microthrombi and surface fibrin.   Assessment of microthrombi demonstrated that it was statistically equivalent in the control and the TAXUS groups at all time points.  There were rare microthrombi as depicted here and shown here at high power with no evidence of gross thrombi.

            And importantly, there was neointimal coverage on all struts with an early, neointimal organization pattern.

            Our early histology correlated with our preclinical safety patterns.

            With respect to late histologic safety, we're demonstrated that it is maintained out to 360 days for the moderate release formulation, demonstrating a mature neointima shown here as a compacted, thin, fibrocellular layer, indicative of a quiescent response.  Vessel stability has been demonstrated using morphometric measures of EEL, which are equivalent to control and demonstrating a stable stent to IEL relationship.

            And then finally, there are no aneurisms.  There is medial cell loss , as I'll show you, but it's replaced by structural fibrous tissue.  There was no incomplete stent apposition.  Vessel wall integrity was maintained, and there were no aneurismal changes.

            Hence, the late and sensitive histologic analysis supports long-term safety.

            We've explored safety in overlapping stent models, looking at the moderate release formulation out to 180 days and 132 stents.  The advantage of this model is that one could compare the proximal single and distal single strut layers with the overlapping two strut layer regions.  Endothelial cell coverage was present by day 28 for the single layers, by day 90 for the overlap layers.  There was stable neointemal and vessel structure out to 180 days, and there was healing in the overlap zone where there was a double strut and theoretically a double Paclitaxel dose.

            Hence safety has been shown for overlapping stents using the moderate release formulation.

            We've looked at histology patterns over time, again, with the moderate release formulation from 28 out to 360 days.  Endothelialization is complete as to the presence and maturation of the neointima and then the maintenance of vessel structure as measured morphometrically, indicating healing over time with the moderate release dose.

            We characterized the observation of para-strut amorphous material, which is present in control and Paclitaxel eluting stents.  It's shown here as an acellular fibrin rich area immediately adjacent to the stents with a connective tissue component.  There's more in the moderate release than the slow release, than the control.

            It involves less than five percent of the vessel wall in the slow release.  It's not inflammatory in that there's a paucity of white cells.  It's sequestered within a maturing neointima in all cases, and importantly, there is no associated luminal thrombus, erosion or vessel wall necrosis.

            If we look at the para-strut amorphous material over time, we see that it resorbs and heals with a moderate release stent.  At day 28, it localizes to the struts.  It's most apparent at this time, but highly variable in size.  There's less in the slow release than the moderate release.  There's associated early medial cell loss, and it's sequestered within the neointima.

            Between day 90 and 180, there's cellular in-growth adjacent to the struts due to smooth muscle cell and white cell accumulation.  The PAM zones themselves decrease in size and intensity.  They're not inflammatory during this resolution phase, and there's a maturation of the neointimal covering with stable medial and adventitial regions.

            By 360 days, there's resorption of the PAM, fibrous in-growth, clearly healing at the implant site and stable vessel and neointima.

            We've concluded that the early  Paclitaxel burst release does not inhibit the later healing pattern.

            With respect to strut associated medial changes for the moderate release, between the struts we have a normal media and a moral neointima.  Adjacent to the struts for both control and the TAXUS program there is compression; there is medial cell loss, more apparent in the TAXUS than in the control.

            However, this is followed by fibrous replacement and stable vessel morphometry, and we've demonstrated that vessel stability is maintained.

            With respect to calcium deposits, shown here are the worst case scenario.  Calcium deposits are present in both control and moderate release.  they're scarce in both groups.  An indicator of healed prior focal necrosis in response to implant injury, they tend to be larger in the control and strut based.  They're more frequent in the TAXUS group, but have a more diffuse pattern, as shown here by the black staining.

            They're variable, but small in size, typically less than 50 microns, and they are sequestered, inert, and stable over time.

            So to summarize our preclinical histologic findings, we have healthy endothelial cells and strut coverage by day 28 for the slow release stent.  There is early medial changes followed by fibrous replacement.

            Neointima forms around, over and between the struts.  There are small areas of an inert para-strut amorphous material that resorb progressively through 360 days, and we're left with an artery that's pacified, has generated a stable neointima, and is structurally intact.

            With respect to the potential concern raised by FDA, histologic observations of PAM, medial loss, calcification did not  develop into higher risk of stent thrombosis, aneurisms, incomplete apposition, or dissection, in long-term assessments, both preclinically and clinically, with and without overlapping stents.

            Safety has been established for the slow released TAXUS Express in that there were no safety events with progressive healing patterns demonstrated histologically.  Safety demonstrated for overlapping moderately stents and an acceptable safety margin established for the slow release based on the safety of the moderate release.

            I'd now like to focus briefly on the TAXUS clinical development program.  This has involved six randomized trials with an anticipated 3,400 patients by the end of 2003.   We've evaluated both the slow and moderate release formulations, and we've progressed from low to higher risk lesions, as I'll show you.

            The program has as its foundation TAXUS I, a feasibility study on slow release; TAXUS II, an international study looking at both slow release and moderate release; and then a single arm in stent stenosis registry using slow release.

            The focus of our program and the focus of today's presentation is the U.S. IDE pivotal TAXUS IV study using the slow release formulation to be presented shortly by Dr. Stone.  The program has expanded to include the randomized studies, TAXUS V, a de novo stent study involving smaller reference vessel diameters and longer lesions; TAXUS V in-stent restenosis randomized against brachytherapy; and TAXUS VI, a European long lesion study on the moderate release formulation.

            In addition to that, we have two active registries, WISDOM and MILESTONE II.  In terms of the clinical development program, the foundation, as I just mentioned was TAXUS I through III that used the mere stent platform.  Since then we have focused on the Express stent platform in TAXUS IV through V-ISR.

            We've moved from work horse stent diameters of three and three-five to TAXUS IV, where we used a two-five stent in addition, and then more recently TAXUS V with a stent diameter of 2.25 as well as 4.0. 

            Lesion lengths have increased from the focal ten to 12 in the foundation studies to up to 28 in TAXUS IV and up to 46 in TAXUS V  and V-ISR.

            We've moved from single lesion stenting studies to multiple lesion overlapping stent studies in the forms of TAXUS V and VI, and we've moved from allowing single vessel stenting procedures to multiple vessel stenting procedures to mimic the real world practice.

            We've studied the slow release in six studies and the moderate release in two studies.

            TAXUS I was a feasibility study focused on the TAXUS slow release formulation.  We evaluated standard risk de novo lesions.  It was randomized one to one.  IVUS was used post placement and in follow-up in all patients.  This was targeted at improving our sensitivity and picking up vascular changes both for safety and efficacy assessments.  It was double blind and the core labs remain blind to date.  It was initiated in September of 2000, enrolled 61 patients, 31 of them receiving the slow release stent.  Clopidogrel was protocol mandated for six months, as it is in all of our studies, and we are evaluating those patients out to five years.

            The baseline demographic and lesion characteristics were well matched with a reference vessel diameter of 2.9 and a lesion length of 11.3 millimeters.  At six months there were concordant and significant QCA and IVUS improvements of greater than 50 percent.  Binary restenosis was zero percent without evidence of any edge restenosis, and at 12 months the MACE was three percent in the TAXUS group versus ten percent in the bare metal control, and we are now in the middle of our third year of follow-up.

            This slide depicts the TAXUS I target revascularization rate with control shown here in gray, with ten percent TLR at 24 months.  TAXUS has not had a single target lesion revascularization out to 24 months.

            TAXUS II is our international slow and moderate release study.  It's focused on standard risk de novo lesions.  Again, randomized one to one; again, double blind; again, the core labs remained blind.  It used IVUS post placement and in follow-up in all patients and remains the largest IVUS database in the drug eluting stent arena.

            It has as its primary endpoint percent net volume obstruction by IVUS, was initiated in June 2001, enrolled 536 patients at 38 international centers into two cohorts, the TAXUS slow release cohort with its own control, with DMC clearance to move into the taxus moderate release cohort with its own control.

            Clopidogrel, again, mandated for six months and follow-up for five years.  Well matched demographic and lesion characteristics with a reference vessel diameter of 2.8, a lesion length of 10.4, and for both the slow release and the moderate release the primary endpoint was met.

            There were also significant improvements in six month QCA and IVUS parameters, as well as edge parameters and ongoing benefit demonstrated at 12 months for TLR, TVR and MACE.

            This study is now ending its second year of follow-up.

            This slide depicts the TAXUS II target lesion revascularization responses.  As you can see, at six months there was significant improvements between control and the slow release, as well as the moderate release group.  At 12 months these benefits were sustained with a TLR rate of 14.4 percent in the control, 4.7 percent in slow release, and 3.8 percent in moderate release, indicating stable and preserved target lesion revascularization out to 12 months.

            Systemic Paclitaxel levels have been measured in TAXUS I through III.  Blood levels were quantitated by a commercial lab with a lower limit of quantitation of ten nanograms per mL.  All 2,188 samples had no measurable Paclitaxel.

            With respect to the issue of potential systemic drug interactions, particularly with the pharmacotherapy common in coronary artery disease, patients that you're familiar with, the antiplatelet regimens, anti-coagulations, statins, and other cardiovascular meds, there has been no precedent for interruption of these meds with Paclitaxel, despite systemic Paclitaxel's infusions for oncology patients with levels up to 3,650 nanograms per mL and a vast clinical experience since approval of Paclitaxel in 1992.

            This is to be contrasted with the plasma levels with the TAXUS stent that are below the lower limit of quantitation of the commercial assay, ten nanograms per mL.  Hence, it is unlikely that metabolism or clearance of pharmacotherapy common in coronary artery disease patients will be affected by the TAXUS stents or vice versa, given the contrast in these plasma levels.

            With respect to the issue of potential hypersensitivity reactions, the principal culprit for Taxol associated hypersensitivity is the Cremophor, a poly-ethoxylated castor oil that's used to solubilize Paclitaxel and other agents.  It generates a histaminic response.

            In the oncology arena, this can be minimized by protracting the infusion over a 24-hour period and using lower systemic Taxol dosing.

            Probably  most important for our technology, we do not use the Cremophor on the TAXUS stent system.  We solubilize the Paclitaxel with our polymer, and then, of course, as I've just shown you, plasma levels with the TAXUS set are below level of detection.

            Hence, we believe hypersensitivity reactions are unlikely with the TAXUS stent.

            To summarize, TAXUS I and TAXUS II serve as a foundation.  They studied the standard risk, focal, de novo lesions; demonstrate improved efficacy with concordant benefits clinically and by QCA and IVUS, and durable patterns out to two years.

            The safety profile we believe is acceptable.  There's improved MACE.  No measurable systemic Paclitaxel levels, no vessel enlargement, aneurisms or dissections in spite of intensive IVUS analysis, demonstrating structural stability post implant.

            These TAXUS I and II findings support the TAXUS IV findings that you're about to hear about from Dr. Stone.

            And it is now my pleasure to introduce Dr. Stone, the principle investigator of the TAXUS IV study.

            Thank you very much.

            DR. STONE:  Thank you.

            Good morning.  It's my pleasure to present to you the pivotal data from the TAXUS IV trail in the United States.

            In the way of disclosure, I'm a consultant for Boston Scientific.  I've also served as consultant for several of Boston Scientific's competitors, including Guident and Medtronic.  I hold an equity position in a subsidiary company, which was purchased by Boston Scientific, which I liquidated before the study was complete, and finally the not-for-profit organization that I work for, the Cardiovascular Research Foundation, which also puts on the annual TCT meeting, has received either research or educational grants from most of the companies that produce both devices and pharmacology products related to interventional cardiology.

            The TAXUS IV pivotal trial of the TAXUS slow release stent was a study that enrolled 1,326 patients undergoing elective stenting at 73 United States sites.  For enrollment in this trial, the patient had to have a single lesion, which was de novo, that is, never before treated, coverable by a single stent.

            The reference vessel diameter had to be between 2.5 and 3.75 millimeters by visual assessment, and to cover these vessel diameters, we had stent diameters present of 2.5, 3.0, and 3.5 millimeters.

            The lesions length had to be visually estimated by the operator to be between ten and 28 millimeters, and we had stent length available of 16, 24, and 32 millimeters, and we had a full matrix of diameters and stents, except there were no 32 millimeter long, 2.5 millimeter stents.

            In order to insure that several of the very important variables that can affect restenosis were equally distributed in the two groups, the randomization was stratified by the presence of medically treated diabetes and by visually assessed vessel diameter, either less than or greater than three millimeters.

            Patients were pre-treated with aspirin, 325 milligrams, and given a Clopidogrel of 300 milligram loading dose, and then were randomized one to one before predilitation to either the one microgram per millimeter squared slow rate release Paclitaxel-eluting Express TAXUS stent or an identical appearing uncoated Express stent.

            I think it's an important point to note that the stents were absolutely indistinguishable, and the stents and patients were randomized by serial number.  So to date we still don't know which stent was put in which patient, and the study is being completely blinded through the five-year follow-up period.

            Patients were treated after stent implantation with Clopidogrel for six months.  Clinical follow-up was planned at one, four, and nine months, and yearly for five years with the primary endpoint of the study being clinical at nine months.

            Angiographic follow-up was planned in a subset in nine months of 732 patients, and intravascular ultrasound follow-up was planned in a subset of 268 patients.

            Before we get more into the trial, I'd like to go over a few of the definitions that are used to understand the safety  and efficacy of this device.

            The principal endpoint of the study was target vessel revascularization, and the two most important determinants of efficacy of drug eluting stent are both target vessel and target lesion revascularization called TVR and TLR.

            TVR and TLR refer to ischemia driven, repeat, percutaneous coronary intervention or bypass surgery of the target vessel or the target lesion, respectively.  And so to be able to meet the definition of a repeat procedure that counts as a TLR or TVR, a separate independent clinical events committee, which was blinded, of course, to randomization assignment, had to say that there was significant restenosis either at the lesion or at the edges of the lesion. 

            So it included the entire analysis segment that we'll talk about later with angiography, and they had to state that there was ischemia that was driving that repeat intervention or that there was a severe diameter stenosis present by angiography.

            Target vessel failure was defined as either cardiac death, target vessel revascularization or myocardial infarction, which was adjudicated by the clinical events committee as being related to the target vessel.  If it wasn't clear whether or not and event was related to the target vessel or not, it was adjudicated as being target vessel failure.

            Major adverse cardiac events were more encompassing relating cardiac death, myocardial infarction, or target vessel revascularization whether or not related to the target vessel.

            And finally, stent thrombosis, of course, is an important indicator of safety, and we used a very sensitive determinant of stent thrombosis being either angiographic, that is, vessel renarrowing occlusion with the presence of thrombus or without the presence of  angiography, any death within the first 30 days without an obvious clause, such as pneumonia or sepsis, was adjudicated as being due to stent thrombosis even without angiographic confirmation, and any acute myocardial infarction without angiographic confirmation was also adjudicated as stent thrombosis.

            Of the 1,326 randomized patients, the analysis plan prespecified that the evaluable population would be those patients in whom a study stent was attempted to be placed.  Whether or not it was actually placed or not, those patients are still in the evaluable population.

            There were a total of 12 patients, a very small percentage in which for a variety of reasons no study stents were attempted to be placed.  Either the patient withdrew on the table or there were no study stents available or there was a left main or a wire dissection when those stents were implanted, and again, by analysis plan, we for statistical purposes de-register those patients.

            However, these patients were followed for the full nine-month period in ongoing, and while we'll be presenting the data today from the 1,314 patient evaluable study population for the analysis plan, we also have generated data which is in your panel packs for the complete 1,326 patients, and there are no significant changes in any of the conclusions.

            Of the 1,314 patients, they were equally assigned, again, 662 to the TAXUS group and 652 to the control group, and nine-month follow-up was complete in 97 percent of the patients.

            Now, of these 1,326 patients, we had originally planned a cohort of 1,172 patients, but we had prespecified to FDA that we would enroll at least 200 patients, receiving 2,400 millimeter stents and have follow-up on them, and 200 patients with 32 millimeter stents. 

            The 32 millimeter stents became available late within the clinical trial.  The clinical trial enrolled actually faster than we had anticipated, and thus, as was prespecified in the protocol, we expanded the enrollment in an additional 154 patients to include enrollment of additional 32 millimeter patients into the study, which brought us up to 1,326.

            Looking at the 1,314 evaluable patients, you can see that they have the pretty much typical characteristics of populations that undergo percutaneous coronary intervention.  Approximately 28 percent of the patients were women.

            Important to note about a quarter of the patients had diabetes, and eight percent were insulin requiring diabetics, mostly Type II diabetics.

            And, again, the majority had hypertension, hyperlipidemia, with significant numbers of current smokers, prior myocardial infarction, and unstable angina, and there was equal representation of all the major baseline demographic characteristics between the control groups and the TAXUS groups.

            The lesions that were treated were in the left anterior descending artery in approximately 40 percent of the cases, the right coronary artery in about 32 percent of the cases, and the circumflex artery in about 28 percent of the cases.  So the vessel distribution of the lesions were similar between the two groups.

            By quantitative coronary angiography, the reference vessels were 2.75 millimeters in diameter in both groups, so moderately small vessels.  The minimal luminal diameters in diameter stenoses were what we've come to to expect by QCA in trials of this type, and the lesion lengths were moderately long at 13.4 millimeters, with a standard deviation of 6.3 millimeters, meaning we had short lesions, moderate length lesions, and long lesions, and there were no differences in the baseline QCA lesion characteristics between the control groups and the TAXUS group.

            This somewhat busy slide shows the procedural details within the trial.  Glycoprotein IIb/IIIa inhibitors were used in about 57 percent of the patients, IVUS within a quarter.

            We allowed percutaneous coronary intervention of non-target vessels prior to randomization and for randomization to continue, this had to be a successful, uncomplicated PCI of the non-target vessel.  So this made this a higher risk population than in other drug eluting stent studies, but it made much more of a real world study.

            And about 20 percent of the patients had a second lesion in a non-target vessel which was treated with a stent or FDA approved stents.

            Stents were deployed at 15 atmospheres, and as you can see the balloon-to-artery ratios were about 1.2 to one, and the stent-to-lesion length ratio is about 1.6.  There were no differences in any of the procedural characteristics between the control groups and the TAXUS groups.

            The mean number of stents implanted per patient were about 1.1 in both groups.  So this was a single lesion, single stent study.  However, as always, there are some times when an operator would choose a stent which is a little short for the lesion, and there are other times where there would be an edge dissection where a second stent had to be implanted.

            As you can see, most of the time, about 98 percent of the time, as prespecified by the protocol, there was an appropriate size study stent that was available, and of course, if the patient randomized to the TAXUS stent, the IVRS system would give them a new serial number for another TAXUS stent, and similarly if it was bare metal stent.

            Non-study stents were used in about two percent of patients, and this typically occurred when the right size taxus stent is not available.  The data, of course, that we'll show you will be intention to treat, which would include all of these stents, both study stents and non-study stents.

            So thus, the vast majority of patients received one study stent.  You can see a smaller percentage of patients with two study stents, and a very small number that had three study stents.

            This is the matrix of the stent sizes that were actually implanted.  You see the control group and the TAXUS group, 16 millimeter stents, 24 millimeter stents, and 32 millimeter stents.  These are absolute numbers broken down by in the brown, 2.5 millimeter diameter; in the orange, three millimeter diameter; and in the yellow, 3.5 millimeter diameter.

            Approximately 60 percent of the stents implanted were the 16 millimeter stents, and about 20 percent were the 24 millimeter stents, and 20 percent were the 32 millimeter stents.  And there was good representation of both small stents, intermediate stents and large stents, and as you can see again, there were no 32 millimeter/2.5 millimeter stents in this trial.

            Aspirin was, of course, prespecified in all patients throughout the follow-up period, and thienopyridines, Plavix being the first choice for six months, both pre-load and then for six months.

            If you look at the aspirin scale here, this shows usage pre-procedure, post one month, four months, and nine months, and the scale here is very expanded, 90 percent to 100 percent, and most patients certainly before the procedure were on aspirin.  There were a few patients who developed reactions that may have been due to aspirin, and because they were also on Clopidogrel, the physician stopped them, but approximately 98 to 99 percent of patients were taking aspirin at most points of the study.

            If you look at thienopyridine use, the vast majority of which was Clopidogrel, you could see it was extremely high, near 100 percent, both at baseline and lost procedure, and at one month and at four months it was still approximately 99 percent, 98 percent.  The drug was allowed to be stopped at six months, and at nine months there was still about 45 percent of patients that were still taking Clopidogrel.

            The primary endpoint of the     trial was target vessel revascularization at nine months, and going forward now, I'll be showing you bar graphs which will show the control stent in red and the TAXUS stent in green.  Target revascularization at nine months occurred in 12 percent of the control treated patients, and this was reduced approximately 61 percent, to 4.7 percent in the TAXUS treated patients.

            Now, the target vessel revascularization consists of actually restenosis of the target lesion which could be at the lesion itself or at the edges, and that is best represented by the clinical variable of target lesion revascularization.

            LTR is actually the best clinical surrogate for drug eluting stent efficacy, and TLR was reduced from 11.3 percent in the control group to three percent with the TAXUS stent group, a relative 73 percent reduction, and you can see the confidence interval here.

            The difference from TLR to TVR is other events that could occur even from side branches or new lesions that develop four or six centimeters away from the drug eluting stent that, of course, we don't expect a drug eluting stent to be able to either prevent or treat, and there were a few such events in both groups that weren't statistically significant.

            When one looks at the reduction in target vessel revascularization, it's important to note that it was caused by both a reduction in repeat percutaneous coronary intervention and repeat bypass surgery.  I think this is significant.  To my knowledge, it's the first time I've seen one stent compared to another where there's actually a reduction in the need for, of course, a much more invasive bypass surgery because of using one stent versus the other.

            This was highly statistically significant.  It's about a 69 percent reduction.

            Thus, summarizing the clinical events, the primary endpoint, nine month target vessel revascularization, that was met as evidenced by a 61 percent reduction in the TAXUS group versus control, and there was a 73 percent reduction in the nine-month target lesion revascularization rate, again, which is the best clinical surrogate for efficacy of a drug eluting stent.

            Let's now go to the QCA and IVUS analyses to learn more about the efficacy and safety of these devices.  Of the 1,314 patient sin the study, the nine-month angiographic cohort was prespecified actually pre and post specified -- I'll describe that in a minute -- as 732 patients.

            There were actually 536 patients from the original 1,172 patient cohort for which angiography was prespecified.  These patients signed a consent form up front for follow-up angiography.

            Now, in addition, because we wanted to get further evidence of efficacy and safety of these additional 32 millimeter stents, and these patients were for the most part enrolled after the follow-up angiography cohort was closed, we submitted a protocol amendment and received permission to get additional follow-up in an additional 196 patients in the 32 millimeter cohort.  This brought the total to 732 millimeters. 

            Now, these patients had not originally signed a consent form for follow-up angiography.  So these patients had to be recontacted and asked if they would again participate in the follow-up angiographic phase. 

            Overall, nine month angiographic follow-up was completed in 76.4 percent.  This included 82.5 percent of the original cohort, who you had agreed up front  that they were going to come back for follow-up angiography, and approximately 60 percent of the patients who were contacted in a post hoc pattern to ask them if they would come back with 32 millimeter stents for follow-up angiography.

            When one looks at the baseline characteristics of the angiographic subset versus those not in the angiographic subset, you can see that the age, gender, and the presence of diabetes mellitus, which is, of course, an important determinant of restenosis, was similar in both groups.

            When one looks at the two most important quantitative coronary angiographic determinants of restenosis, that is, vessel size and lesion length, you could see that the vessel size was similar in the patients prespecified and not prespecified for angiographic follow-up.

            Because, however, we enriched the angiographic subset with more of these 32 millimeter long lesions, there was an overall lesion length of 14.4 millimeters in those patients prespecified for angiographic subset versus 12.1 percent in those not prespecified.  So we would expect, again, a somewhat higher restenosis rate because of longer lesions in this group.

            When one looks now at the patients in the angiographic subset broken down by the control and the TAXUS groups, you can see similar reference vessel diameter, similar minimal luminal diameter and diameter stenosis, and the lesion length was 14.4 millimeters in both groups.

            Now, to show you the analysis methodology, which is used at the Core Laboratory, we break down the stented segment into the stent itself, which goes from the edge to the edge of the stent, which one can see by angiography, and then there's five millimeters on both sides of the stent, the proximal and the distal edge, which is also analyzed.

            So we will refer to "in-stent" measures, which goes from edge to edge.  We will refer to the proximal edge and the distal edge, and then we will also refer to the entire analysis segment.

            When we talked about target lesion or vascularization, that was due to events over the entire analysis segment, and of course, any event occurring over this area will relate to restenosis, clinical restenosis for the patient.

            Looking in stent is actually a very pure way to look at the efficacy of a drug eluting stent, but looking overall at the analysis segment shows the efficacy of the stent, the drug itself, how it diffuses at the edges, takes into account balloon overhang issues and edge trauma issues, et cetera, and again, this is what affects clinical restenosis.

            This is the cumulative frequency distribution curves for the TAXUS stent in green and the control stent in red.  So to orient you here, this is an increasing percentile of patients with this value of percent diameter stenosis both at pre and at post and then at follow-up.

            You can see it pre.  If you go here to the 50th percentile, this how you measure the mean percent diameter stenosis at pre.  Immediate post you can see it is reduced significantly, and the pre and the post measures are identical for the TAXUS and control stents.

            At follow-up, you can see that the diameter stenosis increases significantly in the control stent and it increases to a much lesser degree in the TAXUS stent.  You can actually see that the TAXUS stent for any level of measurement has a very nice, small amount of tissue growth or neointima covering the stent, but much less than the control stent.

            Thus, when one looks at the binary restenosis rates at nine months, which, of course, has stood the test of time as being the single angiographic parameter that most predicts clinical restenosis or the need for repeat procedures.  You can see in stent, again, looking at the potency of the drug just in the stent, restenosis was reduced from 24.4 percent to 5.5 percent, and then looking at the more important analysis segment, restenosis was reduced from 26.6 percent to 7.9 percent.  So a relative 73 percent reduction and 70 percent reduction, of course, both highly statistically significant.

            When one looks at the mean percent diameter stenosis, this is the entire analysis segment, and while you can see there's a marked reduction in the mean percent diameter stenosis within the stent, you can also see a positive effect at both edges.  Both at the proximal edge and at the distal edge, the mean percent diameter stenosis is smaller with the TAXUS stent than it is with the control stent showing some drug efficacy and system efficacy in those locations.

            This slide shows the late loss at nine months.  Late loss is actually the amount of tissue that regrows by quantitative coronary angiography after the immediate post implantation result.  You can see the overall late loss over the analysis segment was .61 millimeters in the control stent, and that was reduced to .23 millimeters.

            When one breaks it down into three independent sections, you can see within the stent there was a reduction in late loss from .92 to .39 millimeters, and there was also a significant reduction in late loss both at the proximal edge and at the distal edge, mirroring what we showed you for the mean percent diameter stenosis measurements.

            The Mehran classification looks at patients who do restenose to see how long and severe the restenosis is, and it is broken down into four patterns either focal restenosis, diffuse restenosis along the length of the stent but maintained within the margins of the stent, proliferative restenosis, which is diffuse, but also extends outside the stent and then total occlusion.

            One can see that when restenosis does occur with the TAXUS stent, a major difference compared to the controlled stent is that it is almost always focal.  There's a marked reduction in the combination of diffuse and proliferative patterns of instant restenosis.

