1

 

                DEPARTMENT OF HEALTH AND HUMAN SERVICES

 

                      FOOD AND DRUG ADMINISTRATION

 

                CENTER FOR DEVICES AND RADIOLOGIC HEALTH

 

 

 

                    CIRCULATORY SYSTEM DEVICES PANEL

 

 

 

 

 

 

 

                       Wednesday, March 17, 2004

 

                               9:00 a.m.

 

 

 

               Hilton Gaithersburg Washington D.C., North

                           Salons A, B and C

                           620 Perry Parkway

                         Gaithersburg, Maryland

 

                                                                 2

 

                              PARTICIPANTS

 

      Cynthia Tracy, M.D., Acting Chair

      Geretta Wood,Executive Secretary

 

      MEMBERS

 

                Salim Aziz, M.D.

                Mitchell Krucoff, M.D.

                William Maisel, M.D., MPH

                Christopher J. White, M.D.

 

      CONSULTANTS

 

                Clyde Yancy, M.D.

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

                John W. Hirshfeld, M.D.

                Thomas B. Ferguson, M.D.

                Norman S. Kato, M.D.

                Brent Blumenstein, Ph.D.

                Charles Bridges, M.D.

 

      Industry Representative

 

                Michael Morton

 

      Consumer Representative

 

                Chrissy Wells

 

      FDA

 

                Bram Zuckerman, M.D.

 

                                                                 3

 

                            C O N T E N T S

 

      Call to Order:

                Cynthia Tracy, M.D.                              4

 

      Conflict of Interest Statement:

                Geretta Wood                                     4

 

      Introductions                                              6

 

      Open Public Hearing:

                Leo Corbet                                       9

 

             Sponsor Presentation: SynCardia Systems, Inc.

                 P030011: Syncardia Systems CardioWest

                         Total Artificial Heart

 

      Introduction:

                Marvin J. Slepian, M.D.                         15

 

      CardioWest TAH Overview:

                Richard G. Smith, MSEE, CCE                     18

 

      IDE Clinical Trial:

                Jack G. Copeland, M.D.                          26

 

      Clinical Perspective:

                Walter E. Pae, M.D.                             54

 

      Conclusion:

                Marvin J. Slepian, M.D.                         64

 

      Questions and Answers                                     67

 

                            FDA Presentation

 

      Summary:

                Eric Chen, M.S.                                 83

 

      Statistical Summary:

                Lilly Yue, Ph.D.                                88

 

      Clinical Review:

                Julie Swain, M.D.                               98

 

      Questions and Answers                                    108

 

      Open Committee Discussion                                130

 

      Open Public Session:

                Robert Jarvik, M.D.                            290

                Aly El Banayosy, M.D.                          292

 

      Recommendation and Vote                                  297

 

                                                                 4

 

  1                      P R O C E E D I N G S

 

  2                          Call to Order

 

  3             DR. TRACY:  Good morning everybody.  Happy

 

  4   St. Patrick's Day.  I would like to call to order

 

  5   this meeting of the Circulatory System Devices

 

  6   Panel.

 

  7             The topic today will be a discussion of a

 

  8   premarket application for the Syncardia Systems

 

  9   CardioWest Total Artificial Heart P030011.

 

 10                  Conflict of Interest Statement

 

 11             MS. WOOD:  The following announcement

 

 12   addresses conflict of interest issues associated

 

 13   with this meeting and is made a part of the record

 

 14   to preclude even the appearance of an impropriety.

 

 15             To determine if any conflict existed, the

 

 16   agency reviewed the submitted agenda and all

 

 17   financial interests   reported by the committee

 

 18   participants.

 

 19             The conflict of interest statutes prohibit

 

 20   special government employees from participating in

 

 21   matters that could affect their or their employers

 

 22   financial interests, however, the agency has

 

 23   determined that participation of certain members

 

 24   and consultants, the need for whose services

 

 25   outweighs the potential conflict of interest

 

                                                                 5

 

  1   involved, is in the best interests of the

 

  2   government.  Therefore, a waiver has been granted

 

  3   for Dr. Clyde Yancy, and a waiver was previously

 

  4   granted for Dr. Judah Weinberger, for their

 

  5   interests in firms that could potentially be

 

  6   affected by the panel's recommendations.

 

  7             Dr. Yancy's waiver involves consulting

 

  8   services with a firm that has a financial interest

 

  9   in the product at issue for which he receives an

 

 10   annual fee of less than $10,000.  His services are

 

 11   not related to the subject matter before the panel.

 

 12             Dr. Weinberger's waiver involves a

 

 13   stockholding in a firm that has a financial

 

 14   interest in the product at issue.  The value is

 

 15   between $25,001 and $50,000.  The waivers allow

 

 16   these individuals to participate fully in today's

 

 17   deliberations.

 

 18             Copies of these waivers may be obtained

 

 19   from the  Agency's Freedom of Information Office,

 

 20   Room 12A-15 of the Parklawn Building.

 

 21             We would like to note for the record that

 

 22   the agency took into consideration other matters

 

 23   regarding Drs.   Thomas Ferguson, Mitchell Krucoff,

 

 24   Cynthia Tracy, Judah Weinberger, and Clyde Yancy.

 

 25   These panelists reported past or current interests

 

                                                                 6

 

  1   involving firms at issue, but in matters that are

 

  2   not related to today's agenda.

 

  3             The agency has determined therefore that

 

  4   these individuals may participate fully in the

 

  5   panel's deliberations.                       

 

  6                                 In the event that the

 

  7   discussions involve any other products or firms not

 

  8   already on the agenda for which an FDA participant

 

  9   has a financial interest, the participant should

 

 10   himself or herself from such involvement, and the

 

 11   exclusion will be noted for the record.

 

 12             With respect to all other participants, we

 

 13   ask in the interest of fairness that all persons

 

 14   making statements or presentations to disclose any

 

 15   current or previous financial involvement with any

 

 16   firm whose products they may wish to comment on.

 

 17             DR. TRACY:  Thank you.

 

 18             Can I ask the panel members to introduce

 

 19   themselves, please.

 

 20                          Introductions

 

 21             MR. MORTON:  My name is Michael Morton.  I

 

 22   am the industry representative and I am employed by

 

 23   Carbomedics.

 

 24             DR. WEINBERGER:  I am Judah Weinberger.  I

 

 25   am an interventional cardiologist at Columbia.

 

                                                                 7

 

  1             DR. YANCY:  Clyde Yancy, UT Southwestern

 

  2   in Dallas, cardiologist.

 

  3             DR. WHITE:  Chris White, Ochsner Clinic

 

  4   Foundation in New Orleans.  I am an interventional

 

  5   cardiologist.

 

  6             DR. KATO:  Norman Kato, cardiovascular

 

  7   surgeon, private practice, Encino, California.

 

  8             MS. WOOD:  Geretta Wood, Executive

 

  9   Secretary.

 

 10             DR. TRACY:  I am Cindy Tracy.  I am from

 

 11   George Washington University, electrophysiologist.

 

 12             DR. FERGUSON:  Tom Ferguson,

 

 13   cardiovascular surgeon, Washington University, St.

 

 14   Louis.

 

 15             DR. KRUCOFF:  Mitch Krucoff.  I am an

 

 16   interventional cardiologist at Duke Medical Center,

 

 17   and Chair of Clinical Trials for devices at the

 

 18   Duke Clinical Research Institute.

 

 19             DR. MAISEL:  William Maisel.  I am an

 

 20   electrophysiologist at Brigham and Women's Hospital

 

 21   in Boston.

 

 22             DR. BLUMENSTEIN:  I am Brent Blumenstein,

 

 23   biostatistician in private practice.

 

 24             DR. BRIDGES:  Charles Bridges,

 

 25   cardiothoracic surgeon, University of Pennsylvania.

 

                                                                 8

 

  1             DR. AZIZ:  Salim Aziz, adult cardiac

 

  2   surgeon, clinical associate professor, George

 

  3   Washington University.

 

  4             MS. WELLS:  Chris Wells.  I am the

 

  5   consumer representative on this panel.

 

  6             DR. ZUCKERMAN:  I am Bram Zuckerman,

 

  7   Director, FDA Division of Cardiovascular Devices.

 

  8             MS. WOOD:  Pursuant to the authority

 

  9   granted under the Medical Devices Advisory

 

 10   Committee charter dated October 27th, 1990, and as

 

 11   amended August 18th, 1999, I appoint the following

 

 12   individuals as voting members of the Circulatory

 

 13   System Devices Panel for this meeting on March the

 

 14   17th, 2004:  Clyde Yancy, M.D., Judah Z.

 

 15   Weinberger, M.D., Ph.D., John W. Hirshfeld, M.D.,

 

 16   Thomas B. Ferguson, M.D., Norman S. Kato, M.D.,

 

 17   Brent Blumenstein, Ph.D., Charles Bridges, M.D.

 

 18             For the record, these individuals are

 

 19   special government employees and are consultants to

 

 20   this panel under the Medical Devices Advisory

 

 21   Committee.  They have undergone the customary

 

 22   conflict of interest review and have reviewed the

 

 23   material to be considered at this meeting.

 

 24             In addition, I appoint Cynthia Tracy, M.D.

 

 25   to act as temporary chairperson for the duration of

 

                                                                 9

 

  1   this meeting.

 

  2             Signed by David W. Feigal, Jr., M.D.,

 

  3   M.P.H., Director, Center for Devices and

 

  4   Radiological Health, and dated March the 11th,

 

  5   2004.

 

  6                       Open Public Hearing

 

  7             DR. TRACY:  Thank you.

 

  8             Now we will have the open public hearing.

 

  9   I have a statement to read before we invite our

 

 10   guests up.

 

 11             Both the Food and Drug Administration and

 

 12   the public believe in a transparent process for

 

 13   information gathering and decisionmaking.  To

 

 14   ensure such transparency at the open public hearing

 

 15   session of the advisory committee meeting, FDA

 

 16   believes that it is important to understand the

 

 17   context of an individual's presentation.

 

 18             For this reason, FDA encourages you, the

 

 19   open public hearing speaker, at the beginning of

 

 20   your written or oral statement, to advise the

 

 21   committee of any financial relationship that you

 

 22   may have with the sponsor, its product, and, if

 

 23   known, its direct competitors.

 

 24             For example, this financial information

 

 25   may include the sponsor's payment of your travel,

 

                                                                10

 

  1   lodging, or other expenses in connection with your

 

  2   attendance at the meeting.  Likewise, FDA

 

  3   encourages you at the beginning of your statement

 

  4   to advise the committee if you do not have any such

 

  5   financial relationships.

 

  6             If you choose not to address this issue of

 

  7   financial relationships at the beginning of your

 

  8   statement, it will not preclude you from speaking.

 

  9             Mr. Leo Corbet of Phoenix, Arizona, has

 

 10   requested time to address the panel, I believe.

 

 11             MR. CORBET:  Good morning, ladies and

 

 12   gentlemen.

 

 13             I normally don't get very nervous at this,

 

 14   but you can imagine, as a transplant patient, how

 

 15   nervous people with your credentials make me.  I

 

 16   had a heart transplant in 2001 and prior to that, I

 

 17   was on the machine that is under consideration this

 

 18   morning.

 

 19             I am a lawyer by education, I was a

 

 20   politician by accident, and I ended up in business

 

 21   for the last 20 years and was president of an

 

 22   ethanol company in Nebraska for the time until I

 

 23   got so sick that in 1998, I was told that I had to

 

 24   have a transplant or I would die.

 

 25             Being St. Patrick's Day and I am an

 

                                                                11

 

  1   Irishman with an Italian heart, I will try not to

 

  2   speak with my hands too much.

 

  3             When I went to the hospital in 2001, I had

 

  4   never been in the hospital at all until 1998, and

 

  5   then I had the misfortune of all at once, I came in

 

  6   contact with a lot of medicine, a lot of medical

 

  7   opinions, and finally, they told me I had to get on

 

  8   the heart transplant list, and I went to Arizona,

 

  9   the University of Arizona.

 

 10             I was lucky enough to be put on that list

 

 11   where I was for two years, and I went to the

 

 12   hospital in March, oddly enough, March 10th of

 

 13   2001, where I just thought I would have to go down

 

 14   there and get a checkup and a cleanup, and then I

 

 15   would be leaving, but I didn't leave the hospital

 

 16   for another four months.

 

 17             I got so sick and all my parts were going

 

 18   down the  drain, I guess, so it was determined that

 

 19   I would be put on this CardioWest machine, which I

 

 20   came to know fondly as Big Blue.

 

 21             Big Blue and I were partners for 3 1/2

 

 22   months in the hospital.  It's a 300-pounder

 

 23   machine.  I had some picture, but my wife told me

 

 24   that it wouldn't be within the rules if I started

 

 25   walking around to the jury and started handing them

 

                                                                12

 

  1   out to you, so I left them with her in the

 

  2   audience.

 

  3             I was on that machine for 3 1/2 months

 

  4   until we could find the human or they could find

 

  5   the human heart to give me, and after that, I was

 

  6   in the hospital another two weeks.

 

  7             Since that time, I have retired as an

 

  8   attorney, I have cashed out of my business in

 

  9   Nebraska, and now I am on the board of directors of

 

 10   the Arizona Coalition for transplantation, the New

 

 11   Heart Society.  My wife and I work jointly with the

 

 12   donor network.

 

 13             This machine is so vitally important to

 

 14   people like myself.  If I could have brought these

 

 15   pictures up here, you would see six of us on this

 

 16   same machine, all of us who just would not have

 

 17   made it without this bridge to a time when you can

 

 18   get a heart.

 

 19             According to the donor network, 50 percent

 

 20   of all of the donors that actually are considered

 

 21   aren't qualified.  This is for organs in general or

 

 22   whatever.

 

 23             So, we have a big job there trying to get

 

 24   people to know about transplants.  So, what do you

 

 25   do?  You have to have machines similar to this or

 

                                                                13

 

  1   some other way to keep these people alive, and that

 

  2   is what this particular instrument does, and it

 

  3   does it well.

 

  4             The only negative some people say it has

 

  5   is the sound, the whirring sound that it makes

 

  6   while you are there, and for 24 hours a day I knew

 

  7   I was alive as long as that sound was being made,

 

  8   and most of the guys that were involved did.  I say

 

  9   "guys," not because I am sexist, but because it's

 

 10   the people with large chest cavities were the ones

 

 11   that were prime candidates for this particular

 

 12   machine.

 

 13             When I got out of there, I felt it was

 

 14   incumbent upon me, because of a number of factors,

 

 15   my ability to dump everything else I was doing and

 

 16   get involved in this thing.

 

 17             My wife and I do own shares of stock in

 

 18   this company.  That is not very much of my net

 

 19   worth at all, but I am told I had to disclose that

 

 20   to you today.

 

 21             We believe that this staff that sits

 

 22   behind me, and this machine, are a vital part of

 

 23   the technology that has paved the way to success

 

 24   stories like mine and others, where we could be

 

 25   here today to stand and tell you what we have been

 

                                                                14

 

  1   through.

 

  2             I didn't want to talk very long, so I

 

  3   won't, I will just stop.  If you have any

 

  4   questions, I was there, I can tell you the thing

 

  5   ran 24 hours a day for probably, I think it was

 

  6   about 110 days, and then I was on it, and when it

 

  7   was time to get off, it was kind enough to let me

 

  8   have this human heart, which is working very well.

 

  9             I appreciate the opportunity to be here

 

 10   today. Well, actually, I appreciate the opportunity

 

 11   to be anywhere, and if there any questions, I will

 

 12   be happy to answer them. I know you have a long

 

 13   day, so I am going to cut this and wait if you have

 

 14   any questions.

 

 15             DR. TRACY:  Thank you, Mr. Corbet.  That

 

 16   was a very powerful presentation.  I think if you

 

 17   will be around a little while later, the panel may

 

 18   have some questions for you at a later point.

 

 19             MR. CORBET:  Yes, ma'am.  I have made it

 

 20   67 years, I am not going to stop now.  I will be

 

 21   here.

 

 22             DR. TRACY:  Thank you, sir.

 

 23             Are there any other members of the

 

 24   audience who wish to address the panel today

 

 25   regarding this topic or any other topic?

 

                                                                15

 

  1             [No response.]

 

  2             DR. TRACY:  If not, then, we will close

 

  3   the open public hearing at this point.

 

  4             MS. WOOD:  I would remind the speakers

 

  5   that are about to present to introduce themselves

 

  6   and to state their conflict of interest.

 

  7             DR. TRACY:  Can we ask the sponsor to come

 

  8   forward at this point to discuss the SynCardia

 

  9   Systems, Inc.,  CardioWest Total Artificial Heart.

 

 10          Sponsor Presentation: SynCardia Systems, Inc.

 

 11              P030011: SynCardia Systems CardioWest

 

 12                      Total Artificial Heart

 

 13                           Introduction

 

 14             DR. SLEPIAN:  Madam Chairman,

 

 15   representatives of the FDA, members of the

 

 16   Circulatory Device Panel, good morning.  I am Dr.

 

 17   Marvin Slepian.  I am the CEO of SynCardia, and I

 

 18   have an equity interest in this company.

 

 19             [Slide]

 

 20             Our purpose here today is to seek panel

 

 21   recommendation for the SynCardia CardioWest Total

 

 22   Artificial Heart.

 

 23             [Slide.]

 

 24             Assembled with me here today are the

 

 25   following presenters and representatives.  In

 

                                                                16

 

  1   addition to myself, to my right is Richard Smith,

 

  2   who is the Chief Operating Officer of SynCardia, an

 

  3   engineer, Director of the Artificial Heart Program

 

  4   at the University of Arizona.

 

  5             To his right is Dr. Jack Copeland,

 

  6   Professor of Surgery and Chief of the

 

  7   Cardiothoracic Surgical Section, and Principal

 

  8   Investigator in the trial that we will present here

 

  9   today, also from the University of Arizona.

 

 10             With us is Dr. Walter Pae, behind me,

 

 11   Professor of Surgery in the Section of

 

 12   Cardiothoracic Surgery at Penn State in Hershey.

 

 13             Also, is Mary Dancer, who has been our

 

 14   clinical consultant with Data Management; Dr. Mark

 

 15   Knowles, our statistician; and Sharon Rockwell, our

 

 16   regulatory consultant.

 

 17             [Slide.]

 

 18             We have several additional responders with

 

 19   us:  Dr. Jim Long, who is Associate Professor in

 

 20   Cardiothoracic Surgery, as well as a Professor in

 

 21   Biomedical Engineering, and Director of the

 

 22   Artificial Heart Program at the University of Utah.

 

 23             In the audience is Dr. Walter Dembitsky,

 

 24   Cardiothoracic Surgery, in practice at Sharp

 

 25   Memorial Hospital in San Diego.

 

                                                                17

 

  1             From outside the United States is Dr. Aly

 

  2   El-Banayosy, who is Director of the Artificial

 

  3   Heart Program at the Rhur University Heart Center

 

  4   in Bad Oeyenhausen, Germany, and we have just heard

 

  5   from Leo Corbet.

 

  6             [Slide.]

 

  7             Our presentation will be delivered as

 

  8   follows.  I will provide a very brief introduction.

 

  9   Richard Smith will then discuss the technical

 

 10   aspects of the CardioWest heart. Dr. Copeland will

 

 11   present the IDE clinical trial.  Walter Pae will

 

 12   provide a clinical perspective, and I will make a

 

 13   brief concluding remark.

 

 14             [Slide.]

 

 15             The indication for use that we are seeking

 

 16   is as an in-hospital bridge to transplantation in

 

 17   cardiac transplant candidates at imminent risk of

 

 18   death due to irreversible biventricular heart

 

 19   failure.

 

 20             [Slide.]

 

 21             To provide a brief history, this

 

 22   technology evolved originally from work at Utah

 

 23   with Dr. Jarvik.  There has been significant early

 

 24   experience with the early technological construct

 

 25   of this heart.

 

                                                                18

 

  1             This technology was then transferred to

 

  2   the University of Arizona at the University Medical

 

  3   Center in the early nineties, and it was developed

 

  4   subsequently under the CardioWest aegis.

 

  5             From 1993 to 2002, we have been conducting

 

  6   the IDE trial.  In 1998, we received European CE

 

  7   approval for this, and subsequently have gained

 

  8   experience with this device outside the United

 

  9   States.

 

 10             In 2001, SynCardia was formed as an entity

 

 11   acquiring all this technology under one umbrella

 

 12   with the primary purpose of completing these

 

 13   clinical studies and of submitting this technology

 

 14   for approval to bring it out for clinical use for

 

 15   cardiovascular medicine.

 

 16             I would now like to call on Richard Smith

 

 17   to provide a technical overview of this heart.

 

 18                     CardioWest TAH Overview

 

 19             MR. SMITH:  Thank you, Marvin.

 

 20             [Slide.]

 

 21             My name is Richard Smith.  I am the Chief

 

 22   Operating Officer of SynCardia.  I have an equity

 

 23   interest in this company.  I also am the Director

 

 24   of the Artificial Heart Program in Tucson, Arizona,

 

 25   at the University Medical Center.  I have been in

 

                                                                19

 

  1   that position since 1985, and I have been involved

 

  2   with this project since 1991.

 

  3             [Slide.]

 

  4             The system we are discussing here is the

 

  5   CardioWest TAH System.  There are three main

 

  6   pieces.  The implantable piece is implanted in the

 

  7   chest and exits the chest wall.  There is an

 

  8   external console and there are 7 feet of a

 

  9   driveline that hooks the two together.  I will

 

 10   discuss each one of these.

 

 11             [Slide.]

 

 12             The implantable system--and I have put on

 

 13   the tables two artificial devices, the actual

 

 14   implantable systems themselves--but there are 6

 

 15   components that make up the implantable system.

 

 16             The two inflow connectors are attached to

 

 17   the remanent atria after the natural ventricles and

 

 18   valves are removed.  Those are sown on.  The

 

 19   outflow connectors are sown onto the pulmonary

 

 20   artery and the aorta, and then the actual

 

 21   ventricles are snapped into those four connectors.

 

 22             Once that is accomplished, it actually

 

 23   occupies the space where the diseased heart was

 

 24   removed.  The displacement volume is 400 ml.  The

 

 25   actual weight is 160 grams, which is less than the

 

                                                                20

 

  1   actual heart that was removed in most cases.

 

  2             In that position, the blood flow path is

 

  3   very similar to what the natural heart is.  There

 

  4   is a very short inflow cannula or connector 2

 

  5   centimeters or less, and the actual path of blood

 

  6   between the natural atria to the great vessels is

 

  7   less than 20 cm.

 

  8             The system is designed to be tailored to

 

  9   the individual, so there is an adjustment that can

 

 10   be made after the ventricles are in place, so that

 

 11   we can achieve the best fit, or if there is not the

 

 12   best fit, that we can actually adjust it to make it

 

 13   a better fit.

 

 14             The other unique thing about this system

 

 15   as compared to other systems is there is no

 

 16   surgical pocket, so the only space that is used is

 

 17   the actual chest cavity.

 

 18             [Slide.]

 

 19             Inside the actual ventricles themselves,

 

 20   here is an x-ray to demonstrate that the mechanical

 

 21   valves, Medtronic Hall mechanical valves, the

 

 22   inflows are larger than the outflows, 27 mm versus

 

 23   25.

 

 24             It is hard to see, but there are 4

 

 25   diaphragms that actually separate where the air

 

                                                                21

 

  1   comes in and where the blood comes in, in between

 

  2   there.  There is also a wire-reinforced driveline

 

  3   that actually connects to the outside driveline.

 

  4   So, these three components are part of the

 

  5   reliability aspect of this particular device.

 

  6             [Slide.]

 

  7             The external console is shown here, and

 

  8   incorporates many redundancies in terms of to

 

  9   provide a good reliability.  For example, the

 

 10   primary controller is all that is actually used to

 

 11   operate the device, and there is an actual total

 

 12   backup controller inside the console.

 

 13             In this console is also backup air and

 

 14   power sources, and you can see an alarm panel here.

 

 15   The computer on the top provides information to the

 

 16   user related to ongoing rate of the device when it

 

 17   is set, the actual stroke volume that is caught and

 

 18   calculated, and then from the combination of those

 

 19   two, an ongoing reading of cardiac output.

 

 20             [Slide.]

 

 21             The system is a fairly simple system to

 

 22   use and tries to mimic the Starling Principle of

 

 23   the natural heart. The drive pressures are set in

 

 24   order to achieve what we call "full eject" in both

 

 25   ventricles, so you are overcoming the pressure in

 

                                                                22

 

  1   the pulmonary system and the systemic system on

 

  2   each beat.  Then, the rate is set in order to

 

  3   achieve what we call "partial fill," so we do not

 

  4   want the whole ventricle to fill.