            So, again, these focal restenoses are much easier for interventional cardiologists to deal with, which I think is probably the major reason why bypass surgery was lessened in the TAXUS stent group.

            There was a small incidence of total occlusions in both groups, and there was no significant increase in the TAXUS group.

            We can learn a lot of the patterns of neointimal hyperplasia from looking at intravascular ultrasound.  There were 268 patients that were prespecified for follow-up intravascular ultrasound.  This was completed in 199 patients, or 74.2 percent of the patients.  A very rigorous, precise requirements for the technical adequacy of intravascular ultrasound was required to be able to do detailed volumetric analysis.

            As a result 38 patients were excluded.  So 161 patients comprised the IVUS follow-up analysis, broken down in to 80 control patients and 81 TAXUS patients.

            When one looks here at the results, again, in the control in red and the TAXUS in green, looking at the vessel volume, this is the actual vessel itself.  One can see a similar vessel volume in the TAXUS in the controlled stent.  One can see that the stent itself was the exact same size.  The main difference that we saw was that the lumens were larger in the TAXUS group compared to the control group.

            The reason the lumens were larger, of course, given the stent size was the exact same, was that the amount of neointimal volume was decreased.  So this is a volumetric analysis now, and looking at the millimeter cubed of neointimal volume, this was reduced from 41 to a mean of 18 with the TAXUS stent, and looking perhaps at what has become accepted as the best IVUS indicator of drug eluting stent efficacy, the percent in-stent net volume obstruction over the length of the stent, this was reduced from 29 percent tissue accumulation within the controlled stent to 12 percent tissue accumulation within the TAXUS stent.

            I think it is actually important to point out that this is not zero or one percent, but there is some neointima covering the metallic stent struts.  It's just significantly less than what one would see with a bare metal stent.

            Thus, the QCA IVUS summary supports that in-stent  and analysis segment binary restenosis rates were reduced by 77 percent and 70 percent, respectively in the TAXUS group compared to the bare metal stent group.

            There was lower late lumen loss and a greater mean diameter stenosis at the edges, as well as in stent.  All of the QCA variables were significantly improved in TAXUS and all of the IVUS variables were also significantly improved with the TAXUS stent compared to the bare metal control, and the IVUS improvements were consistent with the QCA results.

            Now, let's move to some of the safety outcomes and what we'll look at overall now, taking clinical angiographic and IVUS parameters is looking at major adverse cardiac events, stent thrombosis, incomplete apposition, aneurisms, and other non-MACE serious adverse events.

            This slide shows the overall MACE rates in the trial.  There was no difference in cardiac death between the two groups, and there was no difference in myocardial infarction within the two groups, and what we have learned is that stents compared to PTCA and drug eluting stents compared to non-drug eluting stents don't change the rates of death or myocardial infarction.

            What they do do is reduce restenosis as evidenced by the reduction in TLR and TVR, which I showed you before.  As a result, a composite MACE rates adjudicated to being due to the target vessel were reduced from 14.4 percent to 7.6 percent, an approximate 50 percent reduction, and the overall MACE rates related to any target vessel were reduced from 15 percent to 8.5 percent, both highly statistically significant reductions.

            This slide shows the subacute thrombosis rates, and as I mentioned, we tried very diligently in a very sensitive manner to look for all episodes that might be related and adjudicated to subacute thrombosis.

            In hospital, there were very few subacute thromboses.  There was two in the control arm and one in the TAXUS arm.  Out to 30 days there was two and two, and from 31 to 180 days, there was one and two.  So overall -- and there were none, very importantly, from 180 to 270 days.  This is the period after which the Plavix was stopped.  There was zero percent subacute thrombosis in both groups.

            When one sums all of this up, the overall incidence of stent thrombosis was very low in this study in both arms.  Even using the sensitive definition, .8 percent, or five patients, in the control group; .6 percent, or four patients, in the TAXUS group, which was not statistically significantly different.

            Looking at incomplete apposition, and this is an IVUS analysis, you can see that after the procedure, there was statistically no difference in the stents that just by operators happen to be placed slightly and completely opposed to the vessel wall.  Numerically this appears like it might have been slightly more by chance in the TAXUS group than the control group.

            However, what's more important is that over the nine-month period, the gaps get filled in, and there was no difference in the overall late acquired incomplete apposition being very low in both groups.

            Now, when one does a paired analysis, when you have an incomplete apposition immediately after the procedure, it can either resolve; it can be persistent.  What we're most concerned about with drug eluting stents is there can be late acquired incomplete apposition, and that is the incomplete apposition was not there at the end of the procedure, but it develops over time possibly as a response to the drug effect causing the vessel to grow in an etatic (phonetic) or aneurismal fashion.

            And one can see the most feared type of incomplete apposition, late acquired apposition was present in 2.2 percent of control patients and 1.1 percent of TAXUS patients.  There was no correlation within the study of incomplete apposition with MACE or safety events.  So this does not seem to be much of a concern with the TAXUS stent.

            When one looks at aneurisms, we use quantitative coronary angiography to have a very sensitive definition of aneurisms.  Aneurism was defined as being any growth within the lesion 20 percent or greater than the reference vessel diameter.  So 1.2 to one ration, which is very, very sensitive.  Our average stent, that balloon ration was 1.2 to one.

            And as you can see immediately post procedure, aneurisms were called by the core laboratory in a very small number of patients, no difference between the two groups, as one would expect, and in a nine month follow-up, there were two aneurisms in both groups, and again, you can do the same sort of paired analysis, and the type that we would most worry about is the late acquired aneurism, and as you can see, there was one late acquired aneurism with a bare metal stent and zero late acquired aneurisms with the TAXUS stent, and it was very difficult to relate aneurisms with events at all.  There was only one event in any patient with an aneurism.  So this also does not seem to be a safety concern.

            We looked at site reported serious hypersensitivity reactions.  They were mild reactions and may have been due to hypersensitivity according to the site and about three percent of patients in both groups, but these were mild and easily resolved.  Looking at serious adverse events due to hypersensitivity reactions, there was only one patient that had a purported drug hypersensitivity reaction.  This happened to be in a TAXUS patient.  It's the same patient up to 30 days.  There were no additional patients from 30 to 270 days.

            This was a patient who had an allergic reaction, which was a rash and a headache that the site physician felt was due mostly likely to clopidogrel, and this resolved after discontinuation of the clopidogrel.

            There was one other patient in a TAXUS patient that had a nonspecified hypersensitivity reaction, again, occurring within 30 days.  This was an allergic reaction with a swollen tongue and shortness of breath, which was felt to be due by the site to either an iron supplement or an ACE inhibitor and the patient completely recovered after discontinuing those medications.

            Thus, looking at the safety data, comparing the TAXUS stent to the control bare metal stent, the nine month overall major adverse cardiac event rate was reduced by 43 percent; target vessel failure a little bit greater.  Stent thrombosis rates were rare and comparable, less than one percent at both groups, the TAXUS and bare metal stent group.  There were reportedly no stent thromboses reported after six months, after the protocol mandated clopidogrel was discontinued.  The incomplete apposition rate was low and comparable through nine months.  Aneurisms even using extremely sensitive definition were rare and comparable through nine months.  There were importantly no increase in either late acquired incomplete apposition or later aneurisms, and hypersensitive reactions were extremely infrequent, comparable to control, and seemed to be related to study medications.

            Finally, I want to show you some of the important subgroup analyses, and these were all prespecified analyses within the protocol.  This shows you the main subset analyses for target lesion revascularization, thus looking at efficacy, and what you see here are the main parameters which tend to affect restenosis.

            The most important three are the diabetic status of the patient, the reference vessel diameter, and the lesion length, and in some studies LAD versus non-LAD vessel has been related to restenosis.

            You can see here the relative risk reduction in favor of TAXUS compared to control for reduction in clinical target lesion or vascularization, with a confidence intervals in yellow and the line of unity in one.  And as you can see, there's an approximate 70 to 75 percent reduction in target lesion revascularization at nine months independent of all these different parameters.

            Of course, you will note here, and I'll get to some of this in upcoming slides, but the rates are significantly higher, for example, in much longer lesions compare to much shorter lesions, and as a result, the absolute degree of benefit is greater in the more complex lesions, but the relevant degree of benefit extends across all subgroups that were examined.

            Now, when one does what becomes standard  as a Kalon-Ho diagram, and that is it looks at the interaction between lesion length and vessel diameter, the longer and longer the lesion and the smaller and smaller the vessel by quantitative coronary arteriography, you'll see the clinical restenosis rates as well as the angiographic restenosis rates rise exponentially, and that's what we saw here.

            This shows you with longer lesions and smaller vessels broken down by tertials (phonetic) here into these nine separate groups that in general with the bare metal stent you can see restenosis rates significantly increased, and this is, again, clinical restenosis, the need for repeat procedures adjudicated on the basis of angiographic restenosis.

            You can see in the TAXUS group there is a significant and in some cases marked reduction in all subgroups regardless of lesion length, regardless of vessel size.  It certainly on an absolute measure is greatest in the most complex lesions, but even in the more simple lesions to treat by bare metal stents within the confines of this study, there's a reduction in target lesion or vascularization with all lesion lengths and vessel diameters.

            This is the exact same slide, but now looking at angiographic restenosis along the entire analysis segment, and one can see again about a 70 percent to 75 percent reduction in angiographic restenosis, which is consistent across diabetic status, LAD, non-LAD, small and large vessels, and short and long lesions.  And I'll show you some of these specifically in slides coming up.

            This looks at a function of angiographic restenosis, and the analysis segment is a function of reference vessel diameter, and one can see with bare metal stents as the reference vessel diameter decreases, the angiographic restenosis rate almost linearally increases along this scale, along with the TAXUS stent though, you can see there was a significant reduction in angiographic restenosis which did not quite reach statistical significance in the smaller group, although there's still the same trend here for an approximate 65 percent reduction in angiographic restenosis.

            You can see the marked benefit especially in these small vessels.

            Similarly, when one looks at lesion length broken up into approximate tertiles (phonetic), less than ten, ten to 20, and greater than 20, you can see that as the lesion lengths increase in length, the angiographic restenosis rates increase at least linearally, and you can see here significant reductions even with the short focal lesions with the TAXUS stent, with the intermediate lesions, and with the longer lesions greater than 20.  And these patters were also exactly replicated when one looked at the stent diameters, either 2.5, 3.0, or 3.5 TAXUS stent, or when one looked at the length of stents, 16, 24, and 32 millimeter stents.

            Finally, the diabetic subgroup is very important, and here you can see is the overall angiographic restenosis in the patients without diabetes, showing a marked and significant reduction.  In the overall diabetics, if anything, the relative reduction is even greater, and a marked reduction from 34 and a half percent to 6.4 percent, and in our most difficult group to treat, the insulin dependent diabetics, you can see with the bare metal stent the restenosis rates go up from 24.4 in no diabetics to 42.9 in insulin dependent diabetics, but the angiographic restenosis rates with the TAXUS stent were approximately level in this study.

            There are certainly relatively small numbers of patients in the insulin dependent diabetic group.  So we can't make firm conclusions, but at least with this very large, approximately 80 percent reduction showing the treatment effect, I think this is highly likely to be very significant.

            There was a lot of question raised based on an earlier clinical trial of looking at Paclitaxel using a nonpolymeric based system, that there may be a significant interaction between glycoprotein IIb/III inhibitor use, and so we wanted to put this to rest and looked at patients who had IIb/III inhibitors versus those who did not have glycoprotein IIb/III inhibitors, and there was no difference in either clinical or angiographic efficacy on the basis of using IIb/III inhibitors.

            And finally, while this trial was a single lesion, single stent study, as I showed you, approximately ten percent of the patients had multiple stents implanted, and that's important to be able to implant multiple stents.  These were for the most part study stents when required.  You can see there were a total of 45 patients who had multiple study stents in this trial, including 20 patients who had a 16 plus a 16; 13 patients, a 16 plus a 24 millimeter long stent; 11 patients, a 16 plus a 32; and one patient had three study stents.

            When one looks here at the TAXUS improvement in patients with multiple study stents, with a single study stent you could see the reduction in clinical restenosis, and we are gratified to see that while the numbers are relatively small, there was no instance of clinical restenosis or target lesion revascularization in any of the patients in the TAXUS group that had multiple study stents implanted, and similarly, while the angiographic numbers are even smaller, only 15 patients in the TAXUS group, there was no instance of angiographic restenosis in TAXUS patients who had multiple stents.

            You can see with bare metal stents, even with the small numbers, you put in more stents for, you know, complications or for longer lesions; the restenosis rates go up, but there does not seem to be at least with these small numbers increase to the TAXUS group.

            Thus the conclusions from the pivotal TAXUS IV trial were that the nine-month results demonstrate safety of the slow release TAXUS stent with no increased risk of stent thrombosis and no increased risk of aneurism formation, late total occlusions or incomplete stent apposition, and a reduced overall risk of both major adverse cardiac events and target vessel failure.

            There was a reduction in 70 percent and 77 percent, respectively, for analysis segment and in-stent angiographic restenosis and 61 percent and 73 percent, respectively, for TVR and TLR, resulting in lower rates of bypass graft surgery and repeat percutaneous coronary intervention.

            Thus, I think the clinical results mirror the preclinical results that are shown to you.

            And finally, effectiveness was demonstrated in this large pivotal trial across a wide range of complex patients and lesions, including small vessels, long lesions, and patients with diabetes.

            Thank you very much.

            DR. RUSSELL:  Thanks, Gregg.

            And now a few concluding remarks.

            Clinical safety for the TAXUS slow release stent has been evaluated in three randomized control trials involving more than 1,900 patients.  Clinical safety profiles were comparable to control with low rates of stent thrombosis, myocardial infarction, incomplete apposition, aneurism, and other adverse effects.

            Boston Scientific has a commitment to the ongoing tracking of safety events, in particular stent thrombosis.  Here is our most recent records, as of November 14th, 2003.  We track these results in our preapproval studies, post approval studies, and also in customer reporting.

            You've heard about the results from our completed studies where TAXUS  is unblinded from control, and there were eight out of 987 TAXUS patients that had a reported stent thrombosis.  In the ongoing studies, TAXUS IV, V, ISR, and TAXUS VI, there have been 13 reports of stent thrombosis out of 1,747 patients.  That, of course, includes the control and the TAXUS group.

            Of note, one patient had a periprocedural stent thrombosis complicated by two recurrent stent thromboses.  So that accounts for three of the stent thromboses in the 13.

            In our post approval studies, we have programs that are sponsored by Boston Scientific.  That includes the WISDOM and MILESTONE II registries.  There have been 11 reports out of 3,072 patients that have received the TAXUS stent.

            And then we have investigator sponsored initiatives where we have personal communication as of this week of the latest stent thrombosis events, and there are seven out of a total of 932 patients.

            With respect to customer reporting, there have been eight reported stent thromboses out of more than 69,000 units sold.  So the total number of individual stent thromboses reported includes 47 out of more than 75,000 stents implanted or sold for a percentage of 0.06 percent.

            The majority of these stent thromboses which we carefully tracked occur in the periprocedural period.  The ones that have occurred subsequent to that have been examined in detail and often have extenuating circumstances associated with them.

            We're committed to ongoing tracking of long-term safety.  It's evident in our preapproval trial programs.  We have follow-up out to five years, and we've implemented dynamic safety reporting.  This was done in our trials, typically Web-based systems where we can track the events with ongoing reporting in a relatively frequent basis to FDA.

            We're committed to post approval surveillance registries, including transitional registries, such as WISDOM and MILESTONE II that were specifically set up to track such events and our proposed ARRIVE U.S. periapproval registry.

            ARRIVE stands for TAXUS periapproval registry.  It's a U.S. Web-based, two phased program, multi-center and focused on safety surveillance to capture real world safety profiles.  The principal investigators are Dr. John Lasalle and Dr. David Cox.

            There are two phases, a preapproval phase.  We have a submission in for an expanded access registry as part of our IDE, and then Phase 2 will involve post approval, including 2,000 patients in 50 community type centers, reflecting the real world use of drug eluting stents.

            We're focused on consecutive patient enrollment per participating physician, and that would be any use of a TAXUS stent.

            We will be tracking TAXUS stent related thromboses MI, reinterventions, death and other events, and using an independent committee to classify these events.  This will allow us to have dynamic updates of both enrollment and safety status.

            You've heard about the effectiveness of the TAXUS slow release stent from Dr. Stone in the form of TAXUS IV, further supported by TAXUS I and II in more than 1,900 patients that have been evaluated.  We've demonstrated TAXUS superiority compared to control with concordant improvements in clinical outcomes, QCA and IVUS endpoints.

            Taken together, the taxus data demonstrates safety and effectiveness for the TAXUS slow release stent, and we believe that the benefit of the TAXUS slow release stent out weighs any potential risks.

            Thank you very much.

            CHAIRMAN LASKEY:  Thank you for a most comprehensive presentation, and congratulations on completing ahead of schedule.  That, of course, leaves more time for the panel to ask pertinent questions prior to our break.

            So, again, I'll remind the panel that they will have another opportunity to query, but right now are there any questions from panel members to the sponsor on matters presented in the last hour?  Starting with Dr. White.

            DR. WHITE:  Thank you.

            Dr. Russell and Gregg, a very nice presentation.

            I want to ask a question focused on the histology because I'm just a little confused, and I don't know whether you have the same panel packet I have, but I'm looking at in our packets called the "Summary of FDA Review Memos," and I referring to page 13 in that section where there's a summary of some pathology.

            And it says RVF 01-049, and ten there's several other as we go through, and they summarize histology, and in virtually every one of the histologic reports it says that at 28 days the endothelialization is partially complete.

            And you said several times in the presentation that at 28 days it was complete, and I haven't been able to find any data in the submission that says that that's true.  So I'm wondering why the disparity.

            DR. RUSSELL:  Well, I'm going to refer this question to our study pathologist, as well as our study preclinical expert, Drs. Wilson and Schwartz.  Part of this has to do with the use of statistics.  So we used statistics to analyze differences between control and TAXUS stents for ordinate grading scales, and when one compares across multiple programs with ordinate scales, it can be difficult to determine if there's a difference.

            But I think Dr. Schwartz and Dr. Wilson.

            DR. WILSON:  So I'm Dr. Greg Wilson.

            In terms of disclosure, I'm a paid consultant to Boston Scientific Corporation.  I have never and do not now hold stock or stock options in the company, either directly or through family members or related entities.

            Now, I'm the pathologist who developed the histological grading systems for the preclinical animal studies.  I authored some of the pathology reports on those studies, and more importantly, I have reviewed all of the pathology on all of the preclinical animals studies, including looking through over 4,000 glass slides with mounted histological preparations.

            The question you asked about the endothelialization really comes out of a modification that was made in our grading system.  Initially the grading system was simply complete or incomplete, and complete initially was interpreted to really truly be absolutely complete, that is, 99, 100 percent endothelialization as assessed on histology.

            Now, when you look at a histology slide as distinct from looking at scanning electron microscopy, and there was some scanning electron microscopy done in the studies as well on some of the explanted stents, but when you are grading endothelialization based on histology, you're basically looking for the nuclei of the endothelial cells on the lumenal surface.  The endothelial cells are very thin, tenuous cytoplasms.  So, in essence, you're really looking at the numbers and how close the nuclei are to each other for the endothelial cells.

            And so the result of this is that there is a bit of uncertainty in terms of actually saying whether there is truly 100 percent endothelialization.

            Now, as these animal studies progressed, we modified the grading system to have three categories:  less than 75 percent, 75 percent to 90 percent, and over 90 percent endothelialization.  And when we looked at the data in that way, at 28 days with the drug at both the slow release and moderate formulations compared to control was equivalent.  That is always over 90 percent endothelialization as reflected in that grading, and I'll comment that for the controls, just for the drug, there would be instances where it would be hard to tell whether it was, you know, 99 percent, 95 percent.

            There was scanning electron microscopy done on a number of stents that were explanted.  It's not possible to do the scanning EM and the histology on the same stents because, of course, this histology is destructive.  You have to cut through the stents.

            But on those stents that were submitted for scanning EM, which has the advantage that you're looking at an on fast preparation, you're looking at the entire flow surface, and so that you can make really a better evaluation of endothelialization and be more confident about whether it's really complete.  We found there was equivalence between the control, the slow release and the moderate release at 28 days, and subsequently, and indeed, the endothelialization was remarkably complete.

            I can't say absolutely perfect in every specimen, but certainly no detectable differences between control and the drug.

            So I'm afraid the material, I think, that you've received reflects the sort of earliest just complete, not incomplete grading.

            DR. RUSSELL:  Thanks, Greg.

            Next question?

            CHAIRMAN LASKEY:  Chris?  Dr. Somberg.

            DR. SOMBERG:  I appreciate the difficulty.  I just think that the presentation -- I mean, one of the concerns about the stents that we're all listening to and obviously the presentation takes pains to point out is that there is  healing and that there's intimal thickening and the IVUS and everything else shows that.  So I don't have any doubt about that.

            I just think it bothers me when the data is not congruous.  I mean, the data that we're given to review says that there's not complete endothelialization, and the presentation says there is.  I understand that the pathology is not black and white, but maybe the presentation could have reflected that a little bit  better just to be clear.

            DR. SOMBERG:  John Somberg.

            I have a couple of questions.  The first one for Dr. Russell is dose consideration both for the SR and the MR.  I would imagine, as with the pharmaceutic, you would have looked at a number of different doses with the polymer and then try to establish a minimal or a borderline.

            Are we at that borderline with the SR and the MR formulations or are we far beyond that?  There's no mention of any sort of dose studies, and this matter was sort of magic polymer on the stent and it works, but I just wondered where are we along the minimal effective concentration gradient.

            DR. RUSSELL:  Yes, that's an excellent question.  What we discovered in the early foundations of our program is the importance of preclinical models, and perhaps I can get Dr. Schwartz to comment on this following my brief comments.

            And we are very confident about the reproducibility of our preclinical implant model for safety evaluations, but our model produces a very thin neointima, which was inadequate for efficacy evaluations.

            So our dosing decision was made by establishing formulations that were clearly safe by detailed histologic analysis, and efficacy is being studied in the actual human atherosclerotic disease milieu, so to say.


            DR. SCHWARTZ:  Thank you.

            My name is Rob Schwartz.  I'm from the Minneapolis Heart Institute Foundation.  My conflicts of interest, I own no stock, no options in Boston Scientific or any subsidiary.  My laboratory receives research grants from Boston Scientific as well as from all of the other major competitors.  I am a paid consultant, but the funds go to my laboratory and not to me personally.

            Boston Scientific did reimburse my air fare and hotel here.  I have received honoraria for talks about Paclitaxel and stents in the past that have been personal, but I also receive honoraria and give talks for all of the other major competitors.

            Concerning the issue of animal models, a terrific question.  In terms of safety, there are several things that are important.  Preclinical safety represents animal death.  I think that we all know, those of us who do preclinical work, that if a stent or a technology put into the coronary artery is not safe, the animal typically dies in the case of a pig within 12 to 24 hours.

            It was rather remarkable in this study.  I find that there were zero stent related deaths.  Typically in the guidelines that we published and are working out with FDA, in fact, typical mortality is ten percent, but there were no mortalities in this.

            So I think in terms of safety during the endothelial question, regarding the neointima, I think we're very comfortable that there is a great deal of safety in a preclinical/clinical context.

            From there, as we move to your questions about dose and about the effect of the dose on the tissue itself, what appears to me to be key is cellularity.  Is the media intact?  Are the medial cells killed?

            There was a little bit of cell dropout in the media of these vessels, but not much.  The majority of the media was preserved and certainly the media in between stents looked quite normal.

            The endothelium, as we heard, was virtually intact, mostly intact.  We know very well that as little as ten to 15 percent of a vessel endothelialized can function as a normal vessel.  So from an endothelial standpoint, from a medial standpoint, the doses that were chosen, the doses that were used, I think, panned out very well in a cellular context.  In fact, that translated into a preclinical safety context, animal death, and, indeed, as you heard from Dr. Stone, these safety issues were not issues, were not concerns all the way up to the 1900-plus patients who have received these stents in the TAXUS trial series.

            DR. SOMBERG:  But essentially you're saying you do not know on what point of the dose response curve this formulation is in terms of the minimal effective concentration.

            DR. SCHWARTZ:  We don't know the low end.  We do know the high end.  As a result of a recommendation, we did do a safety margin study with doses between four and ten times as high.  In the highest dose, ten times we, indeed, did find medial cell dropout at that point.  There was perhaps more neointima at those levels, but this was the purpose of the study, was, in fact, to find more the higher dose rather than the lower dose.

            I think, again, the safety issues --

            DR. SOMBERG:  Some place between that range you might have had a greater reduction in restenosis or no change from what you have or you could have used half this dose in the polymer and still gotten a prevention of restenosis.

            DR. SCHWARTZ:  That is true.  Hard to answer that question.  We, again, know the safety end and the high.  We do not know the efficacy end on the low, at least from the preclinical, bearing in mind, once again, that the preclinical is not a good efficacy model at this point.

            DR. RUSSELL:  And what may not be reflected in your briefing document is the foundation for the studies that were presented.  So the program was founded based on evaluation of six different polymers to identify the one that was the most vascular compatible, the one that could retain and release the drug at the low level, which was our design principle, and then there are a number of variables that could be explored with this technology. 

            So one can change the dose density.  One can change the formulation, one can change the coating thickness, and we did a series of early studies looking at various combinations of other formulations and other doses that did not have the extensive preclinical analysis that I reviewed earlier to come to our decision to focus on the one microgram per millimeter dose squared dose density, and on the slow and moderate release formulations where we decided to delve deep into the histopathology so that we could at least understand overdose density with the same formulation and over time.

            And these studies are quite painstaking.  If you're looking at animal studies out to 360 days and the histologic analysis, Dr. Wilson looked at thousands and thousands of sections to establish that these two formulations with this one dose that we selected had a safe profile that was reasonable to pursue, proof of concept in the TAXUS program.

            So that's the history behind our dosing rationale.

            DR. SOMBERG:  You led into my second question to you, Dr. Russell, and that was really why.  I'm a little confused on the issue.  There's an SR and MR formulations.  What is the company's, I guess, plan?  Are we going to have two different formulations?  Are we going to try to get different indications for that?

            You know, you've presented very nice information about the difference in release characteristics, but I'm a little confused on how clinicians are going to make use of that pharmacologic information.

            DR. RUSSELL:  Yes, I can understand and appreciate that question.  TAXUS II was designed to compare based on very sensitive IVUS assessment the difference between the slow and moderate release formulations, and as I mentioned but did not elaborate on, the IVUS findings were very similar between the slow release and moderate release program.

            And I must point out that the study was not powered to show a difference between the slow release and moderate release formulation, but nonetheless the percent net volume obstruction was nearly identical at six months.

            We spent some time analyzing this, and our conclusion was one possible explanation was that we had reached the threshold, the dosing threshold, with the slow release formulation and there was no additional benefit for the three times higher release with the moderate release formulation in this standard low risk lesion set.

            And so we actively pursued TAXUS VI, which is a higher risk lesion subset, longer lesions, has more diabetic patients in it for a European study and are awaiting the moderate release results from this higher risk lesion subset to make a decision about whether the moderate release formulation is of added value to the slow release formulation in more complex lesions

            DR. SOMBERG:  Another question I had and it relates, I think, to the preclinical models:  is there a plan to continue beyond 24 months?  Taken with the stent is going to be there forever and there's about 80 percent retention of the active agent there, we know things last for a while and then all of a sudden the device fails at some point or a polymer may change its release characteristics.