 

  5             It can fill to 70 ml, so we try to set it

 

  6   up and the computer tells us about 55.  In that

 

  7   mode, we don't have to make many adjustments after

 

  8   that, and as the patient then requires more cardiac

 

  9   output as he returns more blood to the heart, more

 

 10   gets filled, and since at each beat, it is ejected,

 

 11   the cardiac output goes up automatically and

 

 12   adjusts both up and down without any adjustments

 

 13   from the user.

 

 14             In achieving this in the chest, and with a

 

 15   short pass, we can accomplish 9 1/2 liters of

 

 16   cardiac output. Typically, we are running at 7 to 8

 

 17   1/2 in most of these patients.

 

 18             [Slide.]

 

 19             This is a schematic of the system

 

 20   simplified for your review.  If you can pay

 

 21   attention to the far right, these are 3 air sources

 

 22   we plug into the wall, and there are 2 internal air

 

 23   sources.

 

 24             All 3 of these are hooked up at any one

 

 25   time, and as the patient is moved around to

 

                                                                23

 

  1   exercise or to walk, the wall is detached and then

 

  2   these other systems take over.  I will go through a

 

  3   couple of cycles.

 

  4             This is what the controller operates.

 

  5   There is valves in there that basically turn from

 

  6   this high compressed air during a systolic cycle

 

  7   and pumps air and pushes the diaphragm up.  So,

 

  8   that is half the cycle.

 

  9             On the second side of the cycle--and it is

 

 10   all timed--the valve turns to room air and the air

 

 11   is then exhausted to the outside.

 

 12             It is a very simple system and these

 

 13   sensors that are inside the console monitor this

 

 14   non-invasively, so on the systole, the pressure

 

 15   sensors provide this wave form.

 

 16             This wave form tells that we have

 

 17   accomplished the pressure gradients against the

 

 18   pulmonary and systemic system, and this little

 

 19   notch here says that we are fully ejecting, so the

 

 20   user has a feedback that basically, you have

 

 21   accomplished rule number 1, which is full eject.

 

 22               On the diastole side, we generate this,

 

 23   and this actually calculates the air coming back,

 

 24   and from that we get a stroke volume measurement

 

 25   estimate, so by multiplying the rate that we set on

 

                                                                24

 

  1   the console and the stroke volume we get indirectly

 

  2   at the cardiac output fed back at all times.

 

  3             There are a series of alarms that are

 

  4   constantly monitoring this system also.

 

  5             [Slide.]

 

  6             Now, to differentiate this from other

 

  7   devices, in the very simplistic of viewpoints, you

 

  8   are taking out the natural ventricles and valves.

 

  9             The VADs are hooked when they are hooked

 

 10   to the natural heart, so issues, such as

 

 11   arrhythmias, for example, if you put an LVAD in,

 

 12   then arrhythmias can actually affect the filling of

 

 13   that LVAD, are alleviated obviously because you

 

 14   don't have the ventricles in place.

 

 15             Any type of right ventricular dysfunction

 

 16   by any cause is alleviated because you have an

 

 17   artificial right ventricle in place.

 

 18              Aortic valve issues, whether it is a

 

 19   prosthetic aortic valve or maybe the aortic valve

 

 20   has insufficiency, affect how VADs are filled in

 

 21   that type of setting and again are not an issue

 

 22   when you put a TAH in.

 

 23             Things like thrombus in the left

 

 24   ventricle, septal defects, all these things are

 

 25   technical issues that if you have a VAD in place,

 

                                                                25

 

  1   they affect the performance of those VADs and are

 

  2   not issues when the TAH is implanted.

 

  3             [Slide.]

 

  4             Here is again a simplistic viewpoint of

 

  5   the circulatory system.  Here I have shown that the

 

  6   right ventricle, artificial ventricle is in place,

 

  7   and the left artificial ventricle is in place.

 

  8              When this occurs, immediately we see in

 

  9   the OR that the central venous pressure is reduced

 

 10   because you now have a good pump in place that

 

 11   basically pushes blood across the lungs, and

 

 12   therefore guarantees in most cases 6 or 7 liters of

 

 13   flow to the left side, which then basically is

 

 14   delivered to the organs.

 

 15             So, there are two effects here that

 

 16   accomplishes. You get an increase in your blood

 

 17   pressure because of your increase in cardiac output

 

 18   on this side, and you get a decrease in the central

 

 19   venous pressure on this side, so the actual

 

 20   differential of pressure across these organs is

 

 21   maximized.

 

 22             So, you get an increase on this side,

 

 23   decrease on this side, and that allows us to have

 

 24   the best situation for potential organ recovery in

 

 25   a lot of these patients that begin with organ

 

                                                                26

 

  1   dysfunction.

 

  2             [Slide.]

 

  3             During the trial, we have worked with the

 

  4   FDA in terms of in-vitro testing.  The original

 

  5   setup was to run 8 systems, 4 of them to 6 months,

 

  6   and the other 4 to failure.  That has occurred.

 

  7             In addition, we have taken 3 systems that

 

  8   were actually placed on the shelf, a sterilization

 

  9   period, for 3 years, and then we put those on the

 

 10   mock circulations also, and those have been run to

 

 11   failure.  To date, we have not had any failures,

 

 12   and these systems have run for over 3 1/2 years.

 

 13             [Slide.]

 

 14             So, in conclusion, the CardioWest TAH

 

 15   provides biventricular cardiac flow up to 9.5

 

 16   liters/minute.  The component and design features

 

 17   provide very good safety and facilitate the implant

 

 18   procedure.

 

 19             There are no reliance or limitations

 

 20   related to any native heart dysfunction, and the

 

 21   CardioWest TAH function is simple and reliable.

 

 22             Thank you.

 

 23             At this time, I would like to invite Dr.

 

 24   Copeland to talk about this study.

 

 25                        IDE Clinical Trial

 

                                                                27

 

  1             DR. COPELAND:  Thank you, Rich.

 

  2             My name is Jack Copeland.  I am the

 

  3   principal investigator in this study.  I am Chief

 

  4   of the Section of Cardiovascular Surgery at the

 

  5   University of Arizona and Professor of

 

  6   Cardiothoracic Surgery.

 

  7             I do have an equity interest in SynCardia.

 

  8             We are going to present the results of

 

  9   this SynCardia CardioWest Total Artificial Heart

 

 10   IDE Study.

 

 11             [Slide.]

 

 12             First of all, I would like to mention the

 

 13   clinical need for the total artificial heart.  This

 

 14   is simply in patients who have biventricular

 

 15   failure, who have either impending or ongoing

 

 16   end-organ failure and decompensation, a very sick

 

 17   patient group who can be transplanted if a heart is

 

 18   available.

 

 19             As we all know, hearts aren't so readily

 

 20   available as they should be or we would like them

 

 21   to be, and most often these patients die unless

 

 22   there is some intervention, the intervention being

 

 23   the total artificial heart which provides

 

 24   hemodynamic stabilization, organ recovery, and

 

 25   allows bridge to transplantation.  So, this is used

 

                                                                28

 

  1   when medical therapy fails.

 

  2             [Slide.]

 

  3             If we look at the most recent data from

 

  4   the UNOS waiting list, we see that in 2002, 14.4

 

  5   deaths per patient year occurred on the waiting

 

  6   list, so there is a real need that has been

 

  7   documented by the United Network for Organ Sharing,

 

  8   our national donor group.

 

  9             [Slide.]

 

 10             The aims of the study are to look at the

 

 11   efficacy and safety of cardiac replacement with the

 

 12   CardioWest Total Artificial Heart in bridge to

 

 13   transplantation.

 

 14             [Slide.]

 

 15             Our hypothesis is this - that patients

 

 16   with irreversible biventricular failure could be

 

 17   saved utilizing the CardioWest Total Artificial

 

 18   Heart as a bridge to transplantation.

 

 19             [Slide.]

 

 20             The study design is a prospective,

 

 21   non-randomized, multi-center study of critically

 

 22   ill patients with irreversible end-stage congestive

 

 23   heart failure, all of whom are in a New York Heart

 

 24   Association Class IV, who are also

 

 25   transplant-eligible.

 

                                                                29

 

  1             Historical controls are patients who were

 

  2   put into the study who met identical entry criteria

 

  3   to the study patients.

 

  4             [Slide.]

 

  5             The study endpoint variables are shown on

 

  6   this slide.  First, there is a multifactorial

 

  7   endpoint that is called "primary efficacy endpoint"

 

  8   or a treatment success.

 

  9             Treatment success means that at 30 days

 

 10   post-transplant the patient is alive, New York

 

 11   Heart Association Class I or II, is ambulatory, is

 

 12   not on a ventilator, nor is he on dialysis.

 

 13             So, the idea is that the patient is up and

 

 14   around and reasonably healthy hopefully.

 

 15             The secondary efficacy endpoints are:

 

 16   survival, hemodynamic recovery, end-organ recovery,

 

 17   and ambulation.

 

 18             Finally, we looked at safety parameters as

 

 19   reflected by adverse event analysis.

 

 20             [Slide.]

 

 21             The key issues for us in the design of the

 

 22   study were as follows:  To define a patient

 

 23   population that needs biventricular bridge to

 

 24   transplantation, and to define the natural history

 

 25   of untreated patients using historical controls.

 

                                                                30

 

  1             [Slide.]

 

  2             In this study, there were 5 centers and 12

 

  3   surgeons including the University Medical Center in

 

  4   Tucson, myself and Francisco Arabia; Loyola

 

  5   University Medical Center, Chicago, Drs. Foy,

 

  6   Sullivan, and Montoya; LDS Hospital in Salt Lake,

 

  7   Dr. Long and Doty--we are hoping that Dr. Long will

 

  8   arrive soon, he is scheduled to be in--St. Luke's

 

  9   Medical Center, Milwaukee, Drs. Tector, Schmahl,

 

 10   and Kress; and University of Pittsburgh Medical

 

 11   Center, Drs.  Griffith and Kormos.

 

 12             [Slide.]

 

 13             The study patients in this study are

 

 14   divided up as you see on this slide.  There were

 

 15   130 patients, 35 controls, 95 patients received

 

 16   implants, 81 of these were core patients or

 

 17   patients who met all of the entry criteria, and

 

 18   there were 14 out-of-protocol patients who did not

 

 19   meet entry criteria.

 

 20             [Slide.]

 

 21             The study inclusion criteria are shown on

 

 22   this slide.  First of all, the patient had to be

 

 23   eligible for cardiac transplantation.  He had to

 

 24   meet the criteria for transplant at the local

 

 25   center.

 

                                                                31

 

  1             He had to be New York Heart Association

 

  2   Class IV. He had to be large enough to have a

 

  3   reasonable chance of fit or an excellent chance of

 

  4   fit of the total artificial heart. By this, in the

 

  5   inclusion criteria, it means a body surface area of

 

  6   1.7 square meters or an AP diameter from the

 

  7   posterior sternum to the anterior spine of 10 cm at

 

  8   T10 on CT scan.  The patients had to have

 

  9   hemodynamic insufficiency.

 

 10             [Slide.]

 

 11             By "hemodynamic insufficiency," we mean

 

 12   either one of two things, either Criteria A or

 

 13   Criteria B.  These criteria are multifaceted.  In

 

 14   Criteria A, it is a cardiac index of less than or

 

 15   equal to 2 L/min/M2 with one of the following:  low

 

 16   arterial pressure or high central venous pressure

 

 17   greater than or equal to 18 mm of mercury.

 

 18             Criteria B, it was two of the following

 

 19   list:  Basically, high dose inotropic support

 

 20   including such drugs as dopamine, dobutamine,

 

 21   amrinone, and others at maximal or near maximal

 

 22   levels.  Also, intra-aortic balloon pumping and

 

 23   being on cardiopulmonary bypass, so two of those on

 

 24   that list.

 

 25             [Slide.]

 

                                                                32

 

  1             The exclusion criteria for this study are

 

  2   shown on this slide.  Use of any ventricular assist

 

  3   device, pulmonary vascular resistance of greater

 

  4   than or equal to 8 Wood units, dialysis in the

 

  5   previous 7 days, serum creatinine of greater than

 

  6   or equal to 5 mg/dl, total bilirubin or greater

 

  7   than or equal to 5 mg/dl, and cytotoxic antibody

 

  8   levels of greater than or equal to 10 percent.

 

  9             [Slide.]

 

 10              Now, how did we decide to choose a total

 

 11   artificial heart rather than an LVAD in this study?

 

 12   This slide shows perhaps better than any other part

 

 13   of this presentation the criteria for using a total

 

 14   heart as opposed to an LVAD.

 

 15             We start with 81 core patients, 15 of

 

 16   those patients were on heart/lung machines, CPS

 

 17   pumps or ECMO at the time they were implanted with

 

 18   a total artificial heart, and had global cardiac

 

 19   dysfunction.

 

 20             Fifty-one of the patients had evidence of

 

 21   right ventricular failure as evidenced by a high

 

 22   central venous pressure greater than 18 mm of

 

 23   mercury.

 

 24             Eleven of the patients had right

 

 25   ventricular ejection fractions of less than 20

 

                                                                33

 

  1   percent.

 

  2             This leaves 4 miscellaneous patients who

 

  3   all were intra-aortic balloon pump support, maximal

 

  4   inotropes, and had failing hemodynamics, and, in

 

  5   addition, 2 had incessant ventricular tachycardia,

 

  6   1 had a mechanical aortic valve, which is a

 

  7   contraindication to LVAD implant, and 1 had a right

 

  8   ventricular injury at sternotomy and was therefore

 

  9   in the embarrassing situation of having a divided

 

 10   right ventricle.

 

 11             [Slide.]

 

 12             This shows a representative heart for the

 

 13   type of heart that is removed from these patients

 

 14   when a total artificial heart is implanted, quite

 

 15   similar to what we see when we remove the heart

 

 16   when we do the transplant.

 

 17             It is very thinned out both on the left

 

 18   and the right sides and has this baggy shape, it

 

 19   has no form, it doesn't stand up and look like a

 

 20   normal heart, and you see, interestingly enough,

 

 21   some left ventricular thrombus that was neither

 

 22   detected by transthoracic nor transesophageal

 

 23   echocardiography.

 

 24             [Slide.]

 

 25             Now, a few words and a few slides about

 

                                                                34

 

  1   the control group.  First of all, the

 

  2   identification of the control group.  There were

 

  3   some historical controls that were found from the

 

  4   period of 1991 through 1993, 22 patients.

 

  5             There were some prospective controls that

 

  6   were added during the study, and there were some

 

  7   controls that were found by looking back in 2002

 

  8   through the UNOS Class I patients from all 5

 

  9   centers, and we found 10 additional patients there,

 

 10   so there are 35 patients in the control group.  All

 

 11   of these controls had to meet the study inclusion

 

 12   and exclusion criteria, and that is how they were

 

 13   selected by retrospective analysis.

 

 14             [Slide.]

 

 15             Now, a few words about the comparison of

 

 16   the controls with the core patients.

 

 17             First of all, there were lots of things

 

 18   that were similar statistically, demographics,

 

 19   things like age, ethnic group, and gender, and risk

 

 20   factors, such as diabetes, hypertension, cardiac

 

 21   arrest, acute myocardial infarction, previous PTCA,

 

 22   the presence of an automatic implanted

 

 23   defibrillator, pacemaker, being on the ventilator,

 

 24   and being obtunded and drowsy.  All of these were

 

 25   essentially the same between the two groups.

 

                                                                35

 

  1             [Slide.]

 

  2             On the other hand, there were some

 

  3   non-comparable baseline characteristics of the

 

  4   controls versus the core patients.  It turned out

 

  5   that the control group had more ischemic patients,

 

  6   a higher incidence of smoking history.

 

  7             More of the patients were anticoagulated,

 

  8   but then as we looked at the laboratory data, there

 

  9   was no difference in the coagulation tests between

 

 10   the two groups.  In other words, the core group had

 

 11   elevated PT and INR presumably from a liver

 

 12   synthetic dysfunction because they weren't on

 

 13   anticoagulation as much as the control group.

 

 14             Prior cardiac surgery, there was more

 

 15   history of that in the control group, as you might

 

 16   expect in an ischemic population, and a higher

 

 17   incidence of use of intra-  aortic balloon pump.

 

 18             On the other hand, 18.5 percent of the

 

 19   core group were on cardiopulmonary bypass when they

 

 20   were implanted with the total artificial heart, and

 

 21   none of the controls were.

 

 22             [Slide.]

 

 23             On the other hand, there were some

 

 24   non-comparable baseline characteristics, a lot of

 

 25   hemodynamic factors that were identical.  The ones

 

                                                                36

 

  1   that weren't are shown here.

 

  2             The mean arterial pressure was not

 

  3   significantly different, but the systolic arterial

 

  4   pressure was slightly lower, and this might reflect

 

  5   the use of intra-aortic balloon pumping.

 

  6             The pulmonary arterial pressure was higher

 

  7   in the core patients or the implant patients than

 

  8   it was in the controls, and the central venous

 

  9   pressure was also higher in the core or implant

 

 10   patients than it was in the controls.

 

 11             [Slide.]

 

 12             If we look at the hierarchy of support at

 

 13   baseline, the characteristics that relate to amount

 

 14   of drug use, intra-aortic balloon use, and use of

 

 15   cardiopulmonary bypass, and you look at the white

 

 16   line, above the white line is greater than or equal

 

 17   to 3 drug support.

 

 18             Sixty-six percent of the core patients

 

 19   fell into that category, the green patients, as

 

 20   opposed to 80 percent of the control patients were

 

 21   below the line.  They were on less than 3 drugs

 

 22   support.

 

 23             [Slide.]

 

 24             We looked at a whole variety of laboratory

 

 25   values and nearly all the chemistry, hematology,

 

                                                                37

 

  1   and blood gases were comparable.  One interesting

 

  2   value that wasn't, was a total bilirubin, which in

 

  3   the core patients was 2, and in the controls was

 

  4   1.3.

 

  5             [Slide.]

 

  6             So, what is the usefulness of the control

 

  7   group?  We are in complete agreement that the

 

  8   control group is not for formal statistical

 

  9   comparison.  What we found in looking at all the

 

 10   different characteristics that we looked at, at

 

 11   baseline, comparing control and core patients, was

 

 12   that 53 of 65 characteristics matched, and 12 did

 

 13   not match.

 

 14             There was more ischemic disease, more

 

 15   previous coronary bypass, and more characteristics

 

 16   of an ischemic population in the controls than in

 

 17   the core patients.

 

 18             [Slide.]

 

 19             So, it is our hope that in this study, the

 

 20   controls will give an approximation of the natural

 

 21   history of patients meeting entry criteria who do

 

 22   not have mechanical support.  That is what we

 

 23   believe to be the value of the controls in this

 

 24   study.

 

 25              We also want to point out that in a way,

 

                                                                38

 

  1   each one of the patients that is implanted in this

 

  2   study serves as his own control, because he starts

 

  3   very, very sick, and then has an implant and has a

 

  4   history following that of hemodynamic and general

 

  5   recovery.

 

  6             Finally, there are comparisons from the

 

  7   published VAD studies that provide a perspective on

 

  8   our core patient results, and I think these are

 

  9   very important.  They will be mentioned in detail

 

 10   by Dr. Pae's presentation, but we may refer to them

 

 11   from time to time in this presentation, as well.

 

 12             [Slide.]

 

 13             So, let's go on then to the study results

 

 14   with respect to efficacy.

 

 15             [Slide.]

 

 16             The primary endpoint in this study was

 

 17   called treatment success.  As you remember, it is a

 

 18   multifactorial endpoint - patient alive, functional

 

 19   Class I or II, ambulatory, not on ventilator, not

 

 20   on dialysis.  Sixty-nine percent of patients in the

 

 21   core group met that endpoint, 37 percent in the

 

 22   controls.

 

 23             [Slide.]

 

 24             Looking at survival endpoints, perhaps the

 

 25   most important is survival to transplant.  After

 

                                                                39

 

  1   all, that is the point of putting in a bridge to

 

  2   transplant device.  Seventy-nine percent of the

 

  3   total artificial heart patients survived to

 

  4   transplant.  The 95 percent confidence intervals

 

  5   are shown here.  The lower limit of our confidence

 

  6   interval was 68.5.

 

  7             Survival to 30 days post-transplant was

 

  8   71.6 percent.  Survival at 1 year from the study,

 

  9   looking at all patients, was 70.4 percent.

 

 10   Survival of the transplanted patients after

 

 11   transplantation at 1 year was 85.9 percent. This

 

 12   does compare favorably with published survival data

 

 13   for bridge to transplantation.

 

 14             [Slide.]

 

 15             Now, what happened to these patients that

 

 16   were implanted?  This shows the core patients.  The

 

 17   mean time on the device was 79.1 days or the time

 

 18   to transplant.  The median was 47, and the longest

 

 19   patient on device was 414 days.  This compares to

 

 20   the controls who had a mean time to transplant of

 

 21   8.5 days and a median of 6.

 

 22             There were 6,411 study days for the core

 

 23   patients and 299 for the controls.

 

 24             [Slide.]

 

 25             These two Kaplan-Meier curves show

 

                                                                40

 

  1   survival to transplant or death, and this really

 

  2   gives the history of what happened to these two

 

  3   groups of patients.                The core

 

  4   patients for the most part survived and got

 

  5   transplanted.  The control patients either got

 

  6   transplanted or were dead within 6 weeks.  The

 

  7   control patients, on the other hand, have a history

 

  8   that goes out 50 weeks.

 

  9             [Slide.]

 

 10             Now, the overall duration of survival,

 

 11   this looks at from the beginning to as far as we

 

 12   have followed these patients, and it goes out, for

 

 13   the control group, to 12 years since some of the

 

 14   controls were enlisted from a time prior to the

 

 15   start of the study, and it goes out to about 9

 

 16   years for the study group, and you can see the big

 

 17   falloff in survival early on, and then parallel

 

 18   curves after transplantation, indicating a

 

 19   uniformity of the result in transplantation, but a

 

 20   high mortality rate early on without the device.

 

 21             [Slide.]

 

 22             These curves show survival from

 

 23   transplantation of both the control and core

 

 24   groups, and we have added the red dots, and the red

 

 25   dots give you an idea of survival that has been

 

                                                                41

 

  1   reported from the UNOS registry for the entire

 

  2   population of patients in UNOS for over a period of

 

  3   years.

 

  4             You can see the core group is right on

 

  5   after transplantation with the UNOS group, and

 

  6   these very sick control patients who were crashing

 

  7   and in trouble did not have as good an early result

 

  8   after the transplant, but then have a parallel

 

  9   result sometime later, starting at about 1 year.

 

 10             [Slide.]

 

 11             Now, if we look a little more closely at

 

 12   the UNOS versus core patients, so these are the

 

 13   implanted patients, 64 implanted patients were

 

 14   transplanted, and we are comparing with 4,481 UNOS

 

 15   patients.  Survival at 1 year, 85.9 for the core

 

 16   patients, 84.7 for the UNOS, or the United Network

 

 17   for Organ Sharing, and at 5 years it was 63.8 for

 

 18   the core patients and 69.8 for UNOS, quite

 

 19   comparable.

 

 20             [Slide.]

 

 21             Now, getting on past the survival results

 

 22   to secondary efficacy endpoints, I want to first

 

 23   discuss hemodynamic recovery, and this slide

 

 24   summarizes the key components of that.

 

 25             First of all, immediately after

 

                                                                42

 

  1   implantation, as Mr. Smith mentioned, the cardiac

 

  2   output goes up with this device.  It went from 1.9

 

  3   to 2.9 L/min/M2, and this was a significant and a

 

  4   sustained change.

 

  5             Systolic arterial pressure rose from 92.8

 

  6   to 121.  Central venous pressure fell immediately

 

  7   from 19.7 to 13.6, and organ perfusion pressure, or

 

  8   the difference between mean arterial pressure and

 

  9   the central venous pressure, or that force that

 

 10   perfuses the end organs, went from 48.6 to 67.5.

 

 11             [Slide.]

 

 12             Here, you see a curve for the cardiac

 

 13   index over 70 days.  You see the immediate rise and

 

 14   sustained level at about 3 L/min/M2.

 

 15             [Slide.]

 

 16             This gives you an idea of what that kind

 

 17   of perfusion and drop in central venous pressure,

 

 18   increased organ perfusion pressure does to

 

 19   end-organ function.  Here is the creatinine

 

 20   starting at about 1.7, an elevated level rising to

 

 21   2.5 and then falling to normal within about 3 to 4

 

 22   weeks.

 

 23             [Slide.]

 

 24             The hepatic function.  There is the total

 

 25   bilirubin starting at 2 mg/dl, rising to about 3.7

 

                                                                43

 

  1   to 4, and then falling to normal within 3 weeks.

 

  2             [Slide.]