            Do we have any idea if that will occur and how long are we going to follow that out?

            DR. STONE:  Well, the patients are being followed for five years.  So yearly for five years in all of the TAXUS trials.  So at least, you know, to that extent and that duration, you know, will have good clinical follow-up.

            You know, we expect that since the stents will be by then, you know, by six months, if not a year, completely endothelialized with a neointima separating the stent struts from both the blood vessel and whenever there's neointima growing on one side, it grows on the other side.  We expect the stent struts, the polymer and whatever Paclitaxel is left to be pretty much sequestered.

            But I think just in case there's any problem, that's one of the reasons why the patients are being followed out to five years.

            DR. SOMBERG:  I appreciate that.  I just would think you could also have a parallel animal evaluation of this since it has been said that it's sensitive to the  potential toxic adverse effects.

            DR. SCHWARTZ:  I might add that preliminary discussions, in fact, with Dr. Zuckerman, as an interest, these are ongoing in relationship to putting a study together to evaluate just that, to understand the relationship between short/long-term animal data with the clinical data.

            So these are under discussions at present.

            DR. RUSSELL:  Yeah, and we do have one group of animals that could be sacrificed at 24 months.  There's some debate about whether we should sacrifice them at 24 months or let them go to 36 months, and I think the ideal time point for that group of animals is something that I'd appreciate feedback on because it's quite a debate.

            DR. SOMBERG:  And two questions I just have for Dr. Stone, if I may just to finish that up, that came directly from your presentation.

            One is I was struck by the small number of diabetics.  There was about eight or nine percent in the population and probably because of a bias, but maybe you could comment on that, and what severity of diabetes.

            You know, you said insulin dependent, but that can be a variability here, and also the additional follow-up.  There's one cut point chose, nine months.  Is there any other follow-up planned?

            DR. STONE:  Well, there were about 25 percent diabetics in the study.  So actually there was a pretty good representation of diabetics.  Probably in our clinical practice we see with all comers much more complex disease that was in TAXUS IV even, probably about 30 percent diabetics.  So it was pretty representative with 25 percent diabetics.  About eight percent of them were insulin requiring diabetics, and I think that's also a reasonable representation that we see in most studies.

            European trials, for example, tend to have much fewer diabetics than was represented in TAXUS IV.

            Regarding the second question was about extended follow-up.  Since you asked, and I understand that since you asked I'm allowed to talk about it, we now have the 12-month TAXUS IV data, and the curves continue to spread between showing extra clinical efficacy.  Maybe I can show you one or two quick slides.

            This actually shows the events now that have occurred between nine and 12 months in the TAXUS IV cohort, and particularly striking to me is that first you can see there was one cardiac death and seven myocardial infarctions in the control patients and no deaths from myocardial infarctions in TAXUS patients.

            And this incremental reduction in myocardial infarction between nine and 12 months was actually significant at the .007 level.

            In addition, you can see if you look at TLR, again, the best clinical surrogate of drug eluting stent efficacy, there were 26 additional TLR events between nine and 12 months in the control group and only nine in the TAXUS group.  So four percent versus 1.4 percent.

            So overall the difference, therefore, in TLR and TVR, in MACE and in target vessel failure continues to spread favoring the TAXUS stent.  And it's not surprising that there's a difference in duration of when we see patients come back for TLR or TVR compared to angiographic restenosis.

            We've known for years for the most part most angiographic restenosis occurs between three and six months or is complete between three and six months, but it takes patients longer.  They have symptoms for a couple of months and then they have to schedule an appointment with their doctor, and then they finally get referred for a repeat angiogram.  Maybe they have a stress test.

            So there's a delay which can be anywhere from three to six months between restenosis and clinical events.  Thus, if you look here at the overall event rates here, 12 month MACE and target vessel failure, there's no difference in cardiac death.  If anything, there's a weak numerical trend now overall from zero to 12 months favoring myocardial infarctions, less in the TAXUS group, a big difference in TLR, a big difference in TVR, and again, a bigger reduction in MACE and target vessel failure than we're seeing just at nine months.

            And finally, I think this curve kind of shows the change in efficacy very nicely.  This is target vessel revascularization from zero to nine months, with an 8.2 percent delta by actuarial statistics, and now when you carry it out to 12 months, there's actually a ten percent difference in target vessel revascularization.

            So all of the evidence that we have so far is that the incremental benefit continues to increase in the patients who receive the TAXUS stent.

            DR. NORMAND:  I have  a clarification of the slides, and it sort of carries over to many of the presentations.

            DR. STONE:  Sure.

            DR. NORMAND:  I notice the sample size is always 652 and six whatever the number is, and is that in 12 months you really have follow-up on everybody?

            I don't want to look at the Kaplan-Meier.  I want to look at your --

            DR. STONE:  No, it was about 97 percent.  In fact, it's a little bit higher at 12 months than it was at nine months, but it's about 97 percent at both time periods.  This curve, of course, takes into account the different time effects.

            DR. NORMAND:  But the other one?

            DR. STONE:  Yeah, the other ones, a true denominator would be approximately 97 percent of those numbers for each time period.

            DR. NORMAND:  No both groups?

            DR. STONE:  In both groups, and I put those on the first slide.

            DR. NORMAND:  So the percentage that you're showing are really the percentages of the true numbers or they're the percentage of the numbers at the beginning?

            DR. STONE:  Those are the percentages of the patients with follow-up.

            DR. NORMAND:  So the percentages reflect the different denominators.

            DR. STONE:  That's correct.

            DR. NORMAND:  And hence your P values reflect the fact that you have different denominators.

            DR. STONE:  Exactly.

            DR. NORMAND:  Thank you.

            Ten more questions?  Dr. Morrison.

            DR. MORRISON:  I'm trying to understand the difference between the slow and moderate release, and perhaps this is just ignorance of polymer chemistry.  So I apologize for that, but it looks to me on your diagrams that it's the same dose density.  What differs is the coating thickness, and in either case you're left with a majority of the drug behind.

            Is that the case?  And I think Dr. Somberg's question was directed at the notion of is this truly inert because there's an endothelial layer.

            Are you concerned about the fact that there's -- with the slow release what we're discussing today primarily, it's less than a tenth of the dose that's actually been delivered and sure about what's going to happen with what's left behind is my first question.

            The second is wouldn't it make more intuitive sense to try and get it out and get it in action the first 30 days and get rid of it so that it's not left behind to cause problems 12, 36, 48 months down the line?

            DR. RUSSELL:  With respect to your first question, we did perform a series of studies looking at the retained drug, as I mentioned, and what circumstances would be required to get the retained drug to come out, and as you may recall from those slides, it took fairly extreme conditions that are not really consistent with living human beings to get the sequestered drug out, and that was in stents that really were not implanted and healed as they typically are in humans.

            And then, of course, we developed sensitive assays to see if we could determine whether there was any measurable Paclitaxel, and we were not able to do that, and hence, we turned our focus to the preclinical studies that allow us to use histologic assessments to see if there was any evidence of Paclitaxel activity even though we could not measure it.

            And what we discovered was that healing progressed in spite of the presence of the retained Paclitaxel and achieved many of our design objectives.

            With respect to your original question about the polymer chemistry and the release kinetics, I would like to refer that question to Dr. Kramer.

            DR. KRAMER:  My name is Phil Kramer.  I'm a paid consultant to Boston Scientific.  I have no other financial interest in that company or in any of their competitors, and as far as I know, I've never done any consulting for any of their competitors.

            To answer your question, basically I think you just have to appreciate what happens with this coating.  There's really three layers as Dr. Russell described earlier.  There's the surface layer, which is sort of here.  There's the immediate subsurface, and then there's the bulk.

            And essentially what happens with either the moderate or the slow release is that particles that are on the surface or drug molecules that are on the surface, which is sort of these purple dots, those are released first essentially by dissolving or by dissolution, and that opens up a channel so that particles that are in the immediate subsurface now have a pathway to get out, and so they diffuse to the surface and then dissolve just as the original surface molecules dissolve.

            Everything else is left in the bulk, really has no pathway to get out.  It's surrounded by polymer, and it's basically locked in place.  There's no way unless you were to do very extreme, nonphysiologic destruction of the polymer to get out.

            The difference between the moderate release and the slow release is the thickness of the coating, and in the moderate release, there's a greater percentage of the load that is either on the surface or in the immediate subsurface because it's essentially a thinner coating with the same amount of drug in it, and so you have a higher initial release because there's more on the surface and a greater overall release because there's more in the subsurface.

            But either way, you still have the majority left behind in the bulk.

            CHAIRMAN LASKEY:  Can I just ask a follow-up on that while you're there?  This is such a remarkably efficient and effective drug where really eight to nine percent of what's applied can do the job.  Yet are these -- just tell me about the variability and the measure here of the cumulative in vivo release kinetics.  It's not a flat eight to nine.  There's some variability around there.

            Are those error bars or SD bars or what am I looking at to see where those amounts released might be more than eight percent and could conceivably be more toxic?

            DR. KRAMER:  Well, those are standard deviation bars.  On the figure that Dr. Russell presented during her presentation, those are standard deviations.

            So there is variability, and the other thing that you have to realize is that that is measured by the percent released is measured indirectly by measuring the percent that remains because you explant the stent and then basically destroy the coating and measure what's left.

            And so because of that process, the process of explanting, removing tissue, et cetera, that's attached to the stent, there is more variability in that.

            DR. RUSSELL:  Yes.  It's a fairly labor intensive and awkward model to use in the long term.  The longer the stents have been implanted, the more imbedded and compacted the neointima is, as well as the adventitial response.  So we actually have to extract the stent physically and then digest off the tissue in order to do these measures.

            So you can imagine there is a moderate amount of variability with that process.

            CHAIRMAN LASKEY:  I understand that, but therein lies the existence or death of the neointima.  So that's just the clinician's standpoint.

            Dr. Yancy and then Dr. Weinberger.

            DR. YANCY:  Thank you.

            I, too, want to commend Dr. Stone for a very nice presentation and a well executed study.

            My two questions, and I'd like one follow-up afterwards, comes from the perspective of what I do, which is dealing with LV dysfunction and heart failure. Most of our clinical trials are very heavily inclined towards clinical events, hospitalizations, relief of symptoms.

            It would seem to me, reviewing the data that we had available, that in fact, these patients presented for elective stenting primarily because of an event or a reasonable clinical indication.  I appreciate the MACE data, but I don't see additional data regarding clinical events or the relief thereof after implantation, especially with your nine month and now 12 month data.

            I was a little bit concerned, looking at the nine month data because death and MI were virtually indistinguishable between bare metal and drug eluting stent.  The 12 month data you've just demonstrated are much more gratifying in that regard.

            I'm wondering what your perspective is on clinical events and whether or not the registries that Dr. Russell owns would, in fact, include clinical events.

            So that is the first point about the clinical events outside of death and MI.

            The second question refers to something I commented on during the presentations, but which was in the preceding data, and that is the exclusion apparently of patients with fairly extensive LV dysfunction.  Again, the rationale for that and what you think the applicability of this technology might be in that specific higher risk.

            So let me ask you to address those two questions, and then I'd like to follow up with one very clinical question.

            DR. STONE:  Well, thank you for the kind comments.

            I think when one looks at what you expect a drug eluting stent to do, it's really the same thing you expect a bare metal stent to do or the same thing you expect a balloon angioplasty catheter to do, and that is you are treating one very focal stenosis in a coronary artery.

            When one looks at the propensity for patients to die or to develop congestive heart failure or ventricular arrhythmias, I think it's a function of multiple different features.

            Certainly if a vessel goes on to close an infarct, but it's obviously a function of overall left ventricular function, ventricular irritability, and inducibility for ventricular arrhythmias, valvular disease, systemic factors such as hypertension and renal disease, et cetera.

            What we've learned over decades now of doing the angioplasty versus surgery studies and now angioplasty versus stent and now stent versus drug eluting stent is that the drug eluting stent we hope will reduce restenosis at that focal site.  We hope we can do that safely without hurting the patient, therefore reduce recurrent angina, reduce hospitalization, reduce recurrent ischemia, reduce repeat percutaneous and bypass procedures.

            Now, to the extent -- and this kind of answers the second question at the same time -- to the extent that in a patient with severe left ventricular dysfunction, and patients were excluded for an LVF of less than 25 percent in this study, but to the extent that patients with severe LVF who continue to have to have recurrent procedures are always at risk in any of the procedures we do for them and are always at risk when a vessel renarrows, you would think if anything, this device would probably have some incremental benefit in those patients.

            But I wouldn't expect it to a priori nor did we see a statistically significant effect on death or myocardial infarction, which one is a function of what's happening at the lesion and the MIs, but also vulnerable plaque and new lesions that appear elsewhere.  This device is not going to affect new lesions elsewhere or the progression of trivial stenoses to acute vessel closure.

            So I think we saw exactly what I would have expected us to see, and I think it's important for people to realize this device is not a panacea for heart disease.  It's treating a very focal stenosis, so relieving ischemia, and then hopefully it will not recur.

            DR. YANCY:  Let me follow up with the one clinical question.  Given the data which appear to be fairly robust, what do you think the remaining indication is for the bare metal stent?

            DR. STONE:  It's a great question.  I think there are certainly many subsets that weren't studied in TAXUS IV, that were excluded, that we don't yet have safety and efficacy data for with drug eluting stents, very small vessels, very long lesions, in-stent restenogic (phonetic) lesions, thrombotic lesions, acute myocardial infarction, chronic total occlusions, brachytherapy failures.

            In fact, in some regard, the universe of patients we don't have the answer for is probably as large as the ones that we do have the answer for now, and that's why, you know, we're strong believers that outside of carefully controlled research settings these devices should be used on label as evaluated within large pivotal trials such as this.

            DR. WEINBERGER:  I just want to echo the congratulations of the other members to Gregg and to the whole team.

            I would like to focus on something that hasn't been spoken about quite as clearly, and that is some of the more minor complications that were recognized and reported by the investigators, namely incidents of CVAs and hematological dyscrasias.

            Now, these are minor and we tend not to think of them as cardiovascular events, but given that we're putting a new pharmacologically active component into the blood stream, I think that I'd like to understand a little bit more about sort of the trend at least in what was reported as hematological dyscrasias.  There was access in the TAXUS group.  There were about ten patients with hematological dyscrasias, and there was excess of CVAs.

            Now, each one by itself doesn't lead to statistical significance, but whether or not that represents a common pathophysiological mechanism or not is something I'd like to hear from the investigators.

            DR. STONE:  Well, I think the first point is that when one looks at these SAEs which were all operator reported, there was no statistical significance between the two groups.  I mean, they were small.  They were infrequent in both groups.

            And then the second issue is:  is there an underlying pathophysiologic relationship between some that look numerically like they might be going in the wrong direction.

            I think it's important to note that for each one that looked like it might be going in the wrong direction we can show you another SAE that looked like it was going in the right direction with less SAEs with the TAXUS stent, and I think whenever one looks at so many different subgroups, first of all, you have to really use a Bonferroni correction to get a significant P value, which would probably bring you down with the hundreds of SAEs that have been looked at down to about a .001 P value or less.

            Regardless, there's no real pathophysiologic relationship that I know of between stent implantation and stroke.  I don't think we think that thrombi and emboli are forming on the stent and then embolizing to the carotid circulation, and again, I have to think from just for no other reason than not being able to at all possibly explain Paclitaxel relationship to stroke, especially given the fact that there is no systemic detectable levels, that that's just, you know, one of those things, small, infrequent events, not statistically significant.

            Now, hematologic dyscrasias, of course, we could be more worried about from any, quote, oncologic agent.  The P value is not statistically significantly different.  There's no systemic levels.  We don't think there should be any systemic accumulation in the bone marrow that might affect the hematologic parameters, and again, it was not a statistically significant P value; just one of those things that happened, which looking at multiple subgroups, you do expect to see some.

            There are other things that Paclitaxel could theoretically do, you know, like rash and whatnot that we saw, if anything, being less in the Paclitaxel group.

            DR. WEINBERGER:  I would just as a comment add that we shouldn't glibly expect that we're just looking at Paclitaxel effect.  It could also be the polymer itself, and we don't have a huge experience with this intravascular polymer, I imagine.

            And then the second also fairly brief question --

            DR. RUSSELL:  Maybe I could further elaborate on Dr. Stone's comments here.  We actually looked at the individual reports for those ten events.  Five or six of them were, in fact, formal CVAs.  The remainder were TIAs that resolved.  Of the five, two of them were related.  One of them occurred in a patient who against the protocol had a history of recurrent CVAs, and the second one was in a patient that actually had a stroke post knee replacement surgery.

            So that's fairly antidotal, two out of the five, but that is the breakdown, and we did look at that closely.

            In terms of the hematologic dyscrasias, a scary word, I would have to say, that reflects how the coating was done for the serious adverse events in our trial.  We used the standard Med DRA coating dictionary, and when we examined the verbatim terms that were classified as hematologic dyscrasia, the majority of them were bruising and rashes.

            DR. SOMBERG:  What was that word?  I didn't hear.

            DR. RUSSELL:  Bruising, bruising and rashes.

            And we have the, you know, superimposed complexity, let's say, of the plavix.

            DR. WEINBERGER:  And the angiographic subset, was that pulled out at random?  In other words, were people randomly chosen to be angiographic subset candidates, or was that some operators volunteered to be the angiographic subset operators?

            DR. STONE:  There was a consecutive series of patients at all sites.

            DR. WEINBERGER:  In all sites.  Okay.  So I would imagine that since we're talking about when we define TLR in the study, that we're talking about clinically driven TLR.  If you looked at TLR rates in the nonangiographic patients, they should be the same as in the angiographic patient.

            DR. STONE:  Well, we always expect, of course, to see somewhat of an oculus stenotic reflex, and that is that patients who come back for follow-up angiography sometimes will be feeling perfectly well, but you'll see a moderate stenosis, and you're right there and the operator feels it's in the best interest to dilate that patient.  That's been well described to us since the Benastent II (phonetic) trial.

            We carefully adjudicated events so that you had to have either confirmed ischemia or a least quantitative coronary angiographic diameter stenosis greater than 70 percent, which correlates to an operator seeing I think it's 85 or 90 percent, to adjudicate it as a TLR.

            When we looked at the TLR rates in the angiographic subset, it was 3.8 percent.  In the non-angiographic subset, it was 2.8 percent.  So we had a very minimal ocular stenotic reflex in this trial.  So I think the results are generalizable.

            CHAIRMAN LASKEY:  We have time for one more.  Dr. Somberg, did you have one final before the break?

            DR. SOMBERG:  I'd like to bring back the polymer chemist, if I could, and just get a little bit more information on the material and the studies that have been done on the material's integrity over time and what that period might be.

            MR. BARRY:  Hi.  I'm Jim Barry.  I'm the Vice President of the corporate R&D group at Boston Scientific.

            So with a low drug to polymer ratio like the TAXUS technology, we would very much expect that we could see some drug retention.  We've been very sensitive to that.  So we've done some very extensive testing.  In vitro we've done accelerated fatigue out to ten years, and we see no difference between control and TAXUS stents. 

            We've done tensile testing.  We've seen on change in mechanical properties.  We've looked in vivo both at six months and at two years for polymer degradation, and we saw none between those that were implanted in the polymer that was not.

            So I think we've done pretty extensive studies there as well.

            DR. SOMBERG:  What about in vitro degradation?

            MR. BERRY:  We have not done in vitro degradation.  Again, the polymer is very, very biostable and couldn't really degrade there.  So that's why we did it in vivo as well.  Much more realistic.

            CHAIRMAN LASKEY:  Thank you. 

            I have 11:30.  We should take a 15 minute break and be back at 11:45, and then we'll have the FDA presentation.

            Thank you again.

            (Whereupon, the foregoing matter went off the record at 11:27 a.m. and went back on the record at 11:47 a.m.)

            CHAIRMAN LASKEY:  As much as I'm warmed by the collegiality and conviviality, we have the rest of the day to get through.  If we can resume with the FDA presentation, please.

            MS. GOODE:  Good morning, ladies and gentlemen of the panel.  I want to thank you for agreeing to participate with this panel deliberation on the TAXUS PMA application.

            My name is Jennifer Goode,  and I'm a biomedical engineer in the Division of Cardiovascular Devices at the Food and Drug Administration.  I'm the lead reviewer for FDA, for Boston Scientific's premarket approval application for the TAXUS Express2 Paclitaxel eluting coronary stent system, PMA No. P030025.

            As panel participants, you're being asked to discuss and make recommendations on the applicant's PMA submission.  Your discussion on the clinical study results and labeling recommendations will be taken into consideration by FDA in the evaluation of the application.

            Finally, you will be asked to vote on the approvability of the application.

            The FDA presentation will identify for you the FDA review team members and their expertise, briefly describe the product, summarize the nonclinical testing that was conducted, summarize the clinical and statistical evaluation of the TAXUS product, and identify the FDA questions for the panel.

            The expertise required to review this product was quite extensive, and this slide identifies for you the names of the individuals in the Center for Devices and Radiologic Health that provided reviews for us. 

            I'd like to thank all of these individuals for their work on this project, and in addition, I'd like to identify for you the three members of the review team who will be speaking this morning after I do.

            First, Dr. Stephen Hilbert, the lead animal study reviewer from the Office of Device Evaluation, will describe the issues identified in the FDA review of the animal data.

            Then Dr. John Stuhlmuller, a medical officer from the Office of Device and Evaluation, will describe the issues identified in the FDA clinical review.

            And finally, Dr. Heng Li, a mathematical statistician from the Office of Surveillance and Biometrics, will describe for you the issues identified in the FDA statistical review of the clinical data.

            I also want to thank our colleagues from the Center for Drug Evaluation and Research for their immensely valuable assistance on this project, and their names and expertise are identified for you here.

            This PMA application was submitted to the agency on June 19th, 2003, and was assigned a PMA file number P030025. 

            The agency issued the applicant a major deficiency letter on September 15th, 2003.  This letter requested additional information from the applicant that was deemed necessary to complete the review of the application.

            The applicant submitted a response to the agency's letter on September 30th, 2003, to address the specific animal study, clinical and statistical issues that would be the subject of the panel review today.

            By November 5th, the sponsor had responded to all of these deficiencies in FDA's letter of September 15th, and this additional information is currently under review by FDA.

            The agency and the applicant are working interactively to resolve these outstanding nonclinical and manufacturing issues.

            As defined by Title 21 of the Code of Federal Regulations, Part 3, the TAXUS expressed two Paclitaxel eluting coronary stent system is a combination product because it is comprised of two regulated components, a device and a drug.  The device component consists of the Express balloon expandable 316L stainless steel coronary stent, and the delivery catheter.

            The bare metal Express stent is currently marketed for improving coronary luminal diameter in patients with symptomatic ischemic disease associated with stenotic lesions in native coronary arteries less than or equal to 18 millimeters in length with reference to vessel diameters from 3.0 to 5.0 millimeters, and this stent was approved through P020009.

            It's important to note that the label for the bare metal Express stent does not include a claim for reducing restenosis, as does the proposed TAXUS stent.  The Express stent is approved on both of the proposed delivery systems, the  Express2 over the wire and the Express2 monorail delivery systems.

            While the proposed Express2 delivery systems were not used in the TAXUS IV trial, the differences are minor and are not expected to affect clinical performance.

            To make the TAXUS product, the Express stent is coated with a non-erodible drug polymer coating.  The drug substance used in this product is Paclitaxel.  Under the trade name Taxol, a Bristol-Myers Squibb product, Paclitaxel in injectable form was approved by FDA in December 1992 for the treatment of cancer. 

            For this stent, a generic form of Paclitaxel is being provided by a company called Indena S.p.A., and drug safety information has been provided by this manufacturer in an FDA drug master file.

            Paclitaxel has not, however, been approved for the treat treatment of restenosis or for use in coronary arteries. 

            In developing safety information for this product, the sponsor studied various formulations consisting of different amounts of polymer and drug that have a range of release characteristics.  TAXUS SR is the product being proposed for market and is a slow release formulation with a dose per unit area of one microgram per millimeter squared.

            This product was studied in the animal trials as well as in the TAXUS I, TAXUS II, and TAXUS IV clinical trials.  TAXUS MR also has a dose per unit area of one microgram per millimeter squared.  However, the drug releases more quickly in this formulation, providing some information on the safety from an overdoes perspective.  This product was used in the animal trials, as well as in the TAXUS II clinical trial.

            For this PMA application, the applicant is requesting approval for stents ranging in length from eight to 32 millimeters and in diameters from 2.5 to 3.5 millimeters as outlined in this table.

            For the 2.5 to 3.5 millimeter diameter stents, the stent is crimped on various size deliver catheter balloons.  Because the identical stent component is used for the entire 2.5 to 3.5 millimeter diameter range, the total drug and polymer per stent are a function of the stent length irrespective of stent diameter.

            Please also note that the proposed indications for use include vessel diameters to a maximum of 3.75 millimeters.  The agency is evaluating the acceptability of expansion of the stent to 3.75 millimeters from results of bench testing.  In addition, use of the product in this fashion was evaluated in the clinical studies and no safety concerns were noted.

            Based on the dose density of one microgram per millimeter squared of metal surface area, the total nominal dosage of Paclitaxel ranges from a minimum of 50 micrograms to a maximum of 209 micrograms for the currently proposed matrix of stent sizes as shown on this slide.

            The total nominal dosage of polymer content ranges from a minimum of 519 micrograms to a maximum of 2,170 micrograms for the currently proposed matrix of stent sizes and is shown in orange.

            As denoted in yellow on this slide, please note that the TAXUS IV study was conducted using the 2.5, 3.0, and 3.5 millimeter stent diameters in lengths of 16, 24, and 32 millimeters.  Although approval has been requested for a wider range of stent lengths, the sizes evaluated in the TAXUS IV study allow for adequate assessment of safety and effectiveness because the amounts of drug substance in polymer are the same or less than the stents implanted in the clinical study.

            The sponsor conducted pharmacology studies and in vivo release studies to assess the elution kinetics and toxicity of the TAXUS product.  In vivo animal testing has been conducted to assess the acute and long-term safety of the proposed stent for the clinically intended dosage and with an overdosage of drug and polymer.

            The detailed arterial histopathology and histomorphometry provide important information that cannot be obtained through human clinical trial experience, and Dr. Hilbert will outline FDA's assessment of this information for you.

            Biocompatibility testing in accordance with ISO 10993 was conducted on polymer coated stents, or coupons, without the inclusion of the drug substance.  Since the applicant did not conduct biocompatibility testing on the finished product with drug substance according to ISO 10993, chronic in vivo animal testing has been used to evaluate the biocompatibility of the finished product.

            Bench testing has been performed to evaluate the mechanical integrity and function of the TAXUS product.  FDA is working actively with the sponsor to finalize the stability protocols and specifications prior to assigning an expiration date for the product.

            The sponsor has indicated that they have set aside adequate commercial product per ICH guidelines for conduction of stability testing.  This product will be stability tested once protocols and specifications are finalized with the agency.

            The applicant has performed drug content, elution, degradation impurity, residual solvent, and particulate testing to assess the coating integrity of the finished TAXUS products.  Sterilization validation and package integrity testing have also been conducted.

            With the exception of information on the in vivo animal testing, FDA has not yet completed the review of the recent responses to questions regarding the nonclinical testing.  However, no data have been presented to indicate a safety concern in the clinical setting regarding mechanical device failure or malfunction.