 

  3             This slide shows ambulation, which was one

 

  4   of the endpoints, functional recovery.  On the left

 

  5   you see the percent of patients that were able to

 

  6   get out of bed.  You see by 14 days, 80 percent of

 

  7   patients were essentially ambulatory or out of bed.

 

  8             On the right, you see the percent of

 

  9   patients who are able to walk greater than 100

 

 10   feet.  At 14 days, it was around 60 percent of this

 

 11   core total artificial heart population.

 

 12             [Slide.]

 

 13             We will go on now from efficacy to safety

 

 14   and review adverse events.

 

 15             The first slide is one that we are not

 

 16   going to  go into in detail.  It covers the 81 core

 

 17   patients prior to transplant.  This data is in your

 

 18   packet in case you want to refer to it in more

 

 19   detail, but I show it just to show you the number

 

 20   of adverse events that were looked for.

 

 21             All of this was part of our study from the

 

 22   very beginning.  We did not find anything that fell

 

 23   outside of these categories of adverse events, nor

 

 24   did we find anything that was outside of what would

 

 25   have been expected with LVADs or BiVADs.

 

                                                                44

 

  1             [Slide.]

 

  2             Let me explain.  First, let's look at

 

  3   infection, the adverse event of infection in the 81

 

  4   core patients during the implant period, the time

 

  5   that the device was implanted.

 

  6             There were a total of 125 events, 48

 

  7   respiratory, 29 genitourinary, some

 

  8   gastrointestinal, some driveline infections, and

 

  9   these were superficial skin type infections. There

 

 10   were 8 blood infections.

 

 11             There were 6 mediastinal infections, all

 

 12   but 1 of these were found incidentally at the time

 

 13   of transplantation, to they were not clinical

 

 14   mediastinitis, the 5, 1 was, and I will explain

 

 15   that in a moment, and there were 6 line infections

 

 16   or intravenous line, central line, that type of

 

 17   thing, infections.

 

 18             [Slide.]

 

 19             Looking more carefully, then, at the

 

 20   clinically adverse event infection, and separating

 

 21   out what were the clinically significant

 

 22   infections, we see that 5 of these infections

 

 23   delayed transplantation.

 

 24             There were no instances of ascending

 

 25   driveline infections.  The types of infections are

 

                                                                45

 

  1   shown here, driveline, blood, respiratory, and

 

  2   mediastinal.

 

  3             There were 7 infections that contributed

 

  4   to death, 5 respiratory or pneumonia, 1 line

 

  5   infection or sepsis, and  1 mediastinal infection.

 

  6   There was 1 infection that caused death, pneumonia.

 

  7   Neither the ones that contributed to death or

 

  8   caused death were related to the device.

 

  9             [Slide.]

 

 10             Let's go on to the adverse event of

 

 11   bleeding.  This is for 81 core patients during the

 

 12   implant period.  We had various definitions of

 

 13   bleeding, so we looked very broadly at bleeding,

 

 14   trying to pick up every event that would reflect

 

 15   abnormal bleeding, abnormal loss of blood, or

 

 16   abnormal replacement of blood.

 

 17             First of all, they are the takebacks.

 

 18   Takeback means take back to the operating room for

 

 19   bleeding or cardiac tamponade, early after the

 

 20   implant, usually.  All but one of those takebacks

 

 21   were at less than 21 days after implantation.

 

 22             There was post-implant bleeding.  This is

 

 23   greater than 48 hours after transplant and the

 

 24   patient needed 3 units of blood within a 24-hour

 

 25   period.  There were 18 such events.  Surgical

 

                                                                46

 

  1   bleeding meant that the patient bled greater than 8

 

  2   units or had greater than 8 units replaced while he

 

  3   was in the operating room having the implant.

 

  4             The abdominal bleeding refers to 1 case of

 

  5   abdominal bleeding that required a takeback

 

  6   operation.

 

  7             There were 2 deaths from bleeding in our

 

  8   experience.

 

  9             [Slide.]

 

 10             We go on then to neurologic adverse events

 

 11   in 81 core patients during the implant period.

 

 12   There were 10 stroke, 4 TIAs, 5 episodes of

 

 13   encephalopathy, 1 transient loss of consciousness,

 

 14   1 metabolic encephalopathy, and 5 seizures.

 

 15             A stroke in this study was defined as

 

 16   neurologic dysfunction of greater than or equal to

 

 17   24 hours duration. There were 25 total neurologic

 

 18   events in this study of the 81 core patients during

 

 19   implant.

 

 20             [Slide.]

 

 21             Now, let's expand our patient population

 

 22   to the 95 patients that were implanted, so that we

 

 23   include every stroke that occurred in this study.

 

 24   This is the core group plus the out-of-protocol

 

 25   group.

 

                                                                47

 

  1             For this total group of patients during

 

  2   the implant period, there were 11 strokes in 10

 

  3   patients or a 10.5 percent incidence.

 

  4             In 6 of the strokes, there was no--that

 

  5   should be "no," I am sorry, it says "mo"--but there

 

  6   was no residual after 48 hours, and there was no

 

  7   delay of transplant, so these were what you might

 

  8   call very minor neurologic abnormalities.

 

  9             There were 4 that had residual, but did

 

 10   not delay transplant.  There was 1 dense hemiplegia

 

 11   only.  The other strokes were of much less

 

 12   magnitude.  This did delay transplant, but later

 

 13   the patient did get transplanted and met treatment

 

 14   success criteria.

 

 15             The events per patient month or the

 

 16   strokes per patient month in this study were 0.05

 

 17   linearized rate.

 

 18             [Slide.]

 

 19             I want to go on and mention device

 

 20   malfunctions because device malfunctions I think

 

 21   give us, looking at this gives us confidence of the

 

 22   reliability of this device.  This is in all 95

 

 23   patients.  This is a greater than 19-year

 

 24   experience on the device, 19 patient years.

 

 25              There were 11 driveline kinks.  That

 

                                                                48

 

  1   means the alarm went off on the device, indicating

 

  2   a driveline kink. These were very transitory,

 

  3   lasting seconds or less. Occasionally, they were

 

  4   associated with a loss of consciousness, but not

 

  5   often, and they caused no effect on the outcome.

 

  6             There were 3 patients that had 5 driveline

 

  7   leaks.  This was due to the method of coupling the

 

  8   lines that come out of the patient with the

 

  9   external plastic lines that connect to the device.

 

 10             This was redesigned and following the

 

 11   redesign and the retraining of all centers, there

 

 12   were no more events in 33 patients.  We discovered

 

 13   this when we were about 50 some-odd patients into

 

 14   the experience, and it only occurred in 3 patients.

 

 15             There were a couple of miscellaneous

 

 16   events.  Once, there was loss of consciousness for

 

 17   just a second while an air tank was being changed.

 

 18   There was no effect on the outcome.  There was 1

 

 19   controller that kept showing a low alarm.  There

 

 20   was never a low output in this patient, the

 

 21   controller was changed, and the patient was fine,

 

 22   there was no effect.

 

 23             Then, there was the 1 major event that

 

 24   occurred of device malfunction that caused major

 

 25   difficulty, and this was a tear in one of the

 

                                                                49

 

  1   diaphragms.  The diaphragm is a 4-layered structure

 

  2   of polyurethane with a little bit of graphite

 

  3   between each of the layers.

 

  4             On day 90, this patient started showing

 

  5   signs of low output and eventually we diagnosed

 

  6   that he did have a tear in this diaphragm.  The

 

  7   reason for this is unknown.

 

  8             We checked the entire history of this type

 

  9   of device, which includes well over 500 implants

 

 10   and probably 50 patient years of experience.  It is

 

 11   the only time this ever happened in this device, it

 

 12   was not catastrophic.

 

 13             The patient finally died on day 124, the

 

 14   event happened on day 90.  He refused implantation

 

 15   of a second total artificial heart and support was

 

 16   eventually withdrawn in his case.

 

 17             [Slide.]

 

 18             I want to go on now to fit complications

 

 19   since patient size is a requirement for entry into

 

 20   the study and fitting a device into the chest is an

 

 21   important concept that has to be appreciated by

 

 22   every surgeon that implants and every cardiologist

 

 23   that refers patients for implantation of this

 

 24   device.

 

 25             We looked at this in 95 implanted

 

                                                                50

 

  1   patients, in other words, all the implanted

 

  2   patients, and there were 5 events.  Two of these

 

  3   occurred in the operating room and were detected

 

  4   immediately when the chest was closed.  It was

 

  5   reopened and the device was repositioned.  There

 

  6   was no effect on outcome.

 

  7             One was detected after the patient was

 

  8   returned to the intensive care unit.  He was

 

  9   returned and the device was repositioned, there was

 

 10   no effect on outcome.

 

 11             A fourth event was compression of the

 

 12   pulmonary veins that was corrected by repositioning

 

 13   the device, but the patient continued to be in

 

 14   pulmonary edema and eventually died of other

 

 15   complications, but the pulmonary edema contributed.

 

 16             Finally, there was a patient who came to

 

 17   the operation in severe pulmonary edema with stiff

 

 18   lungs.  The device never really fit well.  He was

 

 19   left with an open sternum and closed with a PTFE

 

 20   graft, and this was a contributing cause of death.

 

 21             [Slide.]

 

 22             Going on then to causes of death.  All of

 

 23   the causes of death in the 81 core patients are

 

 24   listed here. There were 17 deaths, for a percentage

 

 25   of 21 percent.

 

                                                                51

 

  1             Seven patients died of multiple organ

 

  2   failure, 4 patients died of procedural/technical

 

  3   causes, and those causes are listed on the bottom

 

  4   of the slide.

 

  5             Two died of hypercoagulable states after

 

  6   transplantation.  They were being given aprotinin

 

  7   and, in addition, they received an activated factor

 

  8   VII complex called FEIBA, and they had

 

  9   intravascular coagulation.

 

 10             Two of the patients in this group died

 

 11   from wedging of a central line in the tricuspid

 

 12   valve of the total artificial heart.

 

 13             Two contraindications for the use of this

 

 14   device are the use of FEIBA and aprotinin at the

 

 15   time of transplantation should not be done, and

 

 16   central lines should not be placed in the right

 

 17   atrium.

 

 18             Once we experienced this, we, of course,

 

 19   made it known to all of the implanting centers, and

 

 20   there are strict criteria with respect to where

 

 21   central line tips should be, and they are located

 

 22   radiographically, and this is confirmed in patients

 

 23   at this time and from the time of this experience

 

 24   forward.

 

 25             To carry on with the causes of death,

 

                                                                52

 

  1   bleeding in 2 patients, sepsis in 2,

 

  2   pre-implantation cardiac arrest in 1, and pulmonary

 

  3   edema in 1.

 

  4             [Slide.]

 

  5             In summary, we had immediate hemodynamic

 

  6   recovery with using this device.  We had end-organ

 

  7   recovery, and patients got out of bed and were

 

  8   walking usually within a week, most of them within

 

  9   2 weeks.

 

 10             [Slide.]

 

 11             We had a treatment success of 69.1

 

 12   percent.  This compares favorably with the bar that

 

 13   has been set in the FDA presentation of 65 to 70

 

 14   percent for survival to transplant, and it includes

 

 15   a number of other criteria.

 

 16             [Slide.]

 

 17             Seventy-nine percent of the patients

 

 18   survived to transplant, and the lower limit of our

 

 19   confidence interval again exceeds that bar set in

 

 20   the FDA report.

 

 21             Survival at 30 days post-transplant was

 

 22   71, and 1 year survival from entry into study was

 

 23   70.4.  Survival from the time of transplant to 1

 

 24   year was almost 86 percent.

 

 25             [Slide.]

 

                                                                53

 

  1             These results were generalizable or were

 

  2   found across the 3 major centers in the study.  You

 

  3   can see survival to transplant at the LDS hospital

 

  4   was 75 percent, Loyola 84, and University Medical

 

  5   Center 79 percent.

 

  6             Treatment success was found in 63 percent

 

  7   of LDS, 76.9 of Loyola patients, and 69 percent of

 

  8   UMC patients.

 

  9             [Slide.]

 

 10             In conclusion, we feel that this is a

 

 11   reliable device.  There was 1 serious device

 

 12   malfunction in 19 patient years of support.

 

 13             The performance was excellent, giving a

 

 14   high cardiac output, a low venous pressure, and

 

 15   good organ perfusion.  There were significant

 

 16   adverse events including bleeding, stroke,

 

 17   infection, and other events, but we feel these are

 

 18   acceptable in the face of the efficacy of the

 

 19   device.

 

 20             The efficacy basically is that 79 percent

 

 21   of patients were maintained to transplantation.  We

 

 22   therefore feel that the CardioWest Total Artificial

 

 23   Heart offers significant benefit at reasonable

 

 24   risk.

 

 25             [Slide.]

 

                                                                54

 

  1             I would like to conclude by showing you a

 

  2   slide of another one of our patients.  This one is

 

  3   in the middle.  He had a transplant two years prior

 

  4   to November of 2003 when he rode in a bicycle race

 

  5   for 70 kilometers.  He had the CardioWest Total

 

  6   Artificial Heart for 4 months prior to his

 

  7   transplant.

 

  8             Thank you very much.  I would like to go

 

  9   on now to the presentation of Dr. Walter Pae.

 

 10                      Clinical Perspective]

 

 11             DR. PAE:  Good morning.  My name is Walter

 

 12   Pae.

 

 13             [Slide.]

 

 14             I am a Professor of Cardiothoracic Surgery

 

 15   at the Pennsylvania State University.  I have been

 

 16   involved with circulatory support and the

 

 17   development, as well as the clinical application of

 

 18   devices, throughout most of my adult life which

 

 19   started nearly 30 years ago now.

 

 20             [Slide.]

 

 21             What I would like to do is to give a

 

 22   little bit of a clinical perspective on the need

 

 23   for circulatory support  in bridge to transplant.

 

 24             DR. TRACY:  Sir, could you please state

 

 25   your financial interest.

 

                                                                55

 

  1             DR. PAE:  I don't have any financial

 

  2   interests. Thank you.  I have no conflicts.

 

  3             To go on a little bit about the need for

 

  4   total artificial hearts.  There is no doubt that

 

  5   there is a donor shortage, and this continues to

 

  6   exist and will continue to exist into the future.

 

  7   For example, in the year 2002, there were about

 

  8   3,800 transplants performed, that are listed, and

 

  9   only about 2111 patients transplanted in the United

 

 10   States.

 

 11             Now, a donor shortage is not only the

 

 12   absolute number of donors, but also timing.

 

 13   Patients get sick and hearts aren't available.

 

 14             The mortality rate is substantial among

 

 15   patients that are awaiting cardiac transplantation,

 

 16   and certainly those individuals with severe Class V

 

 17   heart failure have an exceedingly high risk of

 

 18   death.

 

 19             I think it is important to note that these

 

 20   patients are the individuals that are now way past

 

 21   beta blockers and ACE inhibitors, and biventricular

 

 22   pacers.  They are individuals who have failed all

 

 23   of these modalities and need heart replacement

 

 24   therapy.

 

 25             I think that a very good argument has been

 

                                                                56

 

  1   made that there is a need for practical device

 

  2   therapy for not only the prevention of death, but

 

  3   medical stabilization. Many times one can argue

 

  4   that many of these patients that are moribund are

 

  5   better off with a device prior to transplantation

 

  6   than they are going on to immediate transplant if a

 

  7   heart was available.

 

  8             [Slide.]

 

  9             This has been alluded to in the past and I

 

 10   think we can skip over it in the interests of time.

 

 11             [Slide.]

 

 12             There is a few caveats about right

 

 13   ventricular failure.  The incidence of right

 

 14   ventricular failure in individuals who are on left

 

 15   ventricular assist device depends on the

 

 16   definition.

 

 17             It has been reported in the literature,

 

 18   peer-reviewed literature, in anywhere from 11 to 26

 

 19   or perhaps as high as 30 percent of the patients on

 

 20   ventricular assist devices.

 

 21             About one-third to one-half of those

 

 22   patients have required with right ventricular

 

 23   failure fail inotropic support and go on to an

 

 24   additional right ventricular assist device or

 

 25   sequential device therapy, ending up with

 

                                                                57

 

  1   biventricular support.

 

  2             I think that in the clinical arena of

 

  3   those individuals who do a lot of this and report

 

  4   it in the peer-reviewed literature, it is much more

 

  5   likely to occur in the individual who is sicker, in

 

  6   the individual who undergoes an emergent therapy

 

  7   versus what we call elective urgent implantation.

 

  8             Certainly, this affects outcomes.  Right

 

  9   ventricular failure, if you look at the outcomes,

 

 10   it has about one-half of the successful bridge rate

 

 11   as compared to those individuals who do not exhibit

 

 12   right ventricular failure.

 

 13             I think the last thing is that our

 

 14   prospective ability to predict this is inaccurate

 

 15   in many situations, but I think there are certain

 

 16   clinical scenarios that exist that make clinical

 

 17   decisionmaking more often right than wrong in this

 

 18   particular instance.

 

 19             [Slide.]

 

 20             So, we are definitely left with a need for

 

 21   biventricular support.  How do we provide this at

 

 22   the present time?

 

 23             Well, we can use hybrid systems where we

 

 24   use a left ventricular assist device of one

 

 25   manufacturer's or another rigged with a second

 

                                                                58

 

  1   device from another manufacturer, and if you are to

 

  2   go back and look through the literature a little

 

  3   bit, one of my fellows, is now a partner of mine,

 

  4   published or was a co-author on a paper of mine,

 

  5   that indicated throughout the registry experience,

 

  6   when it existed through the ISHLT and the AIO, that

 

  7   only about one-third of those patients who got

 

  8   "hybrid" systems went on to actually get orthotopic

 

  9   cardiac transplantation.

 

 10             Now, your paracorporeal systems that are

 

 11   utilized for biventricular support are flow

 

 12   limited, and they are flow limited by design.  They

 

 13   have inlet and outlet cannulae that are long and it

 

 14   takes a certain amount of time to fill and empty

 

 15   these devices.

 

 16             They also require a competent aortic

 

 17   valve, and they are clearly limited by the

 

 18   liability of native heart pathology in certain

 

 19   instances.

 

 20             The total artificial heart, on the other

 

 21   hand, provides immediate high flow.  It is not

 

 22   limited at all by the native heart pathology since

 

 23   the native heart is extirpated just like it will be

 

 24   in orthotopic cardiac transplantation after

 

 25   successful bridging.

 

                                                                59

 

  1             [Slide.]

 

  2             So, I think there is a clinical need for

 

  3   the total artificial heart both in biventricular

 

  4   failure, but in a number of instances where the

 

  5   native heart presents a liability, where there are

 

  6   large ventricular thrombi, and in our own personal

 

  7   published experience in individuals with large

 

  8   anterior wall myocardial infarctions and

 

  9   cardiogenic shock, the incidence of stroke in those

 

 10   individuals bridged with univentricular support was

 

 11   about 85 percent.

 

 12             Individual with ventricular septal

 

 13   defects, post-myocardial infarction, massive

 

 14   ventricular ruptures post-myocardial infarction,

 

 15   refractory arrhythmias, prosthetic aortic valves

 

 16   and incompetent aortic valves, particularly the

 

 17   failed cardiac transplant.

 

 18             There are a number of instances of adult

 

 19   congenital heart disease which makes standard

 

 20   ventricular assist devices very difficult.

 

 21             We have also actually dealt with patients

 

 22   with cardiac malignancies where this is a useful

 

 23   technique, and I believe Dr. Copeland actually

 

 24   alluded to complex reoperative situations, and some

 

 25   of these are just simply unavoidable tactical

 

                                                                60

 

  1   misadventures or acts of the devil.

 

  2             [Slide.]

 

  3             There is a very large body of

 

  4   peer-reviewed literature, which is labeled here for

 

  5   reference, and was selected for review.  Each one

 

  6   of these is labeled with a number, so that you can

 

  7   go through and follow these as I speak about this.

 

  8             [Slide.]

 

  9             Now, survival to transplant.  In these

 

 10   peer-reviewed studies, basically, ranges between 51

 

 11   and 71 percent.  This compares clinically very

 

 12   favorably with what has been presented with the

 

 13   SynCardia device.

 

 14             [Slide.]

 

 15             Interestingly, out of the registry paper,

 

 16   published in 1995, if you look at the numbers,

 

 17   quite comparable despite the fact that this was

 

 18   obviously a voluntary registry and there are

 

 19   problems with that.

 

 20             [Slide.]

 

 21             When you begin to look at the literature

 

 22   review in terms of adverse events, there are

 

 23   clearly a number of very important issues, things

 

 24   to keep in mind.  The definitions are very, very

 

 25   varied.  There is no set definition to define

 

                                                                61

 

  1   adverse events.  They are broad definitions with

 

  2   subcategories.

 

  3             Much of this is device-related data

 

  4   reporting.  The rates of these events, the numbers

 

  5   of patients with each event is not always provided,

 

  6   and I think that everyone needs to understand from

 

  7   a statistical standpoint that these are not always

 

  8   constant hazard functions, and even when the event

 

  9   rates are linearized, it doesn't necessarily tell

 

 10   the entire story.

 

 11             The time frames that devices are utilized,

 

 12   for example, a device being used for 10 days may

 

 13   not have any events during that 10 days, and if it

 

 14   carries out to 100 or 200 or 300, we begin to see

 

 15   the pile-up of adverse events, so it is many times

 

 16   very difficult to compare.

 

 17             Obviously, all of the registry data

 

 18   previously was voluntary and that makes statistics

 

 19   difficult at best.

 

 20             [Slide.]

 

 21             At any rate, recognizing the limitations

 

 22   of the peer-reviewed literature, if you begin to

 

 23   look at adverse events, infection, for example,

 

 24   tends to range between 2 and 55 percent, and I

 

 25   think that tells us that there is a very broad

 

                                                                62

 

  1   range of infectious problems that occur that may be

 

  2   device related or it may have to do with

 

  3   definitions.

 

  4             If one looks at bleeding, again, a very

 

  5   wide range of 31 to 51 percent.  Strokes, TIAs,

 

  6   neurologic events, again zero to 59 percent.

 

  7             If you eyeball these sorts of things, I

 

  8   think that you will see that the total artificial

 

  9   heart question today falls well within those

 

 10   ranges.

 

 11             [Slide.]

 

 12             The same thing for death during implant.

 

 13   If one looks at multi-organ failure, for example,

 

 14   amongst the various devices, biventricular devices

 

 15   or univentricular devices, it tends to range

 

 16   somewhere between 8 and 29 percent, comparing

 

 17   favorably with 9 percent.

 

 18             Please keep in mind that many, many of the

 

 19   patients that die with multi-organ failure from

 

 20   univentricular support are really dying of right

 

 21   ventricular failure, and that is a manifestation as

 

 22   such.

 

 23             Cerebral events, 8 to 21 percent, again

 

 24   compares favorably.  Sepsis, zero to 14 percent,

 

 25   again comparing favorably.  Bleeding, again, a

 

                                                                63

 

  1   very, very wide range, but again comparing

 

  2   favorably.  The same as right ventricular  failure

 

  3   and air embolism.

 

  4             [Slide.]

 

  5             When we go down to the miscellaneous

 

  6   events, and I am sure there is hundreds of those,

 

  7   once again compares favorably.

 

  8             [Slide.]

 

  9             Lastly, when we begin to look at device

 

 10   malfunction, again from the published literature,

 

 11   device failure, meaning exactly what it means in

 

 12   terms of the device not working, we have ranges

 

 13   between 1 and 20 percent.

 

 14             These are usually referred to as critical

 

 15   device failures versus device malfunctions, which

 

 16   can be things that are very, very simple like

 

 17   driveline kinks, or even things as external

 

 18   controller failures that are easily replaced, but

 

 19   they go anywhere between 4 and 100 percent.

 

 20             When you look at these again clinically,

 

 21   the device in question falls well within those

 

 22   ranges.

 

 23             [Slide.]

 

 24             So, I think from a clinician's

 

 25   perspective, this particular device of CardioWest

 

                                                                64

 

  1   Total Artificial Heart has successfully salvaged

 

  2   patients that have severe biventricular failure.

 

  3             The survival to transplant with this total

 

  4   artificial heart is clinically comparable to the

 

  5   left ventricular assist devices and the

 

  6   biventricular assist devices without limitations of

 

  7   the diseased native heart.

 

  8             The safety of the devices appears

 

  9   clinically similar to the available devices, and

 

 10   the clinical benefit of this seems to outweigh the

 

 11   associated risk.  I think there is definitely a

 

 12   clinical need for the total artificial heart.  It

 

 13   is a useful device that fills a therapeutic gap

 

 14   that will clearly extend our ability to treat

 

 15   cardiovascular disease.

 

 16             Thank you.

 

 17                            Conclusion

 

 18             [Slide.]