            There are two major outstanding nonclinical issues, one of which is important for your consideration today.  First, Dr. Hilbert will outline for you results from the animal studies, and FDA would like your input on whether the clinical data adequately addresses any concerns that might be raised by these animal study findings.

            Second, as noted on my last slide, FDA is working interactively with the sponsor to finalize protocols and quality control specifications for product stability testing and adequate product has been set aside to conduct this testing.

            As you heard earlier from Boston Scientific, the applicant has proposed the following indications for use for the TAXUS Express2 Paclitaxel-eluting coronary stent system for improving luminal diameter and reducing restenosis for the treatment of de novo lesions less than or equal to 28 millimeters in length in native coronary arteries from 2.5 to 3.75 millimeters in diameter.

            At this time I'd like to introduce Dr. Stephen Hilbert.

            DR. HILBERT:  I'm Steve Hilbert, and I'm an experimental pathologist and served as the pathology reviewer for the preclinical in vivo studies that were provided in this submission.

            This first slide provides the animal justification for the use of normal swine.  Comparable vascular and cardiac anatomy were present.  Vessel diameter suitable for the delivery and the deployment of the clinical version of the device, and this model certainly allows for comparisons of findings to, you know, historical data.

            I focused my review in the following areas.  Looking at handling characteristics, meaning, you know, delivery and deployment safety, and then within the safety areas, I focused on tissue or -- sorry -- device tissue responses, you know, addressing general healing characteristics, device related pathology, again focusing on device specific findings.

            The experimental design of this series of studies consisted of deployment in all three coronary arteries, dual antiplatelet treatments using ASA and clopidogrel.  These treatments started 72 hours before the implantation and were continued through the duration of the study. 

            The target artery to the balloon ratio is one to one to one to 1.2.  The duration of implant ranged from 28 days to 90, 180, and 360 days.

            The very stent platforms evaluated in this series of studies consisted of a bare metal stent, polymer coated metal stent, polymer drug coated stent, and the drug dose density was one microgram per millimeter squared, using two formulations, the moderate release and the slow release formulation.

            The sponsor is accommodated or is commended for the extensive data set that was provided, as well as the high quality of the studies that were conducted by both the animal study investigators and the study pathologists.

            The following graded pathology endpoints were used:  neural thrombus, endothelialization, neointima, luminal area or -- sorry -- luminal narrowing, a para-strut amorphous material which will be referred to as PAM, medial alterations, IEL or internal elastical lamina disruption, EEL, external elastic lamina disruption, and inflammatory response.

            The studies also included morphometric measurements, luminal area, IEL area, EEL area, neointimal area, luminal equivalent diameter, intimal thickness, medial area, stent profile, and percent stenosis.

            You know, I've estimated that greater than 400 stents have been implanted in over 200 swine and the various safety studies that are directly applicable to this marketing application, which resulted in, you know, well over 1,000 sections being evaluated in this series of safety studies.

            The sponsor has presented this morning micrographs illustrating the findings of these studies.  I will briefly summarize my observations by grouping the findings into similar and different categories, comparing the various stent platforms.

            And the next series of slides really go through some of the criteria of the different studies that were used for the safety studies.  There were studies focusing on the bare metal stent, on the polymer coated metal stent, again the polymer drug coated stent and metal stent, a safety margin study that included the moderate release formulation of one microgram per millimeter squared, and then there was an overlapping stent study using both the moderate release as well as the slow release product at one microgram per millimeter squared.

            And, you know, there were similarities as I mentioned across all of these stent platforms, and the similarities were in the areas of neointimal formation, endothelialization, no neural thrombus or thrombosis, and inflammatory response.

            And continuing on with the similarities, medial remodeling is demonstrated by ingrowth of fibrous tissue and morphometric measurements which demonstrated a stable EEL and luminal area, again reflecting medial remodeling.

            And differences across the stent platforms, there were some qualitative differences that were observed focusing on medial smooth muscle cell loss and PAM.

            And there were also some differences across the platforms that were noticed in trends that were associated with the release formulation, with the moderate release, the slow release as compared to the polymer and the bare metal stent.

            And then we also looked at the segmental analysis, looking at overlapping versus non-overlapping stents.

            So the handling characteristics which I mentioned were looking to deployment and, you know, delivery of the device so that we're satisfactory.  The tissue responses were satisfactory from the standpoint of neointimal formation, endothelialization, medial modeling and inflammatory response.

            There were some device related pathology findings which really require, you know, additional discussion at this time.  You know, medial smooth muscle cell loss was related to the exposure to the arterial wall to Paclitaxel.  Medial smooth muscle cell loss was most apparent in the 28-day explants.  By 90 days, remodeling by fibrous tissue, in growth was apparent in the media, and that continued through the study time periods and was complete at 360 days.

            PAM was resolving with time through 360 days, and microscopic dystrophic calcification is variably present, occurring more frequently in the polymer drug coated stents.  Microscopic regions of dystrophic calcification and polymer drug coated stents were reducing in size through the 360-day period.  The long-term significance of the presence of PAM in microscopic regions of dystrophic calcification are certainly unknown.

            The vessel wall remains structurally intact as demonstrated by histologic findings, and EEO and luminal area.  As within any animal model, the extension of these clinical findings to in vivo findings in the general patient population really should not be made for the following reasons:

            Accelerator rates of healing are seen in various animal models, typically used in the evaluation of cardiovascular devices.

            Neointimal formation, endothelialization, and medial remodeling findings  in nonclinical in vivo studies may not reflect similar rates of healing in humans.

            Lastly, variations in device tissue responses and device related pathology between individuals are always a concern, although a large number of animals and stents have been examined in the series of studies, the sample size is small in comparison to the expected clinical use of the coronary artery  stent.

            For these reasons, FDA would like the panel to consider the following question:  does the combination of nine-month clinical data from the TAXUS IV slow release study and the supporting data from the TAXUS I and the TAXUS II, which was slow release and moderate release, adequately address the issues raised by the animal data.

            Thank you.

            DR. STUHLMULLER:  Good morning.  My name is John Stuhlmuller, and I'm a medical officer in the Interventional Cardiology  Devices Branch in the Division of Cardiovascular Devices.  I'm going to provide a brief overview of the clinical information contained in the PMA.

            In their PMA application, the sponsors provided data for the following clinical studies that are applicable to the evaluation of safety and effectiveness of the TAXUS stent for the proposed intended use:  TAXUS I, TAXUS II, TAXUS IV, and TAXUS V.

            Data from the TAXUS IV study is considered the pivotal data set.  Data from the TAXUS I, TAXUS II, and TAXUS V studies are considered supporting data sets.

            In addition, I will outline the sponsor's proposal for the peri-approval study.

            At this time I would like to highlight the findings of the TAXUS IV study.  The TAXUS IV study is a domestic, prospective, multi-center, and blinded one-to-one randomization of the TAXUS drug eluting stent to a bare metal control stent, the Express stent, in patients with de novo coronary artery lesions.

            Randomization was stratified by clinical site, reference vessel diameter, and the presence or absence of medically treated diabetes.  The intent of the stratified randomization was to  insure that an adequate number of patients were enrolled in the various cohorts to allow assessment of whether a treatment effect was noted that would affect the poolability of the data.

            The study was not adequately powered to reach any specific conclusions regarding safety and effectiveness or to support a specific marketing claim into subgroups such as medically treated diabetes.

            At the time of randomization, a designation was also made as to whether the patient would complete angiographic and/or IVUS follow-up.  A six-month combination post procedure antiplatelet regimen of aspirin and clopidogrel or ticlopidine was utilized.

            The table on this slide interpolates the stent matrix and lesion characteristics.  Three stent diameters and three stent lengths were used in the TAXUS IV study.  Reference vessel diameters of three millimeter or less were considered small, and long lesions were characterized by the use of the 32 millimeter stent.  The bold lines delineate the characterization of lesion length and vessel diameter.

            Multiple stents were used for bailout stenting only.  Where necessary, a 16 millimeter stent was used for provisional overlapping.  Forty-four patients in this study received multiple taxus stents.

            Target vessel revascularization, TVR, at nine months is the principal effectiveness outcome measure.  MACE, stent thrombosis and vessel wall structure are the principal safety outcome measures.

            The TVR rate is 4.7 in the TAXUS arm and 12.0 in the control arm at nine months.  The absolute reduction in TVR is 7.3 percent and the relative reduction is 61 percent.  The difference in TVR is due to reduction in the number of target lesion revascularizations in the TAXUS arm versus the control arm.

            Differences in the rates of death or myocardial infarction in the TAXUS arm versus the control arm are reflected in the secondary composite clinical endpoint of target vessel failure.

            The MACE rate is 8.5 percent in the TAXUS arm and 15.0 percent in the control arm at nine months.  The absolute reduction in MACE is 6.5 percent and the relative reduction is 43 percent.

            The stent thrombosis rate at 30 days is 0.3 percent in the TAXUS arm and 0.6 in the control arm.  There were no cases of in-hospital stent thrombosis in the TAXUS arm. 

            The stent thrombosis rate at 270 days is 0.6 percent in the TAXUS arm and 0.8 percent in the control arm.

            The incomplete stent apposition rate is 11.6 in the TAXUS arm and 6.4 percent in the control arm post procedure.  The incomplete stent apposition rate is four percent in the taxus arm and three percent in the control arm at 270 days.

            Incomplete stent apposition resolved in the majority of cases from post procedure to 270 days.  The late acquired incomplete stent apposition rate is 1.1 percent in the TAXUS arm and 2.2 percent in the control arm.  Otherwise the vessel wall remains structurally intact.

            At this time I would like to highlight the findings of the TAXUS I study.  The TAXUS I study is a prospective, multi-center and blinded study completed at three centers in Germany using a near stent platform.  The primary study endpoints are MACE at 30 days and angiographic restenosis at six months.  Follow-up through two years has been completed.  This includes clinical angiographic and IVUS evaluation in those patients in the TAXUS arm who agreed to complete this evaluation.

            In terms of effectiveness, a persistent drug effect is maintained through two years by angiography and IVUS measurements in those patients in the TAXUS arm who completed the angiographic and IVUS follow-up.

            In terms of safety, the angiographic and IVUS measurements demonstrate that the wall remains structurally intact.  No safety issues are noted.

            At this time I would like to highlight the findings in the TAXUS II study.  The TAXUS II study consisted of two separate studies, meaning the TAXUS II SR cohort and the TAXUS II MR cohort.

            The TAXUS II SR cohort evaluated the slow release formulation on the near stent platform.  The TAXUS II MR cohort evaluated the moderate release formulation on the near stent platform.

            Both cohorts are prospective, randomized, multi-center, blinded studies completed outside the United States.  The TAXUS II SR cohort enrolled 267 patients at 28 centers in 12 countries.  The TAXUS II MR cohort enrolled 270 patients at 28 centers in 12 countries.

            The primary study endpoints were MACE and angiographic restenosis at six months.  Follow-up through one  year has been completed.

            In terms of effectiveness for the TAXUS II SR cohort, clinical benefit is assessed by clinical event rates as maintained at one year.  In terms of safety, the vessel wall remains structurally intact in the TAXUS arm compared to the control arm by angiography and IVUS at six months.  No safety issues are noted.

            In terms of effectiveness for the TAXUS II MR cohort, clinical benefit is assessed by clinical event rates as maintained at one year. 

            In terms of safety, the vessel wall remains structurally intact in the TAXUS arm compared to the control arm by angiography in IVUS at six months.  No safety issues are noted.

            In terms of comparing the results of the TAXUS II SR cohort and the TAXUS II MR cohort, similar clinical benefit is seen at one year.  In terms of safety, similar vessel wall measurements by angiography and IVUS are seen at six months for both groups.  thus, the TAXUS II MR cohort provides preliminary clinical evidence of the safety margin for the SR formulation.  No safety issues are noted.

            At this time I would like to highlight the preliminary findings of the TAXUS V study.  The TAXUS V study is a domestic, prospective, multi-center  and blinded, one-to-one randomized comparison to the TAXUS drug eluting stent to a bare metal, uncoated Express stent in patients with de novo coronary artery lesions.  A broader range of lesions with smaller diameters and longer lesion lengths are being evaluated compared to those evaluated in the TAXUS IV study.

            In addition, overlapping stents are being used as the primary treatment versus being limited to provisional bailout stenting.  Patient enrollment in several prespecified subgroups is ongoing.

            The study remains blinded at this time.  Thirty-day follow-up on 1,080 of 1,103 patients enrolled through November 12th, 2003, has been provided by the sponsor in its safety update to FDA.  A total of 122 MACE events have been recorded.

            Ten episodes of stent thrombosis in eight patients have been recorded.  Stent thrombosis has been characterized as acute in three episodes, subacute in six episodes and late in one episode.

            No safety issues are noted with the broader range of stent use at this time.

            At this time I would like to highlight the proposed peri-approval study.  The proposed enrollment would be 2,000 patients.  The study would be conducted in two phases.  Phase I would enroll up to 500 patients under continued access protocol utilizing the IDE expanded access provisions.  This will allow assessment of the format and function of a Web based registry.

            Phase 2 would enroll up to 1,500 patients as a post approval Web based registry.  The intent of the two-phase design would allow a smooth and efficient transition between pre and post approval.

            The total sample size is appropriate for detecting low rates of adverse events.  For example, with the proposed sample size of 2,000, an adverse event that occurs at a rate of one percent can be detected with an upper 95 percent limit of 1.4 percent.

            At this time I would like to summarize the results of the TAXUS studies.  The TAXUS IV study is considered the pivotal study by the sponsor.  In terms of effectiveness, clinical benefit is demonstrated at nine months.  In terms of safety, the vessel wall is structurally intact at nine months.  No safety issues are noted.

            The TAXUS I study is a supporting study.  Clinical benefit is maintained through two years.  The vessel wall remains structurally intact at two years.  No safety issues are noted.

            The TAXUS II study is also a supporting study.  Similar clinical benefit is maintained through one year for both cohorts.  Similar vessel wall measurements are seen in both groups at six months.  No safety issues are noted.

            The TAXUS V study is a supporting study at this time that is evaluating an extended lesion range in that treated in TAXUS IV study.  The study provides preliminary evidence of safety for a broader range of use in greater than 1,000 patients.

            The proposed peri-approval study will enroll 2,000 patients in two phases pre and post approval.  The proposal is satisfactory from FDA's perspective.

            At this time I would like to introduce Heng Li from the Department of Biostatistics.

            DR. LI:  My name is Heng Li.  I'm the FDA mathematical statistician for this PMA submission.

            In the next three slides I will provide a few comments, some main features of the pivotal clinical trial design conduct and analysis.

            The pivotal clinical trial, the TAXUS IV trial is a multi-center, randomized, controlled, double blind study for safety and effectiveness.  Pretreatment variables are very well balanced between the two randomized arms, indicating that the randomization has been successful.

            Twelve patients were de-registered from the trial after being randomized.  Those 12 patients were not included in sponsor's analysis.

            It's not a good statistical practice to routinely exclude deregistered patients from analysis without considering the reasons for de-registration.

            In this specific case, since the study is double blind, the number of de-registered patients is small and evenly distributed across the two randomized arms.  The analysis excluding de-registered patients is considered by FDA as acceptable.

            This is not the first instance in which an analysis excluding de-registered patients was considered to be acceptable by FDA.

            Among the patients who are not de-registered, those are patients who were included in the analysis.  Reportedly there were quantified deviations.  For example, reportedly there are deviations from treatment protocol, and apparently not every patient has been followed all the way to nine months.

            Those protocol deviations are not very extensive.  Differing approaches to handling those protocol deviations in data analysis do not change the statistical significance of the statistical test preplanned for the primary endpoint.

            The sponsor performed subgroup analysis for this PMA.  Here a subgroup refers to subset of patients, such as males, females, those with diabetes, those without diabetes.

            The subgroup analyses, as reported, have not identified safety issues.  Generally speaking, statistical inference, foremost statistical inference associated with subgroup analyses usually is complicated by the need to account for issues such as multiplicity and is limited by insufficient power due to small sample sizes.

            Here multiplicity is a general term referring to the phenomena such as inflation of Type 1 error due to too many statistical testing.

            Such issues, such as multiplicity, issues such as multiplicity is usually intractable for unplanned analysis.  Therefore, unplanned subgroup analysis can perhaps never be considered as reliable from a statistical perspective.

            This concludes the statistical summary.

            CHAIRMAN LASKEY:  Thank you, people.

            Does the panel have any questions for any of the FDA presenters?

            Dr. Hirshfeld, do you?

            DR. HIRSHFELD:  Yes.  This is a question for Jennifer Goode and maybe BSC and can also answer this.

            I note that the design of this device is that it actually provides a twofold variation in delivered drug dose if you normalize dose per unit vessel area because you're delivering the same dose of drug to a two and a half millimeter vessel as you are to a three and a half millimeter vessel so that the actual dose delivered if you normalize it to vessel area varies over a factor of two.

            And I wondered whether this issue had been addressed in the original design planning and whether, you know, we could see any effect of this in the larger vessel sizes.

            MS. BOAM:  I'm Ashley Boam.  I'm the Branch Chief for this product.  I'm sure Boston  Scientific can speak to this in more detail.

            In terms of some of the in vivo studies that were done and animal studies, a wide range of stent sizes were evaluated so that this type of effect could be looked for.

            DR. STONE:  You know, the Paclitaxel is a very lipophilic agent, and we think there's a different way of looking at this issue.  Rather than the dose actually being double, what happens when you put the same stent size in the smaller vessel is that the metal surface area coverage of the vessel is actually greater.  The stent struts are closer together.

            So I think there's actually greater drug distribution over any given vessel area, and actually when you look at the relative efficacy of the device going from larger vessels to smaller vessels, we actually saw less late loss in the smaller vessels treated with the Paclitaxel stent compared to the control stent.

            We've never seen that before with the bare metal stent.  The amount of late loss is always pretty independent of vessel size.  With the Paclitaxel stent it tended to be smaller, and of course, this is just a hypothesis at this point, but it may be because the stent struts were relatively closer together.  We saw no safety concerns in any of the subsets.

            DR. HIRSHFELD:  It might be for future development that it may be that the dose at the two and a half millimeter level is actually better than the dose at the 3.5 millimeter level.

            DR. STONE:  That is possible.

            CHAIRMAN LASKEY:  Chris.

            DR. WHITE:  I'd like to ask Ms. Goode in her submission or the questions on page 7 for the FDA, summary of FDA review information, she makes a comment that the sufficient information to support the claim of cytostatic versus cytotoxic effect has not been submitted to FDA for review.

            Does she think that's an important issue?

            MS. GOODE:  The labeling that Boston Scientific provided for your review included that claim.  We've talked to Boston Scientific about the fact that we don't believe they provided data to us in the PMA application that supports that claim.  They can speak to the fact that they have been collecting additional data, and I believe may play to address that at a  future time.

            But I wanted to bring it to your attention in the FDA review in the panel pack because we have not yet agreed that we've seen sufficient data to support that claim.

            DR. RUSSELL:  We removed that claim from our DFU until we gather substantially more data to support it.

            MS. GOODE:  I guess I should just make one clarification.  We had talked to Boston Scientific about this, and it was in two places in the label, and they removed it from one place, and it may have been an oversight not to remove it from the second place, and that's why in the FDA review memo I brought it to your attention.  So that may have been an oversight.

            CHAIRMAN LASKEY:  Only if it's a statistical question.  Sharon.

            DR. NORMAND:  I can't ask a clinical question?


            DR. NORMAND:  It will be a statistical question.

            I just wanted to ask the FDA reviewer.  You said there were protocol deviations, and I can sort of guess what some are, but I was wondering if you could just broadly say what those protocol deviations were.

            DR. LI:  Oh, the protocol deviations are inferred from the fact that a protocol analysis was provided by the sponsor, as well as analysis that does not address protocol deviation.

            So from that fact I know that there are protocol deviations.

            DR. NORMAND:  Okay.  So it wasn't like anything.  It was just really the fact of the analytic procedure that they adopted.  That's the way you're characterizing protocol deviations.

            DR. LI:  I believe that a protocol deviation is defined in this PMA as target lesion without not being treated with study stents, but I would like to confirm this with the sponsor.

            DR. NORMAND:  Okay.  Thank you.

            DR. STUHLMULLER:  I'd like to add one clarification.  I mean, from a clinical perspective, part of that comes down to, you know, if you have a two week or 30-day window to complete follow-up in and somebody comes in at day 31 versus day 30,  well, that counts, you know, in terms of bean counting as a protocol deviation, but from a practical perspective it doesn't affect the interpretation of the data, or if somebody is supposed to take a medication for six months and they take it for four months for some reason, that could potentially  also be interpreted as a protocol deviation.  And that's part of what we were referring to.

            Dr. Russell might want to address that as well.

            DR. RUSSELL:  We did extensive monitoring in this study.  The investigators were cited if there were any deviations whatsoever from any of the follow-up windows, including testing such as cholesterol.

            Investigators were cited for having protocol deviations if they did not inflate the balloon to the specified amount listed in the protocol.  So in our analysis of the protocol deviations, there were few of them that had any effect on the outcomes.  We provided FDA with analyses of per protocol.

            All randomized patients, including the 12 de-registered patients, patients that received only single stents, patients that received multiple stents.  We provided FDA with extents of analyses and the findings were replicated irregardless of the analysis.

            With respect to the 12 patients that were deregistered, those patients were deregistered because there was no attempt ever to even place a stent.  Patients that had an attempt and failed were included in our analysis.

            I hope that addresses the questions.

            CHAIRMAN LASKEY:  Dr. Maisel.

            DR. MAISEL:  This is a question for Ms. Goode.

            On page 7 of the FDA review, it mentions that the sponsor had been asked to provide additional information regarding the carcinogenic potential of the entire stent, including the polymer, manufacturer additives, et cetera.  Was that additional information provided?  And what was the result?

            MS. GOODE:  That information has been provided to the FDA.  The review is not yet complete on that.  So we don't expect that there would have been a problem, but I haven't conferred with my consultants.

            CHAIRMAN LASKEY:  One question to Dr. Stuhlmuller and perhaps the statistician.  This certainly was an interesting study from the standpoint of it being a clinical endpoint rather than an angiographic endpoint for the restenosis.  That being the case and clinical endpoints being driven by the patient population, if I read the pack right, about 50 percent of these patients were asymptomatic at the time of the procedure, and I wonder how that might affect rates and certainly differences between the two arms if that were to change in some more real world fashion, for example.

            DR. STUHLMULLER:  Just in terms of clarification, let me make sure we all understanding what we're all talking about.  Can you identify the table specifically?

            CHAIRMAN LASKEY:  Yeah.  The clinical study report, report Table 7, Baseline Demos and Clinical Characteristics.  This is page 68 in Volume 2.  If you just look down to the -- it's actually the top angina class, none.  It's 48.9 percent versus 48 percent.

            I just struck me as being quite unusual for, first of all, an interventional trial, but secondly, a trial driven by a clinical endpoint to have such a rich population of patients without angina and then to use angina as the driver for the endpoint.

            DR. STUHLMULLER:  Yeah, my interpretation of that table is if you add up the unstable and the silent ischemia is that comes out to be 48.9, and it took me a while.  You know, it took several times for me when I went through that table to understand that, and I think perhaps it would be reasonable to ask Dr. Russell to clarify that, as to whether my interpretation is correct.

            CHAIRMAN LASKEY:  I'm misreading that.  That's supposed to add up to 100 percent?

            DR. STUHLMULLER:  Yes.

            CHAIRMAN LASKEY:  Very confusing.

            DR. RUSSELL:  I believe that Dr. Stuhlmuller has addressed the question.  Those were the patients with stable angina only and the breakdown of stable angina.  So the patients that had none fell into either silent ischemia or unstable angina.

            CHAIRMAN LASKEY:  But they were in a minority.

            DR. RUSSELL:  Yes.

            CHAIRMAN LASKEY:  Okay.  Thank you.

            DR. RUSSELL:  The inclusion criteria mandated that they have either stable angina, unstable angina, or silent ischemia.

            CHAIRMAN LASKEY:  Any other questions for the -- oh, Dr. White.

            DR. WHITE:  I'm sorry.  I just wanted to ask a short of one of Dr. Russell.

            The Express2 stent, the FDA says that the Express2 stent is a minor change.  Express2, I suppose, is the delivery system.  Why did you not use that delivery system for the trial?

            DR. RUSSELL:  We were optimizing the minor modifications at the time of the IDE submission and so --

            DR. WHITE:  So it wasn't available before the trial.  Because it makes it seem like it's an approved platform for the express stent; is that correct?

            DR. RUSSELL:  It's been subsequently approved for the Express.

            DR. WHITE:  I see.  It seemed to me like it was available and you chose not to use it as a delivery platform and now have changed your mind.  that's not true.

            DR. RUSSELL:  That's correct.  It's an evolving scenario.

            CHAIRMAN LASKEY:  Well, then we're at that critical point where we can now break for lunch.  So may I suggest that we adjourn for one hour?

            I have 12:40.  If we can call ourselves to order at 1:40, I would appreciate that.  Thank you all.

            (Whereupon, at 12:36 p.m., the meeting was recessed for lunch, to reconvene at 1:40 p.m., the same day.)












         A-F-T-E-R-N-O-O-N  S-E-S-S-I-O-N

                                       (1:49 p.m.)

            CHAIRMAN LASKEY:  If we can resume, that would be much appreciated.

            Okay.  I'd like to call us back to order and lead off the afternoon with the open committee discussion and have Drs. White and Weinberger, respectively, give their reviews and ask questions of the sponsor.

            Dr. White.

            DR. WHITE:  Thank you.

            I guess I will reference the submission panel that I'm talking from.  Do you have that available?

            I'd like to start on page 4 of the safety and effectiveness part of the document.  Page 4, where it says "contraindications," the last of the bullets say "known allergy to stainless steel."  And I'm asking a question because my reading of other products, usually that sentence says "known allergy to nickel."

            And is that what you normally say?  Is that what you say in your other stents?  Is that what it says for the Express or is the specific mention of allergy to stainless steel what you're trying to call attention to or is it more the nickel component of the stainless steel that you want to raise awareness about.

            Because my understanding is it was the nickel allergy that was of concern.

            DR. RUSSELL:  I agree with you, Dr. White.  I'm not sure if this language was FDA mandated, but I would be happy to --

            DR. WHITE:  We should just be consistent with what you've done for your other stainless steel stents, and when I've read them, it has always been a nickel allergy, not a stainless steel.

            DR. RUSSELL:  I guess ours says stainless steel.

            DR. WHITE:  Fair enough.  That's fine.  I just wanted to raise that question.

            And then on the next page, page 5, under precautions number one, general precautions, the final bullet says that "antiplatelet therapy is recommended for a period of six months post procedure."

            And I think this morning in your presentation, and I would suggest I think you acknowledged this this morning in your presentation, but I would like to see you specify that you really want people taking plavix or ticlopidine for six months with their aspirin; that you not just say "antiplatelet therapy."  I think you really do want to ask for a dye.

            DR. RUSSELL:  Absolutely.

            DR. WHITE:  All right.  On the next page, page 6, you talk boulesion and vessel characteristics, and other than the fact that on the fourth bullet there's a typo, you said "overflow" and I think you mean "outflow."  That's not important.

            But the next bullet down says, "Patients with tortuous vessels in the region of the obstruction."  What are you trying to communicate to the physician when you say "tortuosity"?

            I know that in the inclusion criteria of the trial you ask for 60 degrees.  Do you want to be more specific or less specific than the trial criteria for the IFU?

            DR. RUSSELL:  I'm looking at FDA because we complied with the language that we used in our protocol.  We could be more specific?

            DR. ZUCKERMAN:  Dr. Whit, you're referring to the summary of safety and effectiveness.

            DR. WHITE:  Right.

            DR. ZUCKERMAN:  Which if the product is approved would appear on our Web site and would be the FDA document that summarizes the data for a PMA submission.  We can certainly make it a more accurate and truthful document.

            Are you questioning the sponsor and FDA regarding the actual intended labeling, which is a little bit different?  It is in a section "instructions for use."  Maybe you want to refer to that section if you look at how the clinical trial is described.

            DR. WHITE:  I think that as a physician who -- you know, if you ever want to look at the label, I think tortuosity is -- every vessel has got some tortuosity.  So I think you want to make sure that you are giving a reasonable estimate.  If 60 degree tortuosity was what the trial used and you can expect those results and that success, then probably being consistent in the IFU for the physician reader would make sense rather than being more general.

            The next thing, I'd like to jump to the patient guide.  I have a booklet in here that has a colored label called patient guide, and on page 11 of the patient guide, the second question, it says, "Can I walk through a metal detector with a stent?"


            DR. WHITE:  And the answer is of course you can, but your reasoning I think isn't correct, and that is that you said here that you have a nonmagnetic and I believe stainless steel is ferromagnetic.  Nitenol stents (phonetic) are nonmagnetic, but isn't stainless steel magnetic?

            You're saying no.  It's not ferromagnetic?  I didn't know that.  So stainless steel is not or 516 or whatever it is you're making?  No, I didn't know that.  All right.  There you go.  You taught me something.

            DR. RUSSELL:  I believe that that is a residual hangover, Dr. White, from the labeling on all of our other stents.

            DR. WHITE:  Well, I guess I always thought that the stainless steel stuff was ferromagnetic and was an issue, and the Nitentol stents were not.  I thought that was the difference between those stents, but I guess that's not true.

            I would like to see in the patient's guide some emphasis about the need to continue to take their plavix and the risk of thrombosis.  I think the newspapers in the last couple of months have raised everyone's awareness to a probably inappropriate degree over the risk of subacute stent thrombosis, and I think the one thing we can do is instruct patients of the need to continue plavix.  Occasionally I think patients inadvertently stop their medicine without thinking that it's that important or it's expensive or whatever the reason.

            So perhaps in the patient instruction booklet there would be some emphasis about the importance of taking the medicine and if you must stop, contact your doctor, and that would be something they could take away from the encounter.

            I wanted to compliment Boston Scientific and Gregg and Steve on the conduct of this trial.  I think that we were talking about how disciplined and how nice this was done. It is nice.

            The one fly in the whole ointment is the IVUS data, where only 66 percent of the data was actually interpretable, and I mean, the way that you cite investigators for noncompliance, and I get cited every day for other Boston trials, I can't believe people were sending you IVUS data that wasn't interpretable and not getting feedback to fix it.

            I mean, how could only two thirds of the data that went to the IVUS lab end up being interpretable unless it was stacking up somewhere and not being read until the end of the trial?

            DR. RUSSELL:  I'm going to have Dr. Weissman answer that question, but it is one of the complexities of paired IVUS analysis.

            DR. WEISSMAN:  I'm Neil Weissman.  I'm a cardiologist from Washington Hospital Center.  I directed the Core Lab.

            In terms of conflicts, I'm a full-time, salaried employee of MedStar Research Institute.  Boston Scientific gave a grant to MedStar Research Institute to do this.  No stock or anything like that.

            If you look at the compliance with the IVUS, it actually was very good.  It's about 75 percent, which is pretty typical for U.S. multi-center studies.  What you're looking at is how many patients had full, 3D volumetric analysis.  In order to have that, they had to have every single millimeter throughout the stent visualized and the interface was visualized.

            DR. WHITE:  But let me just stop you because it does say 2D, two dimensional analysis.  In the next sentence it's 53 percent for 3D.

            DR. WEISSMAN:  Okay, but the cutoffs even with 2D, okay, we would go through.  If you were missing part of it, okay, and you could still do 2D, you know, on some, you wouldn't go on to 3D.  So I guess there is differential.

            I think the things to look at is the characteristics between the patients that did and did not have IVUS, and if you look at the patient characteristics, the lesion characteristic, angiographic and so forth, they're the same

            The other thing that we set up to alleviate bias is that we present IVUS sites.  So at those IVUS sites, everyone had to have IVUS consecutively.  So I agree with you.  You know, it could have been better, but I think that we still did a good job, and if you look at the IVUS cohort, they are very representative.

            DR. WHITE:  Were you given feedback to the operators that they were giving you unusable data?

            DR. WEISSMAN:  Yeah, we gave them friendly feedback forms usually within about two weeks of the time.

            DR. WHITE:  So they'd get unfriendly feedback.


            DR. WHITE:  While I've got you here, do you have access to these tables?

            DR. WEISSMAN:  I can get them, I'm sure.

            DR. WHITE:  I have a question about Table No. 34.  I'm sorry.  Table No. 35.  It's on page 93 of 145.  The title of the table is "Incomplete Apposition Determined by IVUS."

            My question is that the denominator used in the table is an n of 112, and the denominator in the rest of the IVUS tables are all in the 80s.  So how did you get 112 evaluable patients for that endpoint and not for the other endpoints?

            DR. WEISSMAN:  Yeah, again, I think when you look at the difference between qualitative and quantitative measures, I think that we're talking about quantitative measures, the low ones, but you could still have qualitative measures.

            DR. WHITE:  Okay.  Fair enough.

            I'd like to switch to Volume 2 for just a minute.  In Volume 2, I'm looking at the protocol amendments, and this is page 14 of 83.  Number 14, one of the exclusion criteria with the patient taking colchicine, is that important?  Why was that?

            DR. RUSSELL:  That was set up originally because colchicine also can affect mycrotubulus.  Obviously the number of patients receiving colchicine in modern medicine is relatively small.

            DR. WHITE:  I that going to end up in the IFU or as a contraindication?

            DR. RUSSELL:  I think it should be debated.

            DR. WHITE:  Can you tell me why you didn't include the most important lesions in the LAD that I want to treat with this stent, which are the osteolesions?  Why can I not treat an osetolesion now with a  TAXUS stent, an osteo LAD lesion?  Why did you not include osteolesions in this trial?

            DR. RUSSELL:  Well, osteolesions are routinely excluded from early phase programs, originally, I think, because they were perceived to be of higher risk.  As interventional cardiologists have gotten more technically savvy, I think that concern has gone away, and in fact, it is one of the things that we are looking at modifying in our standard inclusion criteria.

            DR. WHITE:  I was going to say that I don't think you're including osteos in TAXUS V, are you, Gregg?  And you might want to consider modifying that.

            DR. STONE:  Yeah, I mean, there's always the concern in a controlled trial about implanting a stent within the left main coronary when you're dealing with an osteo, and especially when you have a narrow, you know, angle between that.  That can lead to a complicated situation.

            So you try to do these trials to avoid those what could become outlier situations and very rare complications that could skew the data one way or the other.  So for that reason, most studies have excluded osteos, but I agree.  It's not in TAXUS V either, and it's just a reality of doing clinical trials, the same way we exclude definite angiographic thrombus and heavy calcification and other situations.  It's difficult.

            DR. WHITE:  I would say that I think now more than ever in our clinical practices we're being driven by the evidence base, and that driving away from the evidence base is risky and dangerous, and many of us are afraid to do that, and so we really do depend on the trial data to justify our practices.

            So I think it pushes us even more important to test the lesion subsets that are going to become clinically important to us.

            When you say "osteolesions," would that also then apply to a right coronary osteolesion?

            DR. STONE:  Yeah, they were excluded.

            DR. WHITE:  I mean, you could exclude the left main stuff by the bifurcational nature of the lesion, but it still would -- now I guess when I have subacute thrombosis and I put a TAXUS stent in a right coronary osteo and a plaintiff attorney says, "Dr. White, didn't you participate in the panel that said that that wasn't a good place to put that stent?" it puts us in a difficult spot.

            DR. STONE:  No, I understand, and a lot of these complex types of lesions, I mean, bifurcations we probably do more than any other type of lesion subset that weren't included here.  Ultimately we need registry studies and then follow-up carefully controlled studies to examine those lesions.

            DR. WHITE:  That's all I have.  Thank you.

            CHAIRMAN LASKEY:  Thank you.

            We'll go to Dr. Weinberger and then resume around the table.


            DR. WEINBERGER:  I think Chris covered a number of things I was interested in, but I'd like to just go back to the issue of MRI.  In the proposed instructions for use, it says, "Do not perform magnetic resonance imaging scan on patient post implantation until the stent has been completely endothelialized (30 days) to minimize the potential for migration."

            So my question and comment is, number one, if this is non-ferromagnetic, do you care about the patient having an MRI the next day?

            And if it's not -- okay, and if it is ferromagnetic, are you so assured by the animals studies that it's completely endothelialized at 30 days that you're comfortable that that's the situation in humans?

            DR. RUSSELL:  It has been my contention with my staff that this should not be included in the DFU, that there is substantial literature to justify the safety of MRIs after patients have been stented, and I have been instructed that  it is included in all DFUs and needs to be there.

            So I would be happy, delighted to follow the panel's recommendation and work with FDA on that particular issue.


            DR. WEINBERGER:  Okay.


            DR. WEINBERGER:  All right.  Question/comment No. 2.  There was a relatively small number of patients that got two stents in TAXUS IV.  Certainly that was an unintentional event, and that was not prespecified.  In any case, the instructions for use are not silent about that.  They say that use of more than two TAXUS stents has not been fully evaluated, implying that the use of two TAXUS stents has been fully evaluated.

            And I don't know whether or not that can clearly be stated from your study.

            DR. RUSSELL:  I believe that it should read the use of two stents has not been evaluated.  I mean, it has not been evaluated.  That was our intention.

            DR. WEINBERGER:  You correct that because it then goes on to say, "When multiple stents are required, resulting in stent-to-stent contact, stent materials should be of similar composition to avoid the possibility of dissimilar metal corrosion."

            So my question is:  is there any data to support that contention either, or is that just a theoretical issue?

            DR. RUSSELL:  Yeah, I had exactly the same discussion with my staff when the DFU was developed, and I believe that is a theoretical concern.

            DR. ZUCKERMAN:  Dr. Weinberger, we can have Ms. Boam clarify why we have that standard language in.


            MS. BOAM:  That's standard language that goes into IFUs for both drug eluting and bare metal stents because there are now stents of multiple materials available in the market.  Unless the company has actually done testing to evaluate the potential, for example, for galvanic corrosion because of the mix of metals.

            DR. WEINBERGER:  But I assume that's never been seen in vivo in any animal study; is that right, or it has never been tested?

            DR. RUSSELL:  We've only tested two similar stents in our studies.  We'd be happy to work with FDA on clarifying that language.

            DR. WEINBERGER:  Okay, and I would just add my overall congratulations to Boston Scientific on providing some very clear information about this particular stent device.

            CHAIRMAN LASKEY:  Is that it?

            DR. WEINBERGER:  Yeah.

            CHAIRMAN LASKEY:  Thank you.

            It actually may come up because there are some non-316L stents coming down the pike.  So dissimilar metals in an aqueous environment is not a good situation.

            Dr. Hirshfeld.

            DR. HIRSHFELD:  One question relating to the physician instructions for use.  Unless I missed it, I don't believe there's anything in there that describes the maximum diameter to which the stent is approved to be inflated or if there is, I looked for it and couldn't find it, which would mean it's probably not displayed prominently enough to enable us to find it.

            And I was wondering.  I'm sure this data has already been determined for the Express stent, and I think that should be available.

            DR. RUSSELL:  We'd be happy to include that.  It is provided with the Express stent.  So we'll just make sure that it's supplied in this DFU as well.

            DR. HIRSHFELD:  Yeah, because your language right now that's in there says to use an appropriate sized balloon, but it doesn't tell you what the upper limit of appropriate is.

            DR. RUSSELL:  We'll work with FDA to get the proper language on that.

            CHAIRMAN LASKEY:  Dr. Somberg.

            DR. SOMBERG:  Well, I, too, want to congratulate the company on a very excellent presentation and very good database that they have presented.

            With that said, I do have some concerns, one being, as I mentioned earlier, the dose, and I'm going to congratulate you and take you, whoever chose that dose, to Las Vegas next time I go --


            DR. SOMBERG:  -- because it takes a certain amount of, oh, courage to embark upon these types of large clinical studies with really minimal assurity (phonetic) that you're going to choose the right dose.  And having the MR versus the SR really is another issue, and that's the rapidity of release early on, which doesn't have to do with a cumulative dose release.

            So I think we have a polymer here that is retaining most of the material, 80 or 90 percent.  When you take it out, you might lose ten percent, maybe even higher, and you have a potential for an adversity down the road.  So I would strongly suggest that you're going to get this data earlier from animal studies.  If you can carry out the animal studies for a prolonged period of time in vivo and study the reliability of the material over time.

            So I think that's necessary, but not a reason to deny the public the benefits here.

            There are two other areas I have concern about, and that is stent-stent interactions and drug interactions.  The first is a stent-stent interaction, which is in part a stent-drug interaction, and that is that it's well known in the pharmacologic literature that sirumulous (phonetic) inhibits Paclitaxel's activity. 

            So I think if you had a situation where you didn't deliver your stent and they only have one Paclitaxel stent and then they had a sirumulous stent and they mixed them and they put them next to each other, I think you would have a problem since sirumulous may very well inhibit.  Let's say you wanted that minimal concentration effect curve, and it inhibited it.  You could really undo it so that you could have, you know, $10,000 worth of stents, and it would have zero benefit.

            So I think that should be noted in the material that there is a potential drug-drug interaction from the sirumulous to the Paclitaxel.

            Now, there's a whole complexity here because it said in the literature that the eluting substance really just has a local effect, and I do agree, and the systemic effects are not very much, but that local effect does go beyond just the stent because it occurs to both of your edges, and you do have an edge effect, as well, which takes in that whole area that you're considering.

            And I think systemic effects of drugs that interact with the Paclitaxel is something to note.  So there are some drugs that may be inhibitory and there are some drugs that may be facilitory and really don't have knowledge on that.  So I think that has to be looked into.

            The commonest drug used in this population is a torvastatin, and the torvastatin, you know, it's not a hepatic effect.  It's not through your 2C8 or the other polymorphisms, but it's a direct effect on the action of Paclitaxel.

            So I think you have to mention that, and I actually think, you know, if you get a large enough study you find better results than those who are on a torvastatin.

            Also, dicumerol is a rather frequently used compound there, and that potentiates the Paclitaxel effect, and then there is the COX-2 inhibitors, which has an action there as well.  So I think these things have to be noted for clinicians, and I did not see in your database an analysis of, you know, people on, people off.  I'm not sure the database is at this time large enough to look at that.  But these may be confounding factors and certainly should be noted.

            With those said, I want to again congratulate you on this very nice study and presentation.

            CHAIRMAN LASKEY:  Dr. Aziz, do you have it within you?

            DR. RUSSELL:  Dr. Aziz, do you mind if I ask Dr. Rowinsky to address the concerns about drug interaction?  Dr. Rowinsky is an oncologist deeply involved in the Taxol program who may be able to provide us with some insight.

            DR. ROWINSKY:  Thank you, Dr. Russell.

            Good afternoon.  I'm Eric Rowinsky.  I'm a medical oncologist and clinical pharmacologist and Director of the Institute for Drug Development in San Antonio. 

            I'm a paid consultant for Boston Science.  I do own stock.  I do not have any options.  I do perform studies as a clinical investigator with competitors and potential competitors.

            I've worked with Taxol since the initial studies in 1984 and was involved with its approval in the first oncologic indication in 1992, and have been very much involved in clinical research first with this agent since its approval.

            At this point, there really are no consistent evidence of drug-drug interactions reported with Taxol.  It is metabolized by Cyp3A4 and 2C8 very extensively and very efficiently.  It is only a weak inhibitor of those two Cyp systems, and there has been extensive in vitro work with multiple types of metabolic systems, including all of the Cyp enzymes.

            The KI for those two enzyme systems which are most relevant as far as inhibitory effects far exceed the concentrations that we even achieve in oncology patients which are several orders of magnitude higher.

            The only interactions than would be relevant in this population, the only interactions that have been reported thus far are interactions with respect to inducing agents, such as anti-convulsants on the metabolism of Taxol, which I don't really believe is relevant here.

            Thank you.

            DR. SOMBERG:  Well, I'd just like to follow up for a moment.  I'm not talking about the hepatic polymorphism systems, but for instance, a torvastatin has been reported to potentiate the effect of Paclitaxel locally. 

            So what I'm saying is if you have a systemic concentration of a torvastatin and you have a local effect of the Paclitaxel, could you not potentiate that effect, potentiate in terms of its efficacy, which is good, but also potentially potentiate it in terms of further inhibition of the phenomena that may lead to weakening of the imbedding and all of that sort of stuff.

            So that's really my concern.  Do you share in those local effect considerations?

            DR. ROWINSKY:  No, I don't think the evidence is in, but I think that there are some, for example, studies now of systemic rapamycin which is being developed as a chemotherapeutic in Taxol and in combinations, and I think that will actually give us some lead as far as the ability of these agents to augment the other in various tissues, but I believe that won't answer your question that deserves to be answered.

            CHAIRMAN LASKEY:  Salim.

            DR. AZIZ:  I've got to whisper because I've got laryngitis.

            Dr. Russell, the focal calcification that that you see in the animals, do you have any explanation for that?

            DR. RUSSELL:  The calcification, the focal calcification?  We believe and have a fair database that with implant injury you can have small areas of necrosis, focal necrosis in the regions where the stents are, so to be part of the healing response to that focal necrosis.

            DR. AZIZ:  Do we have any data from any patients that died and your autopsy studies?

            DR. RUSSELL:  We have no stents from any patients that have died that have received TAXUS.  We did receive one stent from a patient that died, and unfortunately that was a control stent.

            We did look at calcification scoring by the Core Labs before and after to see if there was any correlation, and we were not able to identify one.

            DR. AZIZ:  And just a matter of theoretical interest, do you have any animal data with a stent intravenous system?

            DR. RUSSELL:  I don't believe we do.

            DR. AZIZ:  Okay.  Thank you.

            DR. NORMAND:  I have a few technical questions for clarification.

            First of all, I wanted to ask both the procedures for randomization.  You didn't mention this.  I'm just presuming it.  I assume you randomized within center.

            DR. RUSSELL:  Yes.

            DR. NORMAND:  It didn't exactly work out that way the whole time, but it looked like, in fact, that's what you had done.

            DR. RUSSELL:  We used an interactive voice randomization system.

            DR. NORMAND:  And the other thing I wanted to get some clarification on is the procedures for blinding.  You indicated it was double blinded.  I wanted to get some sense of it wasn't in the panel packet, the description.  You did mention.  I think Dr. Stone earlier today indicated that the stents looked identical.  So presumably the interventionist doesn't know what it is.

            But could you just say a little bit more about the double blinding procedures?

            DR. RUSSELL:  Certainly.  We began by investing in a process that allowed us to serialize the stents.  So they were packaged under identical conditions.  Then they were serialized.  The serialization numbers were used by the interactive voice randomization system to conduct the randomization.

            And the reason we consider it -- I've been told I inappropriately said this -- triple blind is because the investigators are blind; the patients are blind; all of the core labs remain blind, and all of the study staff working on the study remain blind.

            DR. NORMAND:  Well, it's a beautifully designed study.  There's no doubt about that.

            I wanted to get a clarification on we've heard two different interpretations of intention to treat which I find amusing, but I guess what I wanted to get a sense of, it's 12 people, you know, but in terms of the 12 people, I'm still confused as to why your Boston, the sponsors, doesn't believe that they were part of anything and, hence, you really did use intention to treat principles.

            DR. RUSSELL:  This decision was a decision that we made prospectively in partnership with FDA.  Our study randomized patients before they had predilitation, which is fairly unusual in the interventional cardiology world.  We did that at the request of our investigators to facilitate the conduct of the procedure.

            Normally, we would have waited till there was successful dilatation, and then we would have known that the chances of implanting a stent were high, but we made this decision at the request of the investigators.

            So we had a number of patients where there was failure of the wire to actually cross the lesion and failure to dilate, which led to the need to revascularize with alternative means because of complications.

            DR. NORMAND:  Okay, but they had been randomized in theory, but never made it subsequently.  So they actually had been randomized to two different -- okay.

            DR. RUSSELL:  Yes.

            DR. NORMAND:  So I think in that technical sense you really didn't do an intention to treat analysis.  It doesn't matter, I think, because of the numbers, but I would hate to describe anything as the intention to treat when it really isn't intention to treat, but I think most statistical people would feel that, in fact, those 12 people are important.

            I have some clarifications I would like about the missing data again, and this is actually kind of important, and it relates to some of, well, almost all of the tables I saw presented today, and I still think I may be misunderstanding, but I checked the numbers again, and I think all of the percentages are based on full numbers, six, six, two and six, five, two.

            So you could go and look through your packages and you'll see that in various tables where the numbers are presented you are assuming as if you have complete follow-up on all of the individuals.

            Now, someone help tell me that I'm wrong on that because I just don't understand it, and again, I think you said 97 percent are complete, which is, you know, a pretty high number, but nevertheless, for example, if I look at page 78 of 145 -- I don't know how to describe these to you -- Volume 1, nonclinical and clinical testing summary, and I'm just pulling out one of the tables.  We could go through several more tables, but so if you look down at the numbers, all of the denominators are exactly the same.

            So six, six, two, and six -- I can't see.  I've got to hold it far away from me to see this -- but six, six, two and six, five, two, it looks like you have the same denominators all the time, and that's true for all of your descriptive tables.

            Hence, my inference is that indeed even though you did have loss to follow-up, you included them as if they were in there.  So I think when I asked earlier, the answer was that they were taken out, but could somebody help me with that?

            CHAIRMAN LASKEY:  You're looking at me?


            DR. RUSSELL:  We're looking at the tables ourselves

            DR. NORMAND:  I mean, it's fairly simple if you look on page --

            DR. RUSSELL:  Yes.

            DR. NORMAND:  It all is divided by six, six, two and six, five, two, and surely it looks like you've used that.  I can't imagine -- that's a pretty big typo.

            And it's the same thing in many of the other slides that I viewed.  There's a lot of them.  So I hate to go through all of them, but again, I don't want you to sort of have to look at every single table, but it seems to me with the numbers that are missing, I'd just like to get a sense of where, indeed, it looks like you're assuming you've got complete cases for the six, five, two and the six, six, two.

            The 30-day may be 100 percent, but my understanding is at 270 days you did have some loss to follow-up.  It's not a great big loss to follow-up, and I'm understanding you assume the censoring was uninformative, and that's why you did your cost model.

            I'm rambling.  I'm waiting for you to tell me something and I'll be quite.

            DR. RUSSELL:  Right, and that's pretty much what my statisticians have just told me.

            DR. NORMAND:  I rambled a lot, but that, in fact, everybody was included.

            DR. RUSSELL:  That all of the people that were included at the 30-day window, that was the denominator used subsequent.

            DR. NORMAND:  So you proceeded as if you had information on everybody forward in time.

            DR. RUSSELL:  That's what my statisticians have said.

            DR. NORMAND:  Okay.  So I think that's in contradiction to the earlier answer I had in the morning, just so I understand that.

            Now, with that said, I understand that the -- and I'm sorry.  I have to move this far away.  I need to get different glasses or something, but at the nine month -- it raises two separate issues to me.  One is the follow-up issue in terms of how many observations you had in order to make your comparisons, and I see on the table I'm looking at that, you know, certainly for the major endpoints, you had like 96, 97 percent.

            So I'm not so concerned about that.  It's just that it's somewhat misleading --

            DR. RUSSELL:  Yes.

            DR. NORMAND:  -- in terms of the follow-up information.  I'm okay with it because of the Kaplan-Meier analysis that I believe hopefully they were assumed censored. 

            Somebody is coming up.

            CHAIRMAN LASKEY:  They said in the text they were censored, in their stat analysis.

            DR. CHIACCHIERINI:  I'm Dick Chiacchierini.  I'm a statistical consultant to Boston Scientific.  I'm a paid consultant and have no other financial relationship with the company.

            Dr. Normand, you're absolutely right.  I came to the review of the analysis after it was done and asked the very same question and, in fact, had there not been such complete follow-up, I would be worried about the results.

            DR. NORMAND:  Okay.

            DR. CHIACCHIERINI:  Because in a true imputation analyses the ten patients that were lost to follow-up between implantation and nine months should have had an imputed value in order to use them in the denominator.  That was not done here, but it will make little relevance to the outcome.

            DR. NORMAND:  Okay.  Just for clarification, because as I said I asked it earlier in the day and I realize that I think there was either a misunderstanding how I raised the issue because my issue was correct.

            The reason why I'm raising this issue not only relates to the follow-up for the endpoints, but also relates to the comparisons that were presented comparing the various groups.  So, again, if you're describing smoking history or whatever, I'm surprised that you had complete data on everybody.  It may have happened, but, for example, your baseline demographics characteristics because you use this principle of always having the same denominator, I couldn't tell if, in fact, for the characteristics of your population at baseline, whether or not you were missing any of those points, and so hence, when I looked at the comparability, I couldn't figure out.

            So my question is:  were you also missing data describing the baseline characteristics of the population that you presented in earlier tables?  But yet you still used the full sample size in the denominator.

            DR. CHIACCHIERINI:  That's exactly correct, but the number of patients missing, any single observation was very small.

            DR. NORMAND:  So even the baseline measurements might be --

            DR. CHIACCHIERINI:  Even the baseline measurements.  I mean, the monitoring was extremely well done so that the number of patients that were missing, height, weight, or any of the medical history variables, was so small that it would not have made a difference.

            DR. NORMAND:  Okay.  So the reason why I raise this issue, clearly, as you know as a statistician, but for other people, is that even though on the available observations that they're not different when at least one is reviewing the material, one would like to see whether or not information is missing differentially between the two treatment groups.

            It wasn't there.  So I couldn't tell.

            DR. CHIACCHIERINI:  It turns out that what you're saying is correct regarding the clinical data, that it did use the full denominator. 

            The angiographic data and the IVUS data used the data where there was an actual number of data points.

            DR. NORMAND:  Okay.  Thank you.

            So it's difficult to assess balancing and comparability.  If you don't separate out the missing data, I can't see the denominator.  So given the small number that's missing, maybe it's not such a big deal, but it's difficult to follow.

            DR. CHIACCHIERINI:  Excuse me a moment.  You can do that.

            DR. NORMAND:  I'm sorry.

            DR. CHIACCHIERINI:  There is a table on accountability very early on.

            DR. NORMAND:  There are 3,000 pages here or something.  Okay.

            DR. CHIACCHIERINI:  Yes, it's early on in the document, particularly in the clinical report, and it talks about it starts out with the total number that were enrolled.  It takes the 12 away and --

            DR. NORMAND:  Oh, no.  Yeah.

            DR. CHIACCHIERINI:  -- in each one down tells you how many patients were missing certain relevant pieces of information, and the numbers were very small.

            DR. NORMAND:  Okay, but again, as a principal, typically you don't report it that way.

            Two more questions.  Dr. Russell, I think you had said in an earlier response to a question that was raised that there was -- my interpretation of your answer was that there was a lot of site variability, that you had prefaced this with, you know, statistics was difficult because the measurements were quite varied from site to site.  I'm pretty sure that's what you said.