 

 19             DR. SLEPIAN:  Therefore, in conclusion, a

 

 20   need exists for safe and effective therapies which

 

 21   can save lives for these debilitated patients and

 

 22   stabilize them in imminent danger of dying from

 

 23   biventricular failure.

 

 24             The total artificial heart provides

 

 25   hemodynamic normalization which then leads to

 

                                                                65

 

  1   end-organ recovery.

 

  2             The  Total Artificial Heart System was

 

  3   demonstrated to be safe and reliable, bridging

 

  4   patients to transplantation.

 

  5             The study was a success with demonstrated

 

  6   efficacy in all endpoints.

 

  7                The CardioWest TAH is the first total

 

  8   artificial heart to demonstrate life-saving

 

  9   results, with benefits outweighing risks, in such a

 

 10   sick group of patients.

 

 11             [Slide.]

 

 12             The indication for use therefore is as an

 

 13   in-hospital bridge to transplantation in

 

 14   transplant-eligible candidates at imminent risk of

 

 15   death due to irreversible biventricular failure.

 

 16             [Slide.]

 

 17             TAH candidates would include those that

 

 18   are transplant-eligible, New York Heart Class IV or

 

 19   the new    AHA/ACC Class D heart failure, with

 

 20   hemodynamic insufficiency that is refractory to

 

 21   medical therapy and would be best served with this

 

 22   type of device with situations including RV

 

 23   failure, the presence of a clot in the LV,

 

 24   refractory arrhythmias, prosthetic valve, holes in

 

 25   the heart, stone heart, or a rejected transplant or

 

                                                                66

 

  1   a failed transplant, or unresuscitatable patients

 

  2   following cardiac arrest, failure to wean from

 

  3   bypass with biventricular injury, or due to

 

  4   surgical technical issues with LV situations with

 

  5   massive MI.

 

  6             [Slide.]

 

  7             The contraindications for this system are

 

  8   those patients that are ineligible or not cardiac

 

  9   transplant candidates that are small with body

 

 10   surface areas less than 1.7.

 

 11             [Slide.]

 

 12             Our proposed training will be as follows.

 

 13   It will include a didactic hands-on animal

 

 14   implantation and on-site proctoring component.

 

 15   Didactic training will be provided by experienced

 

 16   surgeons and others on the technical side that have

 

 17   been involved in this study.

 

 18             There will be direct instruction.  A

 

 19   training manual has been developed, study of which

 

 20   will be required. Complete literature review and

 

 21   being familiar with the published literature will

 

 22   be emphasized, and a video of implantation is also

 

 23   part of the training.

 

 24             Surgeons then move on to hands-on

 

 25   experience, as well as technical people with the

 

                                                                67

 

  1   TAH, as well as with the console and drivelines.

 

  2   Then, from there, animal implantation will be

 

  3   required with a minimal implant of two separate

 

  4   acute studies in a pig model.

 

  5             Finally, on-site proctoring will be

 

  6   provided by experienced surgeons and other

 

  7   technical personnel that are familiar with the

 

  8   device.

 

  9             [Slide.]

 

 10             In addition, in discussion with the FDA,

 

 11   we have proposed post-market surveillance which

 

 12   will involve follow-up on enrolled study patients.

 

 13   Fifty additional U.S. patients will be in this to

 

 14   demonstrate generalizability. Less than 10 percent

 

 15   will be from any one center.

 

 16             Adverse events will be captured during the

 

 17   implant period, as well as survival to transplant

 

 18   followed, and one  year follow-up.

 

 19             [Slide.]

 

 20             This study provides reasonable assurance

 

 21   of the safety and effectiveness of the CardioWest

 

 22   TAH for the proposed indication for use.

 

 23             Thank you very much.

 

 24                      Questions and Answers

 

 25             DR. TRACY:  Thank you.

 

                                                                68

 

  1             At this time I would like to ask the panel

 

  2   members if they have any brief clarifying questions

 

  3   for the sponsor.

 

  4             Dr. Aziz.

 

  5             DR. AZIZ:  This is probably for Dr.

 

  6   Copeland.  The  out-of-use of the device, the 14

 

  7   patients that you had who couldn't come off

 

  8   cardiopulmonary bypass, were they mainly in one

 

  9   center or were they in different centers?

 

 10             DR. COPELAND:  That was not restricted to

 

 11   one center.  There were several episodes in each of

 

 12   the three major centers in the study.

 

 13             DR. AZIZ:  These are patients who the

 

 14   operator felt that even in a period of protracted

 

 15   rest like that, peripheral bypass for 24 hours

 

 16   wouldn't allow recovery of the donor heart?

 

 17             DR. COPELAND:  I can only speak for my own

 

 18   experience with respect to the exact situations,

 

 19   and these were situations where there was

 

 20   irrevocable evidence that the heart was permanently

 

 21   and irreversibly damaged either by the patient's

 

 22   disease or the patient's disease with a

 

 23   combination, with the operative intervention that

 

 24   was being done.

 

 25             MR. SMITH:  Could I comment further on

 

                                                                69

 

  1   that?  That group was also people that were put on

 

  2   portable ECMO machines and then transferred to the

 

  3   operating room, so they weren't just patients that

 

  4   were actually in cardiac surgery at the time.

 

  5             DR. BRIDGES:  This question is for Dr.

 

  6   Copeland primarily, but the others could respond,

 

  7   as well.

 

  8             One question is in your presentation, it

 

  9   seemed that you indicated that two of the patients

 

 10   with coagulopathies, that is, those that had

 

 11   received aprotinin and the other factor 7 agent--

 

 12             DR. COPELAND:  FEIBA.

 

 13             DR. BRIDGES:  --had received that during

 

 14   the transplantation?  That is what I was a little

 

 15   confused about, why would they not be included in

 

 16   the transplant group if they, in fact, got to

 

 17   transplantation.

 

 18             DR. COPELAND:  The reason they were

 

 19   included is because that is the way the study was

 

 20   designed.  In other words, the patient is in the

 

 21   implant to transplant period until he leaves the

 

 22   operating room after being transplanted, so all of

 

 23   the adverse events from the transplant operation,

 

 24   the removal of the total artificial heart and the

 

 25   transplanted heart are captured in that period of

 

                                                                70

 

  1   time.

 

  2             Then, the post-transplant to 30-day

 

  3   complications are captured after the patient is

 

  4   post-operatively, so that is why we see that.

 

  5             DR. BRIDGES:  Another question was where

 

  6   were the takebacks bleeding from, was there a

 

  7   pattern to where 25 percent of the patients were

 

  8   taken back for bleeding within the first 48 hours?

 

  9             DR. COPELAND:  Generally, this is, first

 

 10   of all, I would like to point out that the

 

 11   prothrombin time and the INR in these patients on

 

 12   the average was in the anticoagulated range, so

 

 13   these are patients who were very sick, who have, in

 

 14   general, a high incidence of liver synthesis

 

 15   problems, and therefore have coagulopathies.

 

 16             Also, about 30 percent of the patients

 

 17   were reoperative patients, so most of the bleeding

 

 18   was not from anything related to the device, but

 

 19   just simply patient-related soft tissue, diffuse

 

 20   oozing sternal wires, the usual kind of things that

 

 21   we see in routine cardiac surgery.

 

 22             DR. BRIDGES:  One last question.  All of

 

 23   the controls met the inclusion/exclusion criteria,

 

 24   but were all of the patients that met the

 

 25   inclusion/exclusion criteria included in the

 

                                                                71

 

  1   control group?

 

  2             DR. COPELAND:  All of the charts that were

 

  3   reviewed and found patients that fulfilled these

 

  4   criteria were used.  There were something like 635

 

  5   charts reviewed. These were all about UNOS Class I

 

  6   patients.  Over half of these were not used because

 

  7   they weren't sick enough to meet the inclusion

 

  8   criteria.

 

  9             About a fourth of them were excluded

 

 10   because they received a VAD, and another sizable

 

 11   number were excluded because they had some medical

 

 12   contraindication to transplantation.  So, that is

 

 13   the way we found the control patients.

 

 14             DR. BRIDGES:  It just seemed that there

 

 15   weren't as many controls enrolled after 1993 as I

 

 16   would have expected if all of the patients who met

 

 17   the inclusion/exclusion criteria were included in

 

 18   the control group.

 

 19             DR. COPELAND:  Well, when we looked back

 

 20   in 2002 at each of the five centers, at all of

 

 21   their patients who had been listed UNOS Class I,

 

 22   that is all we could find, so our assumption was

 

 23   that in that interval, there was an increased use

 

 24   of other devices that would make them ineligible

 

 25   for this study as control patients.

 

                                                                72

 

  1             DR. BRIDGES:  Thank you.

 

  2             DR. TRACY:  Dr. Yancy.

 

  3             DR. YANCY:  Several points of

 

  4   clarification, and I appreciate the caution that

 

  5   was raised and the answer regarding those 14

 

  6   out-of-protocol patients, because I, too, had some

 

  7   questions about that.

 

  8             One question regarding outcomes.  There

 

  9   were 51 patients in Dr. Copeland's presentation who

 

 10   had a CVP greater than 18 mm of mercury,

 

 11   purportedly the main indication for bypass support

 

 12   in this setting of RV dysfunction.

 

 13             I am curious as to the outcomes in that

 

 14   specific group that had a clear hemodynamic

 

 15   construct for RV dysfunction.

 

 16             DR. COPELAND:  I think it is dangerous to

 

 17   look at one number and assume that that was the

 

 18   reason, the sole reason for putting in a total

 

 19   artificial heart.

 

 20             I mean these are patients who are

 

 21   critically ill, who are declining rapidly, who are

 

 22   on lots of inotropes, who have end organ

 

 23   dysfunction, and in putting together that

 

 24   particular slide for this presentation, we were

 

 25   looking for, within the group of patients, things

 

                                                                73

 

  1   that stood out as indicators of right ventricular

 

  2   failure, and that is how we came upon that.

 

  3             So, all of these patients were severe

 

  4   Class IV, they were all on inotropes.  You know,

 

  5   they might have been on intra-aortic balloon pumps,

 

  6   and so forth, and so on, as well, and they had

 

  7   severe abnormalities in their baseline

 

  8   hemodynamics.

 

  9             Then, you say okay, are we going to put in

 

 10   an LVAD or a BiVAD or a total artificial heart, and

 

 11   basically, the way we decided was looking at these

 

 12   criteria plus those things that you saw recorded on

 

 13   that slide including a high CVP, generally in face

 

 14   of a normal pulmonary artery pressure, which is in

 

 15   an indicator of severe right ventricular

 

 16   dysfunction.

 

 17             DR. YANCY:  You haven't qualified it as

 

 18   such, and I assume that that group did as well as

 

 19   the overall trial result?

 

 20             DR. COPELAND:  Yes.

 

 21             DR. YANCY:  A second question is about

 

 22   adverse events.  There is a slide that Dr. Pae

 

 23   referred to that suggested numbers that were not

 

 24   consistent with data that Dr. Copeland presented

 

 25   and this may just need a point of clarification.

 

                                                                74

 

  1             The slide in reference is Slide 81, and it

 

  2   shows an infection rate for the SynCardia of 28.3

 

  3   percent, which is in variance to the Slide 56, that

 

  4   shows an infection rate of 71.6.  I am assuming

 

  5   that this is a question of definition.

 

  6             I would just like to know the variance.

 

  7   The same can be said for bleeding, 42 percent

 

  8   versus 37 percent, and for stroke/TIA 24.7 for

 

  9   neurologic versus 12.6 for stroke/TIA, but by

 

 10   calculating it, it would still be 17 percent for

 

 11   the events that occurred in the trial.

 

 12             So, are these points of definition, what

 

 13   are the differences here in those two slides?

 

 14             DR. COPELAND:  Yes, what was done--can we

 

 15   have the slide up, please, P81--what was done in

 

 16   this to try to make these adverse event numbers

 

 17   comparable was to redefine infection according to

 

 18   the way it was defined by the majority of articles

 

 19   in the literature.  Bleeding and stroke were

 

 20   treated in the same way.

 

 21             So, the numbers that you see for our core

 

 22   patients or for our combined core and

 

 23   out-of-protocol patients are not going to be the

 

 24   same as what you see here.

 

 25             DR. YANCY:  The third question, if I may,

 

                                                                75

 

  1   had to do with the fact that there is a statement

 

  2   in our packet that says the device has been

 

  3   approved for use in several outside of U.S.

 

  4   countries.