            The reason why I'm asking that question is I'm wondering whether or not in your analysis you incorporated center specific effects or was that ignored in your statistical analysis?

            DR. RUSSELL:  No, we did do an analysis of pooling and then an analysis of centers, but my reference to the variability had to do with the preclinical results.

            DR. NORMAND:  Okay.  So it was really isolated to that.

            DR. RUSSELL:  Right.

            DR. NORMAND:  So you actually did a test to see whether or not the data were poolable across the 73 sites.  Okay.

            And I had one final question, and that has to do with your registry, which I was looking at, and let me see where it is.

            Yes, the sites that will participate in that registry, how are they chosen?

            DR. RUSSELL:  We've used the MEDPAR billing database to identify volume at community centers, at centers.  Then we've used our analysis of centers that are considered tertiary compared to non-tertiary centers by dividing those that are affiliated with a teaching university, and then we've basically tried to select sites that have not participated in our previous studies after the first ten and select sites where we believe that they will have capacity to actually enter the data.  They'll have some sort of research staff or interest in getting research staff.

            So that's how we've done it.

            DR. NORMAND:  Okay.  So it sounds like it's certainly a well thought through strategy on how to select sites to follow through.

            Thank you.

            CHAIRMAN LASKEY:  Dr. Morrison.

            DR. MORRISON:  Well, I'd love to ask a very insightful question that no one else have thought of, but I don't have any.  So I'd like to just make a --


            DR. MORRISON:  -- brief comment and say that what really impressed me first in going through these two volumes before coming here, and secondly, in listening all day to your presentations and the FDA's attempt to dissect all of the work is that you didn't really put together one team to do one study.  You put together multiple teams to do a series of studies going from relatively simple anatomy and single vessel to greater complexity.

            Clearly, there are a million questions that could be asked based on the fact that you don't have all of the world of data, but given what you had, it seems to me a very impressive effort with a lot of planning at each stage, and as I said, all I can do is echo what my colleagues have said because, given the patients that have been studied and the way things are presented, I think it's a beautiful job.

            DR. YANCY:  Once again, I would congratulate the sponsor and investigators for really putting forward a very nice piece of work that should be of some significance in the clinical arena.

            Just as a matter of statement, I would reiterate my earlier comments that it would be my desire to see a more robust clinical profile when we're dealing with what we think are culprit lesions that have caused people to present to an angiographic suite for intervention.  It would seem as if there would be an opportunity to assess the clinical impact beyond just the major adverse cardiac events.

            The specific issue that I'd like to raise just for clarity, because it seems as if this has been dispelled and has been addressed by both the FDA and the company, but we came to this meeting with the understanding that one stated concern was the loss of intimal smooth muscle cells after exposure to the drug eluting stent, and I think in both presentations, the sponsor and the FDA, there was a comment that this is acknowledged to occur, but there was a quick statement that it did not appear to be related to any significant adverse events.

            I'd like to just know a little bit more about your position on the loss of intimal smooth muscle cells, particularly in reference to the data regarding incomplete apposition.  I'm wondering if there is a concern here that over a longer term exposure to the stent that we may see some more significant disease down the road beyond just nine to 12 months related to this loss of integrity of the vessel smooth muscle cells. 

            So that would be my specific concern, if you could address that, please.

            DR. RUSSELL:  Well, let me being with the TAXUS II IVUS analysis.  We were concerned about local drug delivery at the time that TAXUS II was initiated, and elected to use IVUS, postulating that it was the most sensitive way of looking at vascular responses and would allow us to look at the media, unlike any other diagnostic modality.

            We had excellent IVUS follow-up with many paired analyses, and we were able to demonstrate with those studies that there was no significant change within the groups for EEL.  We found that reassuring.  In fact, we were able to demonstrate that the difference from baseline to follow-up was actually increased for what we call peri-stet area.  Some people call it plaque burden.

            We were heartened by that.  We had postulated that if there were to be a significant and ongoing drug effect that didn't heal, that there might be thinning in that region.  So the TAXUS II data argues that the initial medial cell loss is associated with what we saw preclinically, which is the replacement of the lost cells by what we're calling fibrous tissue.

            And on that note, I'd actually like Dr. Wilson to come up because we have asked him to specifically analyze the medial cell loss and provide his comments as the pathologist.

            DR. WILSON:  Unquestionably, there is more medial cell loss with the drug eluting stents.  This is in the preclinical animal studies, than with the controls.  However, in the typical case, the amount of that loss of medial smooth muscle cells is really quite modest with the SR formulation.  It is really with the MR formulation that it is more substantial.

            Now, in all cases, and this is every slide, every stent, that medial cell loss is accompanied by a replacement with collagen so that the media always remains structurally intact right through to the longest follow-up we have, one year in the pigs.

            Now, it turns out that the strength and resilience of the wall of an artery is not dependent on its cellular composition.  One can actually gently remove all of those cells by detergented enzymatic digestion and then do mechanics on the vessel, and you will find that the strength and, in fact, other passive mechanical properties remain the same.

            So the strength of the vessel wall is not determined by the amount of medial smooth muscle cells present, and histologically the vessel wall is always structurally intact in all of the vessels surrounding the drug eluting stents in the animal studies.

            DR. YANCY:  Thank you.

            DR. RUSSELL:  I'd like to ask Dr. Schwartz to comment because his experience with bare metal stents is extensive.

            DR. SCHWARTZ:  Thank you.

            I think Dr. Wilson said it.  It was not absence of medial cells.  It was less medial cells, and even with that, there were still fibroblasts that were there. 

            I think more important physiologically, it's very clear in the animal models that when you lose muscle mass in terms of smooth muscle cells that actually stimulates new intimal hyperplasia.  In fact, we saw the opposite here, that there was less new intimal hyperplasia at the site because of the stent.

            So I think the argument can be made very strongly that, in fact, yes, there was less, but it had little to no physiologic impact.  In fact, the effect of the drug was very positive, as we saw in the preclinical trial as well.

            So I think we have good physiologic information that despite there was less medial cells, they were not absent.  It did not seem to have any deleterious effect at all on the vessel.

            CHAIRMAN LASKEY:  Bill.

            DR. MAISEL:  I just have a couple of questions.  The first relates to the entry criteria for the lesion.  It's mentioned that the lesion length was to be between ten and 28 millimeters, but fully one third or more of the patients had a lesion length that was measured to be outside that range.

            So perhaps you could discuss the rationale for picking a minimum lesion length of ten millimeters at the start of the trial and explain why approximately a third of the patients did not have a lesion length that fell within the predetermined range.

            DR. RUSSELL:  I think that's why there is a Core Lab business.  Investigators make their own assessments of lesion length and reference vessel diameters to make the decision about inclusion of patients, and I'll have Jeff come up and comment on the disparity between investigator assessment and Core Lab assessment.

            DR. POPMA:  I'm Jeff Popma.  As the Director for the Angiographic Core Laboratory for the study at the Brigham Women's Hospital, we did receive research money for the performance of this study as we have done with other competitive trials.  I have done speaking talks for Boston Scientific, as well as the other competitors on stent designs.  All of my speaking engagements have been with under the guidelines for the Harvard Clinical Guidelines Committee for Conflicts of Interest.

            Having said all of that, you're absolutely correct that the investigator at the clinical site doing the angiogram usually comes up with a different lesion length than we do in the Core Laboratory.  The reason for that is that as a clinician looking at an angiogram, we begin to think that we want a stent from fully shoulder to shoulder, wherever there is a luminal irregularity that we see on the angiogram.

            That's very difficult for us to quantitate in the core laboratory, and as a result of that, our automated algorithms us a luminal narrowing of more than 20 percent before it begins to call the shoulder of the lesion.  Just by abligation, it is always shorter when we use a 20 percent criteria than the investigator looking at an angiogram who sees a luminal irregularity.

            And this has been a relatively consistent theme through most of the studies, that the investigator always calls the lesion length a bit longer than we do in the core laboratory.  We want to make sure we have reproducible measurements.  So we use a fixed criteria of a 20 percent narrowing.

            That's how the numbers came out the way they did.

            DR. MAISEL:  Great.  That's helpful.

            Can you comment on the minimum lesion length for the entry criteria?  Why was ten millimeters picked?

            DR. POPMA:  Well, I think ten millimeters was picked because of the need to have a good shoulder of drug eluting stent, proximal and distal.  So if our 16 millimeter stent length and a ten millimeter lesion, we were allowed then to have a three millimeter margin proximally and distally.

            One thing we have learned from the drug eluting stent trials thus far is it's important to have adequate stent eluting drug outside the margins of vessel injury as best one can do.  So I think that was specifically designed because of the stent lengths that were available.

            DR. MAISEL:  So the related question would be in the labeling.  It's mentioned that any lesion less than 28 millimeters would be acceptable, although the trial, you know, set a parameter, a minimum parameter.  Do you think that's --

            DR. POPMA:  Yeah, I think, you know, this is a good thing.  Should we send every film at the investigatory site that gets treated with a TAXUS stent to the Core Laboratory for analysis before we do it?

            I think we should.


            DR. POPMA:  But unfortunately, I don't think that's going to actually be applicable for clinical practice.  So I think what we have to do is to rely upon what was done with the clinical protocol, which is a visual assessment of lesion length, something that's treatable with a 16 millimeter stent as a minimum, and then up to 20 millimeters because of the availability of the longer 32 millimeter stent for clinical practice.

            DR. MAISEL:  Thanks.

            My second question relates to the fact that you're seeking -- oh, I'm sorry?

            DR. STONE:  If I can interrupt, I'll add one other thing, too, on that end.  In TAXUS I and II, which would support the data to this application, they tested very short focal lesions, anything less than 15 millimeters.  So visually they were often three millimeters, five millimeters, et cetera.

            So I think with that as supportive information, in addition to the fact that many of these, quote, ten to 12 millimeter lesions were, in fact, truly shorter lesions by any estimate of the criteria, and given the fact that we did see benefit in all lesion lengths and vessel sizes, then I think it's reasonable to include less than 28 millimeters.

            The sponsor is also seeking approval for several stents that were not used in the clinical TAXUS IV trial, and specifically I'm interested in comments related to the eight and the 12 millimeter stents, which were not used at all.

            Were they available for this trial?  And if so, why were they not used in over 1,300 patients?  And is there any clinical data at all related to stents this short that are drug eluting?  And if not, how about non-drug eluting?

            DR. RUSSELL:  The first question is that they were not available at the time that the study was initiated.

            With respect to the issue of why should they be included, we believe that the most likely use of a second small stent would be to optimize outcomes in the setting of dissection, and this provides a reasonable matrix for operators to use easily to get optimum outcomes.

            There were no issues identified with longer stents.  So it's unlikely that there would be issues identified with the smaller stents.

            Do we have any data on eight and 12 millimeter stents?  We have data that we're collecting in our ongoing TAXUS V studies, but that is short-term data, and now I'm going to let Gregg speak.


            DR. STONE:  I think from a clinical point of view, we always want to match the lesion as close as possible to the stent, given that we probably do want an extra two or three millimeters, as Jeff was saying, to be able to get into the normal reference segment.

            One of the drawbacks, I think, if any, in this trial was that we didn't have shorter stents available.  So when we did miss the lesion length by one or two millimeters or we did have a one or two millimeter long edge dissection, we had to put another 16 millimeter stent in, and the drawback of that is not only is there some relationship between every stent and restenosis, but also you tend to cover more side branches as you're putting in longer and longer length stents.

            So there's a true clinical need to be able to have shorter lengths of stents to be able to match the lesion needs and dissection lengths to the stent length.

            We have a lot of experience with bare metal stents with eight and 12 millimeter stents, and we know that those are appropriate in our hands to be able to accurately place them where we need to, and so while there isn't data yet, this is, I think, an extension of what a reasonable clinical need would be, and given that the doses of both the polymer and the drug are less or no more, that it should be safe and effective.

            DR. MAISEL:  Yeah, I guess I would defer to my interventional colleagues on the panel, but I guess I harbor a little concern that the shorter stents would result in frequent multiple stent uses if there's a little area that needs just another eight or 12 millimeters of stenting.  But enough said.

            I don't have any other questions.

            CHAIRMAN LASKEY:  Well, let me take it just a half a step further.  You're kind of encouraging people to cover dissections with this, the availability of it, or asking for approval, and yet the vast majority of the literature would support leaving these little edge dissections alone, whether it's drug eluting or not.

            So I wonder if we can dredge this a little bit further.

            DR. RUSSELL:  Yes, most of the circumstances when a second stent has been put in all of the studies, TAXUS I, TAXUS II and TAXUS IV, were when the investigator judged the dissection to be problematic for acute procedural success.  So they were not the low grade dissections.

            Our focus on using a second drug eluting stent and mandating it in TAXUS II and TAXUS IV arose from our experiences in TAXUS I and TAXUS III, when it was mandated that if a second stent was required, it should be a bare metal stent.

            And what we discovered with that protocol mandate was restenosis occurring in the bare metal stents when two stents had been used.  So TAXUS IV was put together to prospectively study that, and the protocol, as you know, mandated that a second study stent should be used, and as you saw from the small amount of multiple stenting data that we had, the outcomes looked promising.

            DR. STONE:  Also, what we've actually required in this protocol and most of our other protocols was that while very small NHLBI Type A dissections could be left and those don't seem to have any consequence, anything of a dissection of greater than Type B was stented.  So you know that you've got good luminal in-flow and luminal outflow without any irregularity.

            And I think, number one, that partly explains in both arms, the bare metal stent arm and the drug stent arm, why the stent thrombosis rate was so low because we did have complete lesion coverage with no persistent dissections.  I agree that's never been 100 percent proved, but it seems like the bulk of the suggestive evidence would support that.

            In addition, there's also a lot of situations where you think your lesion length is about, you know, 20 to 22 millimeters.  Maybe I can get by with a 24 millimeter stent, but I'd rather not put in a 32 millimeter stent.   So you try the 24 millimeter stent.  You happen to be a millimeter or two short and then unfortunately you do need a second stent, and it is probably better to be able to match closer even if you do need a second stent than to use one long stent, which will, again, cover more side branches and be more associated with more periprocedural myocon (phonetic) necrosis.

            CHAIRMAN LASKEY:  Which I guess raises the question again.  Chris touched on it, but in the IFU are you encouraging people to get full coverage plus?  Is that stated explicitly, that you need to include the lesion plus? 

            Should you?  Ought you?

            DR. STONE:  I'll let Dr. Russell speak to what's actually in the IFU.  It was a recommendation in the protocol.  Did the lesion coverage go from as near normal reference segment to as near normal reference segment as possible, realizing that these are often diffusely diseased vessels and that any dissection greater than or equal to Type  B gets stented?

            And that is certainly our standard practice, but I'll let Dr. Russell speak to what's actually in the IFU.

            DR. RUSSELL:  We do not comment at this time specifically in the IFU about coverage.  We'd be happy to work with FDA on the right language.

            CHAIRMAN LASKEY:  It seems to be de rigueur in the community.  So maybe we should make it official in the drug eluting stent era.

            DR. RUSSELL:  Works for me.

            CHAIRMAN LASKEY:  I just had one final question.  To follow up on  Bill's point, less about the discordance between the Core Lab angiography, but the CEC evaluation of the events, which of course is the primary endpoint, you had about an 83 percent concordance.  Did I read that right?

            So 17 percent of the time the investigator at the site felt that there was a process requiring revascularization, but the CEC did not feel that way.  Is that right?

            DR. RUSSELL:  Yes, and that had to do most frequently with the percent diameter stenosis criteria that the CEC uses, and occasionally due to the absence of any symptomatology with the percent diameter stenosis that was not very impressive.

            CHAIRMAN LASKEY:  Because I guess a 17 percent miss rate is not too bad,b ut it's a little disturbing, but finally, did it sort out evenly in the two arms?

            DR. RUSSELL:  It was even in the two arms.  It's a pattern that's actually lower than we've seen in some of our other larger studies.  We've seen discordance of up to 45 percent, depending on the study and the rigor, let me say, of the CEC.

            The mandate is on the monitoring staff to actually carefully review the records and collect information about symptomatology.  As Gregg will point out, this study requested that before the angiogram at follow-up was performed, that investigators actually fill out a form indicating if there were symptoms or positive exercise testing and send it in ahead of time to help the CEC in assessing whether it was ischemia driven or not.

            CHAIRMAN LASKEY:  Sorry.  Dr. Aziz.

            DR. AZIZ:  Just one more pathology question.  Did you look at the vasovasorum of the vessels of the stents that were placed in animals?

            DR. RUSSELL:  I'll let Rob answer that question.  It's one of his favorite topics.


            DR. SCHWARTZ:  A superb question.  We have done it now unofficially.  We actually presented it in abstract at TCT, suggesting that there is an inhibition of near angiogenesis in the vasovasorum.  It's currently in preparation for manuscript, but I have a fellow working on that right now.  So there is somewhat a significant inhibition of angiogenesis of additional vaso.

            DR. AZIZ:  And the PAM material, is there a lot of ex cellular matrix?  What exactly is that, PAM, the amorphous material?

            DR. SCHWARTZ:  Excellent question.  Against our advice, we had actually said sequestered PAM, S-PAM, which would have made it SPAM, and that was actually deferred by Dr. Russell.


            DR. SCHWARTZ:  In all seriousness, the important things to know about SPAM are really four salient points that Dr. Russell, I think, pretty clearly outlined.  First, we know what it is.  It's essentially fibrin.  It's very benign in the context.  There was little to no inflammation associated with it.  It goes away over time.  It seems dose dependent.

            It also was found in a few of the control animal sections as well.  So in that context, it has absolutely no effect as far as we can see on the safety of the animals.  It's really based on those four points.

            DR. RUSSELL:  Just to follow up on that point that Rob has covered, we have speculated based on the paucity of white cells, which we believe is an effect of Paclitaxel, that the clearance of fibrin is initially delayed, and then once macrophages return, we see the resorption.

            CHAIRMAN LASKEY:  Thank you.

            We have that precious opportunity to either be on time or perhaps even end early.  Does anybody up here want to take a break or should we proceed with the questions?

            Yes, sir.  In that case, I'm going to ask the Executive Secretary to now read the questions, please.

            MS. WOOD:  Does the combination of the nine month clinical data from the pivotal TAXUS IV SR formulation study and the adjunctive data from TAXUS I SR formulation and TAXUS II SR and MR formulations adequately address the potential concerns raised by the animal studies?

            CHAIRMAN LASKEY:  I will try and summarize the consensus of the panel on this.  As usual, please wade in if you want to clarify or disagree.

            I think at the nine month clinical point, clinical endpoint, the data is convincing.  However, none of us up here are comfortable with a nine month cutoff to establish safety concerns, certainly those raised by animal studies.  I think we would like to see longer term follow-up, particularly given the discordance in progression in animals versus man.

            Are we in agreement?

            DR. ZUCKERMAN:  Okay.  Dr. Laskey, I think you need to clarify your response.  Does it provide reasonable assurance of safety, not absolute assurance?

            CHAIRMAN LASKEY:  This is not my response.  This is the collective response, and I've just surveyed the nodding of heads that we are all in accord with safety out to nine months, but we await additional long-term clinical data to be fully assured.

            MS. WOOD:  Are the clinical studies presented adequate to address concerns about possible adverse bent from interactions with drugs typically administered to the target patient population?

            CHAIRMAN LASKEY:  There is a member of the panel that feels that there -- and I would agree with this -- that there are some theoretical issues about drug-drug interaction which have not been evaluated.  From a practical standpoint, there does not appear to be any evidence of drug-drug interactions with drugs most likely to be used in this population.

            Can we be 100 percent certain?  No.

            Is that fair?

            MS. WOOD:  Please comment on whether the clinical studies adequately address other drug interactions that are likely to be important or of interest.  If not, what other information or studies should be provided?

            Specifically, please consider the potential for the following types of interactions, with anti-neoplastic agents, with chemotherapeutic agents where a hypersensitivity reaction could be endused.

            CHAIRMAN LASKEY:  Well, do you want to summarize that?

            I don't think we spent a great deal of time on this issue.  It's a very small point, and I think it's a very targeted, very small population, that we shouldn't hang ourselves up on.

            DR. SOMBERG:  Well, I would say we do not have adequate information at this time, but it shouldn't preclude the availability of the product, but this should be collected from the planned studies, and I think the panel pretty much would encourage a long-term follow-up in animal studies as well, with potentially some activity.

            Specifically, I think what the question is getting at, and which we didn't discussion, was, for instance, doxorubicin, you know, the chemotherapeutic agent, can antagonize the activity of Paclitaxel, and we talked about some of the potentiating effects.

            So I think that that's long term, and there are very small safety and very small efficacy issues that can come of it, but for instance, in a large -- I'm going on on this, but to make it very brief -- in a large registry one can develop a database.  You know, who's on torvastatin and who isn't on it?  And then see if there's any improved benefit or lack of benefit of the stent in that situation.

            But I think that it's similar to the drug area where not everything is known when something is available, and certain interactions will have to come out, and this is certainly not with the P450 polymorphisms where you can predict something.  So we're going to have to be able to develop a larger database than just 1,000 and a half patients.

            DR. WEINBERGER:  I'd just like to make one comment about this.  I don't think that we're exposing patients even in the worst case scenario to an infinite time of hypersensitivity.  This drug presumably has some finite residence time in the vessel wall, and after that, an initial month or so.  The release rates are very slow, and I don't think that even on a theoretical basis we should be concerned with drug-drug interaction when the second drug is going to be administered out of the first couple of months after implantation.  I just don't think that that's going to be much of a likely interaction time.

            DR. WHITE:  But are you concerned that if you're treating the patient who is going to be actively treated with chemotherapy should there be something in the IFU for the physician that raises concerns?

            I mean, it's an unusual patient, but is there a caution?

            DR. WEINBERGER:  I think that that's what's being asked in this question, whether or not we want to take a position as to whether or not that's contraindicated or let's say that it's unknown.

            CHAIRMAN LASKEY:  Well, I don't think that we can say it's contraindicated if there's no body of knowledge.  So it's certainly unknown.

            DR. WHITE:  I would think that most clinicians are not aware of the potential potentiators or problems, and so simply providing information of the possible issue, informing them and allowing them to ask questions would be appropriate, I think.

            DR. ZUCKERMAN:  So, Dr. White, can you extend your comments?  I think you and Dr. Weinberger have understood the intent of this question.   Is the practical solution that you're suggesting to perhaps beef up Section 5.6 of the label which talks about potentials for drug interaction?

            DR. WHITE:  Yeah, that would make me happy.  I don't see a problem at all, but I think as a user myself, I would like to know if there was a potential problem in a potential patient, and that information could be provided, I think, to us, where an unknown risk, but perhaps a risk, and just raise that awareness.

            CHAIRMAN LASKEY:  I guess we'll come back to that when we vote, but that may be the Pandora's box.

            Can we also be clear about what we mean by hypersensitivity?  I'm not sure we're all on the same page on that one.  You mean IgE mediated events with typical IgE mediated clinical sequelae or what do you mean by hypersensitivity?

            DR. ZUCKERMAN:  Not necessarily.  I think this whole issue needs to be better thought out and discussed with the sponsor.

            CHAIRMAN LASKEY:  As well as for ourselves if we need to.  Okay.

            MS. WOOD:  Do the clinical data submitted from the pivotal TAXUS IV SR formulation study, plus the data from the adjunctive TAXUS I SR formulation and TAXUS II SR and MR formulation studies, provide reasonable assurance of safety?

            CHAIRMAN LASKEY:  I'd say the panel is universally in agreement that there is a reasonable assurance of safety at the 270-day point, perhaps at the 365-day point if we looked at the materials supplied to us from the AHA presentation as well.

            Beyond that we need to be very, very circumspect.

            MS. WOOD:  Does the clinical data at 270 days presented on the TAXUS stent from the pivotal TAXUS IV study provide reasonable assurance of effectiveness?

            CHAIRMAN LASKEY:  That I believe  there's no question.

            MS. WOOD:  Does the evidence presented on the TAXUS product support the proposed labeling indication?

            CHAIRMAN LASKEY:  Well, we can generate some discussion here.  There were several items that were brought up this afternoon that may or may not lead us to consensus.  The first is the short stent is the eight and 12 millimeter stents.  So can we have some discussion on that?

            DR. WHITE:  I think everything the sponsor said and I think in that discussion about using the shorter lengths was appropriate.

            The other thing that I didn't hear said was that there's a deliverability issue that comes with shorter stents, and there are times when that second stent is needed that a longer, 16 or 28 millimeter stent may be very difficult to deliver, and so the ability to use smaller links really helps the doc in the cath lab.

            And so Ii would want to give this issue the benefit of the doubt for the smaller subsets in order to improve safety and usability in the cath lab if there is not real concern or real evidence that there's a problem with the shorter lengths.  I think I would give them the benefit of the doubt for improved cath lab safety.

            I would want those.  I would want those on my shelf.

            CHAIRMAN LASKEY:  As a clinician, I agree, but does the evidence presented on the TAXUS products support that?  We'd have to say no, but we're reacting as clinicians at this point; is that right?

            DR. WHITE:  I think you have to ask yourself, you know, this is the same -- I mean, again, it goes with the different sizes.  You know, it's the same stent mounted from two and a half to three and a half.  So it's not a different device because they're also asking for a diameter that hasn't been tested as well, but it's the same device.

            So to me cutting the device shorter simply means less total drug delivery.  So the whole risk here would be lower dose or maybe loss of efficacy perhaps, but I don't see how it increases risk or makes it unsafe because of a lower dose.  I mean, I don't see any evidence of unsafety.  So I think that would be reasonable.

            DR. NORMAND:  Can I?  I'd like to just make the point that there's no evidence presented.  So just to emphasize that.  As I said, there's no data about that.  So I just want to -- you can guess or give good clinical opinion, but there's no evidence.

            CHAIRMAN LASKEY:  Thank you.

            DR. HIRSHFELD:  I'd just like to echo what  Chris said.  I think that the shorter lengths belong in the tool box and that we would be hampering clinicians if they didn't have access to the entire complement of lengths that are requested.

            CHAIRMAN LASKEY:  Clinicians happy?


            CHAIRMAN LASKEY:  Was there another?

            DR. WHITE:  Let me say that this is different than if we were considering a size that had not been tested.  Let's say that somebody said we'd like to get approval for a one and a half millimeter device and that wasn't tested.  I would have concern about that, but I don't see the fact that the data hasn't been presented is much less worrisome to me in this configuration just because the length gets shorter than it would be in a different diameter vessel.

            I view it as much less of a leap of faith for these shorter lengths, and if I was going to be writing this or if they hadn't asked for approval of these, I'd be the one telling them to cut them up, give them to me shorter because I'm the one who has been struggling for two hours to get that long stent through the other one, and I want a shorter piece.

            DR. ZUCKERMAN:  Okay.  Dr. White, we don't want you cutting up your stents.


            DR. ZUCKERMAN:  But I think you understand the spirit of the question here.  As Dr. Normand has correctly indicated, there are no real data in the submission, but the FDA is asking for an independent expert opinion from expert interventional cardiologists, and I think we've heard your opinion and will act accordingly.

            The second part of Question 5(a) though, perhaps the more important one, is the addition of reducing restenosis to the indication.

            DR. MORRISON:  Well, as a clinician rather than a statistician, I would like to carp about the characterization there's no data.  I mean, we were told that in Texas one and two by design, bare metal stents were used for the purpose that now in TAXUS IV, the drug eluting stent, was used, and what they noted was restenosis in the area of the bare metal stent.

            Now, that's not --

            PARTICIPANT:  A different length.

            DR. MORRISON:  Yeah, but they used overlapping stents, bare metal, and demonstrated restenosis.  So we have the kind of clinical data that we use every day to make inference.  It's not large scale, prospective, randomized trials, but it sounds pretty clear that you don't want to tack up an edge dissection with the express stent.

            And then we've gone through at great length different metals, different, you know, drugs and so forth.

            So it seems to me pretty clear if you have a flow limiting edge dissection, you don't want to use a bare metal stent to tack that up.  So that's within the spirit of labeling.

            CHAIRMAN LASKEY:  Well, just to clarify, those were near stents.

            DR. WHITE:  But I think the answer or to direct Bram's question is I think there's no question the efficacy has been measured in restenosis and TLR, and so I think they deserve labeling for that indication.

            DR. WEINBERGER:  As a practical matter, if you don't supply an eight millimeter stent for the most common use, which would be a edge dissection tack-up, what will happen is people will take a longer stent and just get a larger amount of overlap, which I would argue is a worse thing to do.

            So I think that from a practical point of view, from a real world clinician point of view, you have to make those shorter lengths available.

            CHAIRMAN LASKEY:  Full circle.  We are in agreement then.

            DR. WEINBERGER:  One other comment about the label, about the indications, and that is it's not just that we're being told that the indication includes reduction of restenosis, but there's a statement that there is less restenosis, improved outcomes in nine months compared to uncoated stents, and just in the interest of fair disclosure, I think that you can say that there is less restenosis than Express stents, but not all bare metal stents.

            So I think that making an unqualified statement that there's a better outcome at nine months compared to bare metal stents is slightly overstating what was studied.

            CHAIRMAN LASKEY:  Well, it's easy enough to change that.  It does reduce restenosis in comparison to the bare metal stent comparator.  That's easy to reword.

            DR. WEINBERGER:  That would be fine.

            DR. ZUCKERMAN:  Are there any other comments on the main indication given by the sponsor?  There's no indication, for example, that these are symptomatic patients, symptomatic de novo lesions?

            DR. MAISEL:  I did have one comment on the labeling with regard to total occlusions and acute myocardial infarction, which is not really mentioned anywhere.  I didn't know if that warrants mentioning as, you know, it clearly wasn't included in the trial and whether that should be mentioned one way or the other.

            DR. WEINBERGER:  You know, I think that we get into trouble if we have to try to list every exception because I think Greg listed a whole bunch of things that were not studied.  So I think that --

            CHAIRMAN LASKEY:  Well, the inclusion/exclusion criteria go into the IFU.

            DR. WEINBERGER:  Right.

            CHAIRMAN LASKEY:  So it's clear from the get-go who's in and who's out.

            Dr. Zuckerman, your question about the patient population, again, about symptomatic and asymptomatic?

            DR. ZUCKERMAN:  Yeah.  Ms. Goode, can you put up the proposed indication?  This is perhaps one of the key parts of the label to get right, and we've heard several comments, but we want to get it all right.

            DR. SOMBERG:  Are you suggesting that there was data presented that may have suggested that it might benefit symptoms as well as just this?

            DR. ZUCKERMAN:  No.  If we were to summarize what has been said so far, I think Dr. Weinberger has indicated we should be reducing restenosis compared to the bare metal stent that was used in the trial because other stent designs and materials will be shown in the future and may be better than drug eluting stents.

            The second point is that perhaps, well, just from a regulatory viewpoint we want to correctly describe the patient population that this stent is intended for, and therefore, one might consider using the word "symptomatic" in front of "de novo lesions."  If one has, for example, a 50 percent lesion on an angiogram where the patient is not symptomatic with a very high work load, does one need this device or any other interventional device?

            DR. SOMBERG:  You clarified the point, but it cuts both ways.  If you put the word "symptomatic" in, I think that's what our chairman was going to do, but I don't want to put words in his mouth here.

            But if you say "symptomatic," it suggests that we're doing something to benefit symptoms as well, and while there was a suggestion about the database, I'd be concerned.  So I think the way it stands is pretty neutral.

            I'm a little concerned about saying just the bare metal in the Express stent because that's sort of surrogate for the state of the art as it is now, and I can see somebody else saying, "Well, if you used ours it wouldn't, you know, compare," or whatever.

            So I think, you know, the word just "restenosis" is with bare metal stents at this point.  If something comes along that's different, that would be fine, but I think that's an inroad to deceit.

            DR. HIRSHFELD:  The statement as written has an implied comparison and doesn't state what it's compared to.  The actual data is compared to the bare metal Express stent, and so I think the panel needs to decide whether the bare metal Express stent is a generic reflector of all bare metal stents or whether there may be something different about the Express stent than other stents.

            Now, in favor of considering a generic is that the actual Core Lab restenosis rates are very similar to what we've seen in many other trials of similar lesions, and so I don't think it's to large a leap of faith to say that the data that are presented are relatively generic for other bare metal stents.

            CHAIRMAN LASKEY:  If the inclusion and exclusion criteria are going into the IFU, I believe the inclusion criteria do use the word.

            DR. SOMBERG:  Symptomatic.

            CHAIRMAN LASKEY:  Documented stable angina or unstable angina, ischemia, et cetera, et cetera.  So it is there.  Do you need to repeat it again in the label then?

            DR. ZUCKERMAN:  Where are you referring to?

            DR. WHITE:  I think the question though, Bram is that if the indications that are used for the trial are included in the IFU and you're going to start to abstract certain words as more important, would you -- I mean, I don't know where to stop.

            I mean, the problem with symptoms is that they're subjective, and I don't think I want to.  You might want to say "appropriate for revascularization," but, I mean, there's a whole bunch of language that you could attach to that.

            DR. MORRISON:  I thought a quarter of the patients in TAXUS IV had silent ischemia.  Well, silent ischemia is asymptomatic where I come from.  Is that different from the East Coast?


            CHAIRMAN LASKEY:  It said on page 29, Volume 2 of 2, Section 9.2, 9.3.1, inclusion criteria.  Read number three.  "Documented stable angina CCS Class 1, 2, 3, 4 or unstable angina with documented ischemia, brown wall," et cetera, et cetera, "or documented silent ischemia."  That's the universe.

            DR. MORRISON:  So ischemia is asymptomatic on the East Coast, right?

            CHAIRMAN LASKEY:  Does that help?

            DR. ZUCKERMAN:  Yes.  I think you've defined the patient population.  And is there a consensus as to how much of a comparison statement should be for reducing restenosis compared to what?

            MR. MORTON:  Just if I could point out on the indications, when you look at page 6 of 31, that's rather nicely done in two paragraphs there, and so rather than change this paragraph, we might want to consider the way it's expressed in the second paragraph.

            DR. MAISEL:  I think that concept that there's an implied --

            DR. ZUCKERMAN:  Yes.  I think we need to take a step back and recognize that the agency traditionally would remove that second paragraph that Mr. Morton is referring to.

            And let me try to explain why by just reading the regulatory definition of indications for use.  It's a general description of the disease or condition the device will diagnose, treat, prevent, cure or mitigate, including a description of a patient population for which the device is intended, and that is in the CFR.

            The type statements that are found in that second paragraph would generally go in the clinical trial section of the label to actually describe what happened in the trial in appropriate detail.

            PARTICIPANT:  I think the statement regarding reducing stenosis that there's an implied comparison, I'm a little uncomfortable making the general statement that it reduces stenosis compared to all bare metal stents because evidence hasn't been presented there, and I think the information is provided in the IFU.  So I personally don't think we need anything more than that.

            CHAIRMAN LASKEY:  But I thought we suggested just the addition of the term" the Express" or the "bare metal stent, Express stent."  That seemed to fit the bill.

            DR. WEINBERGER:  So is what you're saying, Bram, that there's a mandate to include a description of the patient population in this one paragraph statement under indications?

            DR. ZUCKERMAN:  Ideally we should in the indications for use have a description of the intended patient population, and is this adequate is the question here.

            DR. WHITE:  I think the patient population that was included in the trial was the universe.  I mean it was.  It was asymptomatic ischemia.  It was stable and unstable angina, right?

            DR. WEINBERGER:  Well, but it boiled down to patients with some sort of documentation of ischemia.  That's what it boils down to.

            DR. WHITE:  So we could say that.

            DR. WEINBERGER:  Fine.  That I'm happy with.  Are you happy?


            DR. WHITE:  Yes.

            CHAIRMAN LASKEY:  For the penultimate time, are we in agreement and is the agency content with that?

            DR. ZUCKERMAN:  Yeah, we received enough feedback on this sentence.

            CHAIRMAN LASKEY:  Yeah.

            MS. WOOD:  Please comment on whether the labeling should specify that multiple stents should only be used for bailout purposes, dissection, insufficient lesion coverage, and whether in these cases the shortest stent available, such as the eight millimeter, should be used.

            CHAIRMAN LASKEY:  I think to summarize the last 15, 20 minutes of discussion, we feel that these stents should be approved, that is, the eight and 12, approved for other reasons, for reasons of expediency and efficiency, and just the general availability of these stents, not necessarily for coverage of dissections for which there was not enough documentation in this particular protocol.

            MS. WOOD:  Please comment on whether the labeling should address the potential combination of the TAXUS stent with an additional drug eluting stent in the same vessel.

            CHAIRMAN LASKEY:  It should.

            DR. WEINBERGER:  Is there any proposal as to what it would say?

            MS. WOOD:  Please comment on whether the label --

            CHAIRMAN LASKEY:  Just a -- I don't know.  That wasn't entirely facetious.

            DR. WEINBERGER:  No.  It's a serious question.

            CHAIRMAN LASKEY:  That will probably come up, and I think it's worth some discussion.  So, Judah, you'll have an opportunity to suggest language when we vote with conditions.

            You're new to the process I know, but you'll have an opportunity to suggest such language, but we need to concern ourselves not only about drug-drug, but stent-stent interaction.  So that will come up.  We'll get to that.

            MS. WOOD:  Please comment on whether the labeling recommendation for post procedural antiplatelet regimen is appropriate and whether additional recommendations on procedural anticoagulation regimens are warranted.

            DR. WHITE:  I think the sponsor agreed to add the six-month Plavix or Ticlid.  I think that needs to be specified, not just antiplatelet.  I don't see any reason to get into the cath lab anticoagulation issues.  There didn't seem to be any difference with different agents that we used.

            CHAIRMAN LASKEY:  But I heard you get the sponsor to agree that the inclusion of six months of Plavix is critical.

            Sorry, John?

            DR. SOMBERG:  I was just going to say, I mean, I feel that way possibly, too, but do we really -- we have to ask ourselves do we have data to support that.  And I think we're reacting to the concerns that there is a thrombogenic --

            DR. WHITE:  The data you have to support it though is the trial was conducted that way, that the investigators were all encouraged for six months.  So we're just basing the recommendation on the data, evidence that we have.

            If you want to get the results they got, you've got to do it their way.

            DR. SOMBERG:  That's good point.

            MS. WOOD:   Please comment on any other aspects of the labeling, such as the contraindications, warnings, precautions, such as use with bracytherapy, in conjunction with other procedures, et cetera, drug pharmacology, pharmacokinetics, or specific drug safety information, such as use in special populations, warnings, precautions.

            DR. WHITE:  I mean, the answer is yes.  We've been talking about all of these, and I think that we need to have -- well, we don't have information, but we have concerns.  The concerns need to be raised and listed and informed, and we may not have a lot of data about a lot of these concerns, but I don't think it's wrong that we should raise the issue that we have concern because it's potentially a difficult thing.

            CHAIRMAN LASKEY:  Okay.  So the contraindications to the use of this stent are?  What do we all agree on as an absolute contraindication?

            There's a history of allergy to nickel or stainless steel?  What other contraindications are there?

            DR. WHITE:  I don't see any reason not to accept the contraindications that were used in the trial.

            CHAIRMAN LASKEY:  Okay.  So keep that.

            And as far as the warnings and precautions, obviously adjunctive or precedent brachytherapy was an exclusion criteria.  So it simply remains good clinical sense that it's a warning and precaution.

            DR. WHITE:  Did the sponsor comment on the presence of a CYPHER stent?  Was a CYPHER stent -- did that kick them out of TAXUS IV?

            So we can say that as well.

            DR. ZUCKERMAN:  Dr. Laskey, in going through this section, which is really a critical section where the FDA needs your expert panel feedback, are you turned to the labeling section?  Right now we're on page 7.

            CHAIRMAN LASKEY:  Page 7 of?

            DR. ZUCKERMAN:  Of the IA instructions for use.  So it's in Volume 1, instructions for use.

            CHAIRMAN LASKEY:  Right.

            DR. ZUCKERMAN:  So the contraindications are listed in Section 3.

            DR. WHITE:  So we would want to add that a serolimus stent would also -- the presence of a serolimus stent would be a contraindication.

            DR. WEINBERGER:  In that vessel.  Is that right, in that vessel?

            DR. SOMBERG:  And we also don't know the interaction of several drugs that have been talked about before.  I wouldn't make it a contraindication.  It's just that we just don't know the interactions.

            DR. WEINBERGER:  That would be the drug interaction section, right?

            DR. SOMBERG:  Yeah.

            And can I just ask a quick question?  There was no comment made.  Do we know anything about radiation therapy, brachytherapy with this material?  I mean were there any patients in the clinical trials who got that?

            DR. WEINBERGER:  That was an exclusion.

            DR. SOMBERG:  That doesn't mean they didn't get it.  I know many trials that have also sorts of things that are excluded.

            DR. RUSSELL:  Yeah, in TAXUS II and IV combined, there have been 17 patients that have had bracytherapy used to treat instant restenosis, and of those, not surprisingly, only three of them were in the TAXUS group, for which there was one reported adverse event, but that was not a MACE event.

            DR. SOMBERG:  Why do you say that?  It was blinded, right?  So they shouldn't know which one.  Why was there a bias towards the plain metal?  Am I --

            DR. RUSSELL:  Because more patients developed restenosis in the control group.

            DR. SOMBERG:  Oh, okay.  Yeah, okay.

            DR. RUSSELL:  In our ongoing studies obviously the investigators have been compliant.  So we haven't had any brachytherapy use.

            DR. SOMBERG:  As a physician, I'd just say list that sort of information because it's most critical because if there's a dire -- you know, we don't know this, but if there's something that's dire, pressing, you know, you can't do anything else and you want to do something,  this is the only experience we have and there's really a paucity of it.

            CHAIRMAN LASKEY:  Again, it was drawn to our attention in Section 5.3 on page 7.  It says explicitly "safety and effectiveness of the use of brachytherapy to treat ISR in a TAXUS Express stent has not been established."  So the language is there.

            DR. MORRISON:  Well, I think the language is there to say it's not known for both CYPHER stent and brachytherapy, but I would not include that in contraindication any more than in five --

            CHAIRMAN LASKEY:  It's a warning.  It's just a warning.

            DR. MORRISON:  Right.

            DR. SOMBERG:  You don't want to put it in the contraindications.

            CHAIRMAN LASKEY:  Yeah.

            DR. SOMBERG:  I agree with you.

            DR. ZUCKERMAN:  Okay, and the same principle applies to another comment that was previously made where a panel member wanted to put a contraindication of a CYPHER and a Boston Scientific product, and again, given the lack of data, perhaps it's Section 5.6 that needs to be expanded regarding all of the comments about potential drugs and their interaction with this particular product that have been discussed today.

            Is there any comment on that notion?

            DR. YANCY:  If you want to do that, it would appear that Section 5.2 might be the better spot for it because it comments on having multiple stents with similar composition, and so perhaps if there are multiple drug eluting stents, you can comment that they should have similar drug profiles.

            DR. SOMBERG:  I must say the section on drug interactions I have a log of problem, specifically Section 5.6, because it implies that because there's such a low concentration there won't be a drug interaction, but that's not what we're talking about.  We're talking about a local effect and then a systemic effect of a drug that might interaction in that local phenomena, and there may actually be a major interaction there.

            So, I mean, this is a nice story, but it's based in I can make up another story.  It's not based in fact.

            DR. MORRISON:  Isn't this a little bit analogous to a legal decision being a precedent for subsequent decisions?  I mean one of the things I can see the agency getting into is consistency as there are more and more drug eluting stents and more and more other technologies, and if you start setting the precedent that everything that isn't known is a contraindication, that really defeats the purpose, it seems to me.

            DR. SOMBERG:  I don't want you to take my statement as a contraindication.  They're just warnings that we don't know what the effects are and let somebody make a decision on that basis.

            But as he said earlier, as Dr. White says, some people are more into the mechanical aspects and don't know the potential interactions.  Some people know a lot about the interactions, but never do anything mechanically.  So that's why you put it in the document.

            But I don't want it to be said as a contraindication.  In fact, I agree with you.  Sirolimus and, you know, unless we have some data, it's not a contraindication, but it's a very strong warning.

            DR. ZUCKERMAN:  Right.  I think that would be the agency's perspective, but the key issue here, as Dr. Somberg noted, is that we want to as best as possible have truthful and accurate statements about this drug eluting stent in the label, given that it's probably the best source of information for this product.

            And what I heard him say is that Section 5.6 needs to be reworked with the sponsor.

            DR. YANCY:  Well, I might speak against that.  I think that from a pragmatic standpoint nothing that I saw today suggested that we were dealing with detectable levels of the agent in the serum, and I also think we've heard evidence that animals proves that if it's a logarithmic scale where there's drug release in the early phase and minimal drug release late, and I think that there is enough ambiguity in this statement to suggest what's unknown, that to go further, the necessity for data then becomes upon us to say that we have hard data that this drug-drug interaction is a real possibility.

            I am comfortable with 5.6 as it's stated.

            DR. SOMBERG:  I just want to say that the systemic level of Paclitaxel has nothing to do with the drug interaction.  We're not talking about a systemic level.  We're talking about local effect and a local interaction.  So I think you have to mention that.  I also feel strongly that if this is not mentioned that I'd like to query the company again in going through each of their studies and asking them how many was on, how much was on a torvastatin, how much wasn't, did you do an analysis by that particular drug, did we do an analysis on the preclinical work, and we can go through all of this step by step.

            That would be demanded on a drug level, and I think this is a drug device.

            CHAIRMAN LASKEY:  I'm not so sure.  I mean, that is an endless road, number one. 

            Number two, this is not a study of 100,000 or a half a million or five million people.  The event rate for the toxicity there is below the detection level.  So I would tend to line up behind Dr. Yancy on this one.  I don't know how much you really want to beat this stone, but there are contraindications which we seem to agree on. 

            There are warnings and precautions, which we would like to draw a line somewhere at the end of that, but the list of drug-drug interactions in patients with cardiovascular disease and their permutations is several orders of magnitude.  So I'm not sure it's useful to go there unless you have a huge population base.

            I would agree with Clyde that what we have here is probably adequate with some fine tuning, but to ask for information on every single conceivable drug interaction with Paclitaxel is unreasonable.


            DR. HIRSHFELD:  I think one way we might be able to address this conundrum is to divide it into two things.  The first is the potential interaction between the Paclitaxel on the TAXUS stent and a systemic drug, another chemotherapeutic agent that a patient might be concomitantly receiving.  We had some discussion about that today, and I think that would be an appropriate issue to raise in the instructions for use, just to make sure that the investigator considers that.

            The second is to just merely make a statement that there is no information available about the interaction between stents that elute sirolimus and stents that elute Paclitaxel, and that the clinician should be aware  that this has not been studied and there's no information about this and that they should take that under advisement in selecting a product.

            CHAIRMAN LASKEY:  And by extension, the use of dissimilar stents.  Okay.

            DR. HIRSHFELD:  That's fine.

            CHAIRMAN LASKEY:  We want to be helpful.

            DR. NORMAND:  I had a comment about the label.  Are we done with the label now or no?

            CHAIRMAN LASKEY:  No.


            DR. NORMAND:  Well, could I just make a -- I don't think there's going to be any discussion about it, but the tables where they report the numbers, it's the issue I raised earlier about the long sample sizes that you report in the tables.  So I think they need to be fixed in the label because you've got the wrong numbers and misleading rates.

            DR. ZUCKERMAN:  Can you clarify also, Dr. Normand, for the primary endpoint should that be a Kaplan-Meier estimate at nine months, given that problem and a comparison that way?

            DR. NORMAND:  It doesn't matter because they're making the same assumptions.  Kaplan-Meier, I'm assuming that they said that it's censored.  The censoring was uninformative, whereas I still didn't have the answer to my question, but if the P value is based on the real numbers, then it's the same.  It's the same inference.  It's just that it's misleading to report the rates that are listed here because they're not the rates, and so I would just clean that up.

            MS. WOOD:  Please discuss long-term adverse events that may be associated with TAXUS stents and whether the proposed five-year follow-up on the clinical trial cohorts and the proposed pre/post marketing study are appropriate to evaluate the chronic effects of the implantation of the TAXUS stent.

            If not, what additional information should be collected?  Specifically discuss how long the patients should be followed and what endpoints and adverse events should be measured.

            CHAIRMAN LASKEY:  So I think we're certainly comfortable with longer term follow-up than what we have and five years is certainly practical and should be informative if there's a signal there.  The devil is in the details about the clinical population in whom these stents are being implanted, but certainly the long-term follow-up we all agree is absolutely mandatory.

            The additional information that should be collected is of the nature of the beast.  What types of patients are being procedurized and under what circumstances is the procedure being done?  In effect, how well is the protocol being adhered to?

            DR. WEINBERGER:  Is the intent of this really to look at low probability  outcomes that are important to clinicians?  I assume that that's what you're really interested in in post marketing surveillance.  So the question is not just who's getting in and who's getting it, but what's happening to them.

            So I think that it's not just MACE, but it's MACE and some of the lower, sort of the noisy things that appeared as possible complications.  We want to know if they pop up during the five years, and you want to know whether people develop aneurisms.  You want to know whether or not they develop pericardial effusions.  You want to know whether or not they develop things that we were theoretically worried about that haven't shown up in the first 1,000 patients, but might show up over five years when you give more time for the clinical syndromes to develop.

            DR. WHITE:  Well, I think rather than to specify the issues though, Judah, since you don't know what those are, I mean, in five years of following these patients, I think that's an adequate amount of time to follow these patients.

            DR. WEINBERGER:  I agree with that.  The question is from a practical matter, you're collecting data.  There's going to be a central data repository.  There's going to be a form.  What are they going to be asking for?

            DR. WHITE:  But I don't think we're prepared to write that form for them.

            CHAIRMAN LASKEY:  We're just discussing the need for it rather than anything.

            DR. SOMBERG:  I'm not sure if it his is implicit in this statement, but it should be, that there has to be some sort of concomitant group that you follow it.  I think some of the industry follows their device and you have 43 of this or 75 of that, and you don't have anything to compare to, and you might see the same number in a concomitant population.

            So it's important to, I think, follow concomitantly your bare metal stent.  Is that built into this?

            MR. WHITE:  You mean following a control population?

            DR. SOMBERG:  Well, some sort of comparator.  Here it would probably be the bare metal stent or something of that nature because, say, you got 100 people had neuropathy over five years in the Express stent and you had 110 in the bare metal stent.  What does that tell you?

            But if you just had 100, you would say, "Hey, let's Taxol it.  It reduces neuropathy."

            DR. NORMAND:  Yeah, I had sort of the same comments.  There just wasn't enough -- I don't think you meant to put a lot of detail in this, but when I looked at the proposal for the registry and sort of the surveillance, there's lots of details that aren't here, and I had the same reaction.  It seems to me you need somebody else to follow up.

            There's lots of questions I had in terms of what you're really looking at, but I just didn't think that you were meant to have that much detail here because actually it says in terms of, let's say, the registry, that you're going to look at all of the people who have the indication.  I guess you're not interested in off indication.

            Again, there were details here that I didn't think we wanted to get into necessarily, that good suggestions could be made to help get some more information, but again, I don't know how much detail you want.

            DR. MAISEL:  I think it's asking a great burden for them to have a five-year controlled trial of people not getting a drug eluting stent.  I think the ball is going to be in a different place in two or three years and maybe everyone is going to be getting drug eluting stents.  I think there has been plenty of people who have gotten stents, and there are historical controls.  I think that would be quite reasonable.

            DR. NORMAND:  We could argue about design issues.  I could tell you how to design it so that it would be efficient, but again, what do you want to gain from this?

            If you want to gain adverse events rates, that's one thing.  If you want to -- I mean, we could think about how to make it better, but again, I think that there's not enough detail here to really --

            CHAIRMAN LASKEY:  Long-term adverse events, Sharon. 

            DR. NORMAND:  Well, that's a pretty simple statement.  So if we say long-term adverse events, we could talk about in an absolute sense versus a relative sense.  We could talk about a design, which is it only for the patients in which the label is written for, or do you want to get, you know, again, detail, detail?

            DR. SOMBERG:  But to keep it simple, you just need some -- and I'm not saying it has to be concomitant.  It may be just this population followed out plus the as they accrue in.  It's not that one type of stent will disappear immediately, but there's a great danger of just having your product being followed and a reporting number which has no comparator.

            I mean, I don't want to get into specifics, but that has just come out with another stent with another problem.

            DR. MAISEL:  Do we have any interest in long-term pathology, such as either autopsy or animal studies?

            CHAIRMAN LASKEY:  Well, we do, but I don't know how you enforce that sort of thing.  It's hard to enforce under trial conditions.

            I think it would be worth hearing Dr. Zuckerman on the whole issue about what the agency needs for post marketing surveillance so that we can perhaps be a bit more pragmatic in terms of what we suggest.

            DR. ZUCKERMAN:  Good.  I think the key word is pragmatic because we want to, first of all, extend upon the point that you made, Dr. Laskey.  We have safety data out to nine months.  In an idea setting we might want safety data out to ten years, but there's a preapproval/post approval balance.  So both the sponsor gave you an indication of the main safety problems they would be looking at.  The agency is in general concurrence, and in addition, the agency gave you an estimate of the precision with which we could look at low frequency events.

            I believe Ms. Goode had a slide where 2,000 patients or a one percent event rate.  Our delta is about .4 percent, and with all of that said, we're looking any other practical suggestions about any other unusual safety concerns that you would want us to look for over the next five years.

            We also have heard the point about comparing to an adequate control, but I think the panelists should recognize that this sponsor has a large, in-house, historical bare metal stent database that can be utilized.

            Any other suggestions?

            CHAIRMAN LASKEY:  No one has mentioned it, but explicitly and specifically the endpoints and adverse events should be those that we  looked at at nine months to start.

            DR. ZUCKERMAN:  Okay.  Any other adverse events that you're concerned about?

            CHAIRMAN LASKEY:  Well, there's the hematologic dyscrasia blip.  I think that needs to be looked at.  Again, that's another one of these if you do 50 comparisons something will show up, but because of the potential hematologic toxicity of this agent, it makes sense to look for that in terms of non-cardiac toxicity.

            Other cardiovascular sequelae may or may not be taken up by the fold, but certainly what Dr. Yancy has suggested, the repeat hospitalizations as an index of either efficacy or safety is a very useful handle on how these products are doing in the real world.

            DR. ZUCKERMAN:  Okay.

            CHAIRMAN LASKEY:  Sorry, Chris.

            The Executive Secretary will now read the --

            MS. WOOD:  I've already done that.

            CHAIRMAN LASKEY:  You did that.  Good.  Good job.


            CHAIRMAN LASKEY:  Well, do you want another two minutes to add additional comments or the end of our discussion?  Dr. Zuckerman.

            DR. ZUCKERMAN:  The FDA has no additional comments.

            CHAIRMAN LASKEY:  Does the sponsor have additional comments before the vote or questions?

            MR. OCWIEJA:  The sponsor would not have any other additional comments except to again extend our thanks to you for taking the time to take a look at our data and information and discuss it as well as you have.

            Thank you.

            CHAIRMAN LASKEY:  Mr. Morton?

            MR. MORTON:  Only to echo what's been said before, to recognize that the sponsor brought a well planned and well managed study and also to acknowledge what they had presented to us for a peri and post approval study, and to compliment the agency for concise and pertinent review.

            CHAIRMAN LASKEY:  Concise, as always.  Thank you.


            CHAIRMAN LASKEY:  Dr. Hughes?

            DR. HUGHES:  Thank you. 

            I, too, would like to compliment the manufacturer, the FDA staff, and the committee for the overall preparation of the PMA package and its review and this dialogue.

            I think I do want to say something that perhaps goes along with the last question on post market evaluation.  Actually, a question or a point of clarification from the consumer's perspective in terms of the post approval registry.

            I guess in the fourth from the last or third from the last slides that the manufacturer presented, you introduced MILESTONE II and WISDOM and also ARRIVE as your long-term registries, and I'm just wondering about the relationship among those different approaches or packages.

            I take it that these are, you know, specific to earlier versions of TAXUS, and you know, the current drug eluting coronary stent, you know, for ARRIVE, and so I'm wondering, you know, what kind of lessons were learned from the previous two and also maybe just a little bit about user interface in terms of entering the data into the registry because with this type of device where, you know, it looks like it's indeed a very good, useful device, you know, safe, but we are going into territory or treading into unchartered waters a little bit when it comes to, you know, drug device or interactions or products, as well as how long to go out and looking for  interactions and data like that.

            So I was wondering if the manufacturer could say a little bit more than what they've already said about that.

            DR. RUSSELL:  Yes, thank you very much for the opportunity to address this issue.

            We've gained substantial experience in how to conduct trials and work at gaining investigator compliance.  Most investigators and studies sign agreements, and we work diligently to gain their compliance, but it requires a moderate amount of effort.

            We will be following the patients five years in the studies that we have ongoing.  So that includes TAXUS IV, V, VI and V-ISR.  Hence, we will have substantial control data, and it will be useful control data because it will be conducted under formal preapproval trial processes.

            I say that as a prelude to the contrast of the issue you raised, which is in the post approval registry world in the two programs we have, WISDOM and MILESTONE II, are what I would term transitional programs.  So they are not the purest form of post market study, and we do request substantial information from participants, and it's not even fair to call them investigators.

            However, if you are in the real world in a busy cath lab as an active operator, it is very burdensome for you to enter the volume of information into a Web-based system that is required in the preapproval study.

            So we have worked diligently with FDA to streamline progressively our data collection to gather the essential information about the safety events, but what one must remember from a post approval study is we can tell you what was reported and we can monitor against the report, but unlike a preapproval study, we cannot tell you what was not reported.

            So I have worked actively with my staff to avoid the use of any of the phrases that are used in the preapproval world.  We are not reporting target lesion revascularization.  That requires some amount of angiographic follow-up.  It requires mandated follow-up.

            We are reporting events that are reported by customers, and we are trying to collect that in an accurate manner and monitor against that, but the information collected is only as good as the people entering the information in the database, and if you want real world data from real world community physicians, we may run into a conflict here.

            So we have worked very actively in designing the study to make sure we are making it user friendly, making it an easy Web base that's intuitive, looking at sites that are willing to commit to staff, but we have had to invest huge amounts of effort to get our other post market registries up to dynamic data collection, and so I just want the panel to consider the practical sides of being a busy practitioner who is using the stent in real life and the kind of data that we're really practically going to be able to provide.

            Along those lines, an important difference in this particular registry, ARRIVE, from the previous registries is we mandated that participation in the registry follow the directions for use.  So that was the inclusion criteria.

            So if you had a patient and they fit the inclusion criteria, they got enrolled in the registry, but that doesn't mean any patient that received a TAXUS stent was included in that registry.

            The purpose of ARRIVE as negotiated with the FDA is to collect consecutive TAXUS stent use, which I think is a different purpose than you may have been addressing in your earlier conversation about the real world use of the scent and its associated safety outcomes.

            And the way we have worked on the case report forms interactively with FDA has been to add the option of reporting other events that the implanter believes are related to the TAXUS stent and to take out some of them for the convenience of reporting as opposed to allowing them to enter verbatim language, which can make it very difficult to capture and categorize events.

            So just to back up, WISDOM was a transitional study, had 778 patients about enrolled.  All stents used at those sites were required to enter data into that study, and we specifically were tracking the worrisome events of stent, MI, death, and use patterns to see if they had shifted.

            MILESTONE II is an extension of MILESTONE I.  MILESTONE I was a use registry collecting information on how the bare metal stent was used in our European experience.  We then hope to use that information as relative control information to the use of the TAXUS express stent.

            We are collecting events that are reported by sites, doing monitoring, but as I point out, this is only of events that are reported by the sites.

            And then ARRIVE, the details are still in development, and it is really an effort to work at a real world registry that works.

            DR. HUGHES:  Thank you.

            That's all I have.

            CHAIRMAN LASKEY:  Thank you.

            Before we proceed to the vote, I'd like to briefly open the meeting to the open public hearing portion again, and we have a request from Jim Gustafson to address the panel.

            MR. GUSTAFSON:  Give me a minute to get set up here.

            Okay.  Mr. Chairman, members of the panel, my name is Jim Gustafson.  I'm Vice President of Research and Development for Possis Medical, maker of the AngioJet thrombectomy catheter system.

            I have  no financial interest in Boston Scientific or any other stent manufacturer or developer, but I will confess to having a compelling financial interest in Possis Medical, which goes pretty much all the way down the line.

            I've asked to address the panel on the topic of resolving thrombus prior to the use of drug eluting stints, which goes back to Question 5(e)(ii) that the panel just considered from the FDA.

            Intercoronary thrombus is recognized as one of the most challenging, complicating factors in performing percutaneous coronary interventions.  In an article published in the American Journal of Cardiology presenting the pivotal VeGAS 2 trial data for the AngioJet catheter, Dr. Richard Kuntz, et al., stated, "Percutaneous coronary or vein graft interventions in thrombus containing lesions carry high risks for complications, including abrupt closure, late reocclusion, periprocedural MI, emerging CABG, and death."

            And we believe thrombus is present in coronary lesions more often than is usually appreciated.  In a study published in Circulation by Dr. Chris White, et al., who just left the room, 122 angina patients were evaluated with both angiography and angioscopy.  The angiography identified thrombus in 20 percent of these patients, but angioscopy identified thrombus in 60 percent of these very same patients. 

            These results suggest that lesion thrombus is significantly under diagnosed in the very kind of patients that were enrolled in the pivotal trials for both the CYPHER and TAXUS stents.

            In fact, the approved indication for the CYPHER stent states that the safety and effectiveness of the CYPHER stent has not been established in patients with unresolved thrombus or in AMI patients with evidence of thrombus, and the clinical trial for the TAXUS stent had the same exclusion.  So actually as we saw today in the draft IFUs, the exact same text is included.

            We believe that unresolved thrombus presents both safety and effectiveness concerns when considering the use of a drug eluting stent.  This animation illustrates two important safety issues.

            First, that stenting in the presence of unresolved thrombus can cause a cheese grater effect, creating many smaller thrombus particles that could embolize downstream.

            And, secondly, that unresolved thrombus in the treatment field today creates an environment conducive to thrombus reformation in the future.

            Did that animation run up on the screen?  It did not.  It's running on my screen.  Well, it starts out, but it doesn't run.  Let's try that.

            I apologize.  We tested this at lunch time.

            CHAIRMAN LASKEY:  That's all right.  We've all been there.

            MR. GUSTAFSON:  But I'm not sure what to do about it.  Anybody who's more computer savvy than I am?  That should be hard.  Well, you could all gather around my screen, but I don't suppose that would be particularly effective.

            CHAIRMAN LASKEY:  Sir, it's not critical that we have the video.   If you can proceed.

            MR. GUSTAFSON:  All right.  I'll do.

            CHAIRMAN LASKEY:  Thank you.

            MR. GUSTAFSON:  I'll go on to actually the previous slide.  This animation illustrates two important effectiveness concerns:  first, that stenting over unresolved thrombus can result in undersizing the stent, producing suboptimal stent apposition; and, second, that pinning the thrombus between the stent struts and the vessel wall may compromise drug delivery.

            Now, this slide. 

            These safety and effectiveness concerns may be minimized by effectively resolving thrombus before stenting.  So what does it mean to effectively resolve thrombus?  This table shows various treatment strategies and their effect on thrombus.

            Glycoprotein 2B3A inhibitors are effective at preventing new thrombus, but they do not have an appreciable effect on already existing thrombus.  Thrombolytics can effectively dissolve thrombus, but lytics can take many hours and some patients cannot tolerate it.

            Conventional angioplasty and stenting displaces thrombus, but as the failed animation showed, it basically just squishes the thrombus against the vessel wall, and that can create some additional and more serious difficulties.

            Thrombectomy will actually remove thrombus from the target lesion.

            This was going to be an animation of mechanical thrombectomy performed with the AngioJet catheter.  AngioJet is an approved and proven safe and effective treatment of intercoronary thrombus.  Specifically the approve indication states that AngioJet is intended for removing thrombus in the treatment of patients with symptomatic coronary artery or SVG lesions in vessels two millimeters or larger in diameter prior to balloon angioplasty or stent placement.

            And this animation was going to show that with intercoronary thrombus removed from the lesion site, a drug eluting stent can more readily achieve maximum desired stent expansion with full strut apposition and more effective delivery of drug to the target lesion and the target tissue.

            And that was going to compare with this next animation that would have shown again what you would have already seen, suboptimal strut apposition and impeded drug delivery to the target tissue when thrombus had not been removed before loading the stent.

            Therefore, we urge the panel to consider a more explicit statement in drug eluting stent labeling regarding the need to remove thrombus prior to using a drug eluting stent to optimize their safe and effective use in such patients.

            Thank you.

            CHAIRMAN LASKEY:  Thank you.

            Is there anyone else in the audience who wishes to address the panel on today's topic?

            (No response.)

            CHAIRMAN LASKEY:  If not, I'd like to close the public hearing and ask the sponsor one final time:  do you have any additional comments?

            DR. RUSSELL:  We have no further comments.   Thank you.

            CHAIRMAN LASKEY:  God bless you.


            CHAIRMAN LASKEY:  Ms. Wood, if you can read the voting options please.

            MS. WOOD:  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 approval applications, PMAs, that are filed with the agency.  The PMA must stand on its own merits and your recommendation must be supported by 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 on 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:

            Approval if there are no conditions attached;

            Approvable with conditions.  The panel may recommend that the PMA be found approvable subject to specified conditions, such as physician or patient education, labeling changes or a further analysis of existing data.  Prior to voting all of the conditions should be discussed by the panel.

            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 if a reasonable assurance has not been given that the device is effective under the conditions of use 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.

            CHAIRMAN LASKEY:  The panel is now prepared to vote.  As Geretta said, the recommendation of the panel may be approvable, approvable with conditions that are to be met by the applicant, or denial of approval.

            I'd like to now ask for a motion on the PMA.

            DR. WHITE:  I'd like to make a motion that the device is approvable with the following conditions and that those pertain to the labeling issues.

            CHAIRMAN LASKEY:  One at a time.

            DR. WHITE:  One is the Plavix-Ticlid continuation for six months.

            CHAIRMAN LASKEY:  Yeah.  Remember we need to do these one at a time.  So you are moving that the PMA be approved with the following conditions.

            Okay.  Is there a second of Dr.  White's motion for approval?

            MS. WOOD:  You first ask for a motion on the main motion.  You have to get a second on that.  then you list the conditions, and then you take a final vote on the entire package.  So right now we're looking for a second on the motion for approvable with conditions.

            CHAIRMAN LASKEY:  Okay. Do we have a second?

            (Show of hands.)

            CHAIRMAN LASKEY:  Do we now want to enumerate?

            DR. ZUCKERMAN:  We need to talk into the microphone for the transcriptionist and the audience.

            DR. WHITE:  I would defer to you, Warren.  Do you want to summarize what we've talked about since your mind is so much better able to capture these things than me?

            CHAIRMAN LASKEY:  No, not at this hour, Chris.  Thank you.  But we need to -

            DR. WHITE:  We've talked about a lot of things in the labeling which I think are important that we include.

            CHAIRMAN LASKEY:  But you need to make the motion and the condition.

            MS. WOOD:  We start with the first condition.

            DR. WHITE:  The first one would be that it be specified in the labeling that it's recommended that Plavix or Ticlid with aspirin be continued for six months.

            CHAIRMAN LASKEY:  Do I hear a second on this?

            (Show of hands.)

            CHAIRMAN LASKEY:  Okay, and is there discussion?

            Are we all universally in agreement on that, all right?  On that first condition of Plavix plus aspirin for six months.

            I am told that we will now vote on Dr. White's motion to approve with condition number one being the addition of Plavix and aspirin for six months.  All in favor, raise hands, please.

            (Show of hands.)

            CHAIRMAN LASKEY:  It's unanimous.

            May I have another condition for the PMA?

            DR. WHITE:  I'm empty.

            DR. HIRSHFELD:  I could offer two.  Should I offer them once or --

            CHAIRMAN LASKEY:  Unfortunately we need to do it one at a time.

            DR. HIRSHFELD:  Okay.  The second one is that the interaction -- that labeling should state that the interaction between the Paclitaxel eluting stents and stents that elute other compounds has not been established, has not been studied.

            CHAIRMAN LASKEY:  Do I have a second?

            PARTICIPANTS:   Second.

            CHAIRMAN LASKEY:  Discussion?

            (No response.)

            CHAIRMAN LASKEY:  I think we're all in agreement.  All in favor of the second condition being the labeling to specifically reflect the precautions or the potential interaction of the Paclitaxel and sirolimus eluting stent?  All in favor --

            MS. WOOD:  Other drug eluting stents.

            CHAIRMAN LASKEY:  Did you say --

            DR. HIRSHFELD:  I thought it should be other drug eluting stents to allow for future contingencies.

            CHAIRMAN LASKEY:  Okay.  All in favor?

            (Show of hands.)

            CHAIRMAN LASKEY:  Again it's unanimous.  You had a second?  You had an additional?

            DR. HIRSHFELD:  Well, I don't know whether -- the other issue that I had raised was that the labeling doesn't describe the maximum expansion diameter for the stent, and I didn't know whether that's a peccadillo or whether that needs to be officially voted on as the labeling.

            CHAIRMAN LASKEY:  Yes, unfortunately.

            DR. HIRSHFELD:  Yes, it is a peccadillo?

            CHAIRMAN LASKEY:  And we need to vote on it.

            DR. HIRSHFELD:  Okay.

            CHAIRMAN LASKEY:  So it needs to be included.

            DR. HIRSHFELD:  Well, then I would propose that the labeling has to reflect the maximum permittable expansion diameter of the Express stent.

            CHAIRMAN LASKEY:  Of?

            DR. HIRSHFELD:  The TAXUS Express stent.

            DR. WEINBERGER:  Second.

            CHAIRMAN LASKEY:  Discussion?

            (No response.)

            CHAIRMAN LASKEY:  Okay.  We'll vote on this condition of the labeling to reflect the maximum permittable stent diameter.   All in favor?

            (Show of hands.)

            CHAIRMAN LASKEY:  Unanimous again.

            DR. NORMAND:  Do I have to say the condition about the labeling and the numbers in the table?

            I'd like another condition.  I'd like the numbers in the table that report on the primary effectiveness endpoints in the denominators to be corrected to reflect accurately the numbers.

            CHAIRMAN LASKEY:  So the rates in all of the tables?

            DR. NORMAND:  Yes.  They have to be corrected everywhere.

            CHAIRMAN LASKEY:  That will be in the IFU.

            DR. NORMAND:  Yes.

            CHAIRMAN LASKEY:  Okay.  Do we have a second?

            DR. WEINBERGER:  Second.

            CHAIRMAN LASKEY:  If there's no discussion, let's vote.  All in favor?

            (Show of hands.)

            CHAIRMAN LASKEY:  Great.  Unanimous again.

            Are there other conditions for us to consider to Dr. White's recommendation?

            DR. MORRISON:  I'd like to suggest that the labeling also state that the interaction between this drug eluting stent and bracytherapy has not been studied.

            CHAIRMAN LASKEY:  As a precaution that note is in there already, is it not?  It's in the IFU as a precaution.

            DR. MORRISON:  Is it?

            CHAIRMAN LASKEY:  Yes.  So it's technically in the label.


            DR. WEINBERGER:  I think that the discussion of MRI should probably be deleted.

            CHAIRMAN LASKEY:  A bold move.


            CHAIRMAN LASKEY:  A second?  We have established the fact that the Express stent is a non-ferromagnetic stent; is that correct?

            DR. WEINBERGER:  That is correct.

            DR. SOMBERG:  Could we have discussion on that?

            CHAIRMAN LASKEY:  Absolutely.

            DR. SOMBERG:  I don't think it has been studied.

            DR. WEINBERGER:  I think it is pretty well known that if you put this particular alloy, 316L, into a magnet that there's no force on the alloy.  That's a matter of scientific fact.

            DR. SOMBERG:  Has it been studied?  I mean, I'll just be the devil's advocate here.

            DR. WEINBERGER:  Why don't we have the bioengineers tell us that?

            DR. SOMBERG:  I'm just saying I would caution the other panel members if you haven't seen the data, I wouldn't endorse one person's statement.

            DR. WEINBERGER:  The problem here is that if there's any instruction in there about the MRI, there is no data to say that at 30 days it is safe.  So either it's never safe or it's safe immediately.

            DR. SOMBERG:  I would prefer to leave this to the FDA biomaterials and radiologic group which is expert in this area and will be able to either mandate tests or make, you know, a generic warning that is standard for things that are not studied because this is a medical legal issue.

            CHAIRMAN LASKEY:  I am happy tabling this and not making this as a condition of approval.  I think it's a theoretical issue as far as I'm concerned.  I don't think there's scientific basis for it.  I think all of us have sent patients to the magnet within a couple of days, but how does the agency feel on moving forward from here?

            DR. ZUCKERMAN:  I think you've given some good advice, Dr. Laskey.  We'll work with the sponsor to see what data there are, and we'll construct a statement that is appropriate for the data that the sponsor has.

            CHAIRMAN LASKEY:  Okay.  So let the discussion reflect that you want to withdraw the motion?

            DR. WEINBERGER:  Yes, I will.

            CHAIRMAN LASKEY:  The motion has been withdrawn

            DR. SOMBERG:  I wanted to remind the Chairman of his biomaterial compatibility that you mentioned.  You might want to mention that.

            DR. WEINBERGER:  That's in there already.

            DR. SOMBERG:  It is?  Different metals?  Okay.

            CHAIRMAN LASKEY:  Additional conditions of approval?  We only have four, folks.  Come on.  We need some more.


            CHAIRMAN LASKEY:  Okay.  We spoke during the day of the metal-metal incompatibility.  Do we want to include --

            DR. WHITE:  It's already there.

            CHAIRMAN LASKEY:  You're there.  It's in the biocompatibility?  You metal to metal?  Fine.

            DR. MAISEL:  We had an extensive discussion regarding the indication statement which I will allow you to summarize if you like, but do we need to make any changes to that?

            CHAIRMAN LASKEY:  Can I paraphrase that?  We need to include the use of the term "Express stent comparator" in the labeling claim of reducing restenosis compared to.  Is that what you're driving at?

            DR. MAISEL:  Yes.

            CHAIRMAN LASKEY:  Can we have a second to that motion?

            PARTICIPANT:  I second.

            CHAIRMAN LASKEY:  Discussion and a vote.  All in favor of including the comparator? 

            (Show of hands.)

            CHAIRMAN LASKEY:  Okay.  Let's raise hands again because not everybody is unanimous.

            (Show of hands.)

            CHAIRMAN LASKEY:  We have one, two, three, four, five in favor. 

            Those against?

            (Show of hands.)

            CHAIRMAN LASKEY:  One, two, three, against. 

            PARTICIPANTS:  Four.

            CHAIRMAN LASKEY:  I'm sorry.

            DR. ZUCKERMAN:  Okay.  For the record, Dr. Laskey, we need to identify who's voting in favor and who's against this condition of approval.

            CHAIRMAN LASKEY:  So in favor of the condition, including the language of the term of comparator, the bare metal stent comparator being the express stent, Drs. Hirshfeld, White, Morrison, Weinberger and Maisel in favor, and Drs. Somberg, Aziz, Normand and Yancy against.

            DR. HIRSHFELD:  Is there an opportunity to discuss this briefly?

            The reason that I think this is a good motion and why I voted for it is that the language that currently exists contains an implied comparison, but doesn't state what it's compared to, and so I think just from a linguistic standpoint the statement needs to be revised.

            DR. SOMBERG:  The reason I'm opposed to this is that it will be said that other bare metal stents that are existing now will not offer the benefit or will offer benefit for a reduced restenosis while only the Express is the one that is problematic.

            I think the Express is a generic standard for bare metal stents and that the rates are comparable from what I understand from the literature, and therefore, I don't want a detail person saying, "Well, if you use this bare metal stent, you know, it's different; it's okay because they limited the warning that this was better than the Express."

            So I think it can cause confusion, and I think there is as good as you're going to get that this is a controlled study against the bare metal stent.

            DR. HIRSHFELD:  I agree with that, and the qualification is that the panel really didn't come to agreement as to whether they felt that bare metal stents were generic with respect to restenosis or not, and that could be subsequently assessed, and that decision could be made subsequently.

            DR. SOMBERG:  We didn't have data presented on that in detail, but I think when this study was put together, it was put together as a drug eluting stent versus bare metal, and there may be bare metal stents that are not as good as the bare metal stent in this particular comparitor.

            So I just think it's fraught with dangers, and we shouldn't get that specific.

            DR. YANCY:  And I tend to agree with that.  I think that by specifically stating the Express stent we are excluding the possibility of that other bare metal stents are as good as, better than, or not as good as, and I am comfortable that the event rate in the reference population was equivalent where it's reported for the majority of bare metal stents, and I think that this creates a precedent that would be very difficult to future generations of these devices.

            DR. WEINBERGER:  So the way you would like to put it, it would include the possibility that the introduction of a new bare metal stent with a different alloy would never be better than the TAXUS stent.  That's what you're saying.

            DR. SOMBERG:  No.  I'm just saying that we're not saying that the only thing this particular intervention does is compare favorably to the Express stent.  As it now stands, I mean, if you want you can ask them to put the data of restenosis in many restenosis trials, and it all seems to me very, very similar.

            If something comes down the line that uses a new epoxy resin and the will be the first ones to differentiate themselves in the future from all other bare metal stents, and they probably will do a comparator to a drug eluting stent.

            So it will not stymie anybody, but it will not give people a segue to mislead.

            CHAIRMAN LASKEY:  I'm not sure it's misleading, John, at all, and in fact, it's just very specific.  A stent is not a stent is not a stent, and if you look at the rates of TVR with NIR, they're different than in this study.  So I don't see what the problem is, but I don't know where to go from here if we're on the fence with a condition.  I need some legislative help.

            It's five to four, but you need more than that?

            So the condition passes.  Okay. 

            Were there other conditions?  I think we've exhausted the list.  Panel members?

            DR. WEINBERGER:  Under indications for use we wanted to say ischemia driven revascularization because there's no reference to the patient population or --

            CHAIRMAN LASKEY:  Except in the patient population in the IFU.

            DR. WEINBERGER:  Right, but in the indications there was a legislative requirement to mention the patient population.

            CHAIRMAN LASKEY:  Is there, indeed, a legislative requirement?  I don't think so.

            DR. ZUCKERMAN:  I think we reviewed the general framework for indications for use, but  you know, are you offering this as a condition for approval to vote on, Dr. Weinberger?

            DR. WEINBERGER:  Yes, I was.

            DR. ZUCKERMAN:  Okay.

            CHAIRMAN LASKEY:  All right.  Well, we tend to shy away from recommending patterns of practice.  I don't think that's in our purview.  I think we need to stick to what we have before us.

            DR. MAISEL:  Can we be a little more vague and say patients requiring revascularization ore percutaneous revascularization?

            MR. MORTON:  Could I just ask is this particular drug eluting stent that much different than another drug eluting stent or from a bare metal stent?

            And what I'm looking for is just kind of some continuity from device to device so that we're not specifying something with this device unless it really is different.

            In other words, are we simply evolving into a matter of specificity that may not be needed?

            CHAIRMAN LASKEY:  May not be needed and may be a dangerous precedent for us.  I wish we had discussed this a bit greater, but from prior years we've tended to back away from telling people what to do.

            DR. WEINBERGER:  Well, if people are uncomfortable, I don't mind leaving it in the more detailed instructions which follow the indication section.  That would be acceptable.

            CHAIRMAN LASKEY:  Which is where it is.  Great.

            So then we have five conditions on Dr. White's motion for approval.  They are, number one, the addition of the Plavix and aspirin treatment for six months; number two, for the labeling to reflect the drug-drug interaction; number three, for the labeling to reflect the maximum inflation diameter; number four, for the tables to reflect reality; and number five, the inclusion of the comparator term "bare metal Express stent" in the indication statement.

            Will all of those voting members in favor of approval of these conditions raise their hands?

            (Show of hands.)

            CHAIRMAN LASKEY:  We have Drs. Hirshfeld, White, Somberg, Aziz, Normand, Morrison, Yancy, Weinberger and Maisel, unanimous.

            And finally, just a brief rendition of why you voted as you did.  Dr. Hirshfeld?

            DR. HIRSHFELD:  I think the clinical data for both safety and efficacy are compelling, and the product should be on the market.

            CHAIRMAN LASKEY:  Chris.

            DR. WHITE:  I would concur with that.  I agree.

            CHAIRMAN LASKEY:  Dr. Somberg.

            DR. SOMBERG:  Clinical data is compelling and while I'm not happy with the mention of the Express stent because I think the control group is a generic control group, I think that's a minor point and will not overly confuse the practicing physicians.

            DR. AZIZ:  It's safe and effective.


            DR. NORMAND:  Ditto, if you heard that.  I agree with safe and effective.

            DR. MORRISON:  I think there's concordance between multiple well done clinical trials and some extremely well done animal data, and I think we look forward to seeing this on the market.

            DR. YANCY:  I agree with all that has been said.  I just hope that future iterations of this device, particularly since other companies are represented, will bring to bear data that will be a bit more robust with bedside clinical endpoints.  I think that that's very much needed.

            DR. WEINBERGER:  And I think that clearly efficacy and safety were demonstrated so that there was no question that this would be a valuable addition.

            DR. MAISEL:  Very well conducted clinical trials with evidence of safety and efficacy.

            CHAIRMAN LASKEY:  Mike or Allen, any final comments?

            MR. MORTON:  None from me.  Thank you.

            DR. HUGHES:   No additional comments.

            CHAIRMAN LASKEY:  Well, in conclusion, I want to thank the sponsors for an absolutely superb presentation; thank my fellow panel members for sticking to the schedule; and this concludes the report and recommendations of the panel on PMA P020039 -- you gave me the wrong script -- this is P030025.

            Thank you much.  The meeting is adjourned.

            (Whereupon, at 4:34 p.m., the meeting in the above-entitled matter was concluded.)