 

  5             I am wondering if there can be a

 

  6   statement, too, about the general experience of the

 

  7   device, I think it's France, Canada, and Germany,

 

  8   is there any statement that can be made about the

 

  9   experience with the device in the out-of-U.S.

 

 10   countries.

 

 11             DR. COPELAND:  First of all, I can

 

 12   summarize by saying the survival to transplant in

 

 13   the out-of-U.S. experience has been about 60

 

 14   percent, that the involvement of the centers

 

 15   utilizing this device in Western Europe primarily

 

 16   have been people who were not constrained by a

 

 17   protocol, namely, a protocol that is this type and

 

 18   that limits the use to transplant candidates, for

 

 19   instance, which is a pretty strict criterion, and

 

 20   also that limits it to use in people who have not

 

 21   had a VAD in place.

 

 22             There have been many uses of this device,

 

 23   for instance, in patients who have failed either

 

 24   BiVAD or univentricular support and then gone on to

 

 25   total artificial heart.

 

                                                                76

 

  1             Now, if it's permissible, I could have Dr.

 

  2   Aly Banayosy from Bad Oeyenhausen comment on that

 

  3   question, as well.

 

  4             DR. TRACY:  Actually, there is going to be

 

  5   a lot more time for discussion after lunch.  Maybe

 

  6   we can hold off and just any other brief clarifying

 

  7   questions.

 

  8             Dr. Krucoff.

 

  9             DR. KRUCOFF:  Dr. Copeland, I just want to

 

 10   make sure that I heard what you said about the

 

 11   identification of the control population, because

 

 12   at least what I heard you say was not in our pack.

 

 13             The charts that were reviewed, is the

 

 14   implication they were triggered by UNOS I

 

 15   categories?  How did you identify the 600-some

 

 16   charts that you reviewed?

 

 17             DR. COPELAND:  The 600-some refers to sort

 

 18   of the total review that was done, and the

 

 19   definition of UNOS Class I, as you may know,

 

 20   changed over the course of time, but, in general,

 

 21   that is a true statement that the patients did meet

 

 22   UNOS Class I, which means on inotropic support

 

 23   either out of hospital or in hospital, perhaps on

 

 24   hemodynamic monitoring in hospital.

 

 25             That would cover both what we call UNOS

 

                                                                77

 

  1   Class IB, which is the out of hospital, and UNOS

 

  2   Class IA, which is the in hospital with hemodynamic

 

  3   monitoring and inotropes.

 

  4             DR. KRUCOFF:  So, the 600 were identified

 

  5   presumably out of thousands of charts in the

 

  6   medical center using what?

 

  7             DR. COPELAND:  That is correct.

 

  8             DR. KRUCOFF:  How did you identify those

 

  9   600?

 

 10             DR. COPELAND:  By going directly to the

 

 11   UNOS Class I patients.

 

 12             DR. KRUCOFF:  So, that is what triggered

 

 13   that list.

 

 14             DR. COPELAND:  Right, exactly.

 

 15             DR. KRUCOFF:  Within that, if I heard you

 

 16   correctly, of the patients who would have fit the

 

 17   criteria, for a control patient, about half of

 

 18   those patients had had VADs, so you excluded them,

 

 19   is that what you said?

 

 20             DR. COPELAND:  About 130 some-odd had had

 

 21   VADs, about nearly 300 were not sick enough by the

 

 22   hemodynamic insufficiency criterion.

 

 23             DR. KRUCOFF:  But of the ones who were

 

 24   sick enough, about half of them had VADs?

 

 25             DR. COPELAND:  That is correct.

 

                                                                78

 

  1             DR. KRUCOFF:  Have you looked at those

 

  2   data?

 

  3             DR. COPELAND:  We actually have, and I am

 

  4   not--do you remember that data?  We can--

 

  5             DR. KRUCOFF:  If you have it this

 

  6   afternoon when we have time, I think that would be

 

  7   very relevant.

 

  8             DR. COPELAND:  We can get that for you,

 

  9   yes.  I recollect that--let's wait until this

 

 10   afternoon, we can get you the right numbers, but

 

 11   the experience just to the best of my memory was

 

 12   about a 40 to 50 percent survival at transplant in

 

 13   that group.

 

 14             DR. TRACY:  Dr. Hirshfeld.

 

 15             DR. HIRSHFELD:  Dr. Copeland, I would like

 

 16   to try to get a better handle on what the

 

 17   thromboembolic potential of this device is.  You

 

 18   reported a total of 11 strokes.  That sounds as

 

 19   though they were likely to be thromboembolic

 

 20   events.

 

 21             There is also a report of 9 peripheral

 

 22   thromboembolic events, which weren't really

 

 23   discussed this morning, and I wonder--I assume,

 

 24   first of all, those are different events than the

 

 25   strokes, it's not just another rubric for the

 

                                                                79

 

  1   strokes.

 

  2             DR. COPELAND:  No, it's not.

 

  3             DR. HIRSHFELD:  I wonder if you could tell

 

  4   us what the nature of these peripheral

 

  5   thromboembolic events were, what the consequences

 

  6   were, and whether you think it is appropriate to

 

  7   combine the frank strokes and the peripheral

 

  8   thromboembolic events as a measure of the overall

 

  9   thromboembolic potential for the device.

 

 10             DR. COPELAND:  Most of those events were

 

 11   Amaurosis fugax.  They were transient retinal

 

 12   problems that lasted seconds to minutes and

 

 13   disappeared.  There were a few that were emboli to

 

 14   either the spleen or the kidneys, and we have that

 

 15   data, and we have reviewed that data, and to the

 

 16   best of my memory, that is a summary of the

 

 17   findings with the peripheral embolization.

 

 18             DR. HIRSHFELD:  The second question is it

 

 19   seems to me that there was a high frequency of

 

 20   perioperative bleeding at the time of the

 

 21   transplantation, as well as at the time of the

 

 22   initial implant of the device, and it seems to me

 

 23   plausible that this is due to the fact that these

 

 24   patients were on full warfarin anticoagulation and

 

 25   fairly aggressive antiplatelet therapy at the time

 

                                                                80

 

  1   that they underwent transplantation.

 

  2             I wondered whether your experience is that

 

  3   this is a real phenomenon that doing the transplant

 

  4   is technically more challenging because of the

 

  5   anticoagulation status of the patient.

 

  6             DR. COPELAND:  Yes, my impression, and,

 

  7   you know, I am coming out of an experience of over

 

  8   200 pulsatile pump bridge to transplant, so over

 

  9   two-thirds of our experiences with LVADs and

 

 10   BiVADs, and I will have to say that my experience

 

 11   is that they all bleed at the time of transplant

 

 12   and usually very badly.

 

 13             I don't really see that the total

 

 14   artificial heart is any different in any major way

 

 15   from those devices.

 

 16             DR. HIRSHFELD:  One last point of

 

 17   clarification, if I might.

 

 18             DR. TRACY:  I would like to remind the

 

 19   panel, though, that there is plenty of time for

 

 20   discussion all afternoon, so unless these are

 

 21   really to clarify something that you need to

 

 22   discuss things further this afternoon, I would like

 

 23   to hold off, so if anybody has just a brief

 

 24   question.

 

 25             Dr. Maisel.

 

                                                                81

 

  1             DR. MAISEL:  Perhaps you could clarify a

 

  2   little bit about who the patients were who were

 

  3   being enrolled in the study.  I certainly recognize

 

  4   that they are critically ill and have horrible

 

  5   heart failure, but, for example, about 50 percent

 

  6   of the patients were ischemic patients.

 

  7             Were they chronic ischemic patients, how

 

  8   long from their initial heart failure diagnosis to

 

  9   the time that they ended up getting their device,

 

 10   were they presenting with acute myocardial

 

 11   infarction, were they post-CABG and didn't do well,

 

 12   who were the patients that were enrolled in this

 

 13   study?

 

 14             DR. COPELAND:  You are asking with respect

 

 15   to the ischemic patients?

 

 16             DR. MAISEL:  With respect to everyone, but

 

 17   particularly with respect to the ischemics who

 

 18   seemed to have a worse outcome.

 

 19             DR. COPELAND:  I am just pausing for a

 

 20   moment.  I am trying to get you as quantitative an

 

 21   answer as I can.

 

 22             Let's go to B3D, please.

 

 23             [Slide.]

 

 24             This is a slide on the etiology of disease

 

 25   of these study patients.  First of all, with

 

                                                                82

 

  1   respect to acute myocardial infarction, I can tell

 

  2   you that 5 of the core patients had just

 

  3   experienced acute myocardial infarctions. None of

 

  4   the control patients had just experienced an acute

 

  5   myocardial infarction, and that distant myocardial

 

  6   infarction incidence was about the same in both

 

  7   groups.  It was around 25 to 30 percent for both.

 

  8             Unfortunately, the printing on this slide

 

  9   doesn't show up as well as we would like, but it is

 

 10   idiopathic, ischemic, viral, acute myocardial

 

 11   infarction, acute rejection, failure to wean from

 

 12   bypass, and other are the major categories, and we

 

 13   can provide you with this slide if you would like

 

 14   to examine it in more detail, but in terms of is

 

 15   your question directed towards--

 

 16             DR. MAISEL:  That answers my question,

 

 17   thank you.

 

 18             DR. COPELAND:  Okay, thanks.

 

 19             DR. TRACY:  I think in the interest of

 

 20   time here, we will pause at this point for a brief

 

 21   break.  Let's try to reconvene at 11 o'clock,

 

 22   please.

 

 23             [Break.]

 

 24             DR. TRACY:  At this point, we are going to

 

 25   move on and I will ask the FDA to begin their

 

                                                                83

 

  1   presentation.

 

  2                         FDA Presentation

 

  3                             Summary

 

  4             MR. CHEN:  Good morning, Madam Panel

 

  5   Chair, panel members.  I would like to welcome you

 

  6   to the Circulatory Support Devices Panel today.

 

  7             [Slide.]

 

  8             The PMA application to be discussed today

 

  9   is PMA No. 030011 for the SynCardia Systems

 

 10   CardioWest Total Artificial Heart.

 

 11             [Slide.]

 

 12             My name is Eric Chen and I was the lead

 

 13   reviewer for this application.

 

 14             [Slide.]

 

 15             I would like to present a brief overview

 

 16   of the presentation  today.  I will be presenting a

 

 17   history of the clinical study along with the

 

 18   preclinical evaluation of the device.  The

 

 19   statistical evaluation will be presented by  Dr.

 

 20   Lilly Yue, and Dr. Julie Swain will be addressing

 

 21   the clinical evaluation.

 

 22             At the end of the FDA presentation today

 

 23   will be the Questions to the Panel, which we will

 

 24   discuss in the afternoon session.

 

 25             [Slide.]

 

                                                                84

 

  1             Due to the complexity of the device, a

 

  2   wide variety of specialists were called upon to

 

  3   review this application.  The FDA Review Team

 

  4   consisted of engineers, statisticians, and

 

  5   clinicians, and their names have been listed here

 

  6   for their recognition.

 

  7             [Slide.]

 

  8             The following is a picture of the

 

  9   CardioWest Total Artificial Heart System.  I will

 

 10   not go into too much detail with the device's

 

 11   function since the sponsor has already presented a

 

 12   full explanation of the device, however, the

 

 13   device, as you can see, consists of two implantable

 

 14   artificial ventricles, two pneumatic drivelines

 

 15   that exit the chest and attach to an external

 

 16   console.

 

 17             The external console contains a computer

 

 18   that monitors the device's function.  The external

 

 19   console also contains a controller and a backup

 

 20   controller along with an alarm panel.  As well, the

 

 21   external console has backup compressors and a

 

 22   backup power supply in the event of an emergency.

 

 23             [Slide.]

 

 24             Shown in this cartoon is the anatomical

 

 25   fit of the device inside a patient.  As you can see

 

                                                                85

 

  1   from the picture, the native ventricles of the

 

  2   patient have been resected in order to implant the

 

  3   device.

 

  4             Once again, you can see the pictures of

 

  5   the left and the right ventricles along with the

 

  6   pneumatic drivelines and driveline exit sites.

 

  7             [Slide.]

 

  8             The proposed indication for use for the

 

  9   device is as a bridge to transplant in cardiac

 

 10   transplant-eligible candidates at risk of imminent

 

 11   death from non-reversible biventricular failure.

 

 12   The device is solely intended to be used inside the

 

 13   hospital.

 

 14             [Slide.]

 

 15             Presented in this slide is a history of

 

 16   the clinical study.  The clinical study was

 

 17   approved in October of 1992 as a two-arm

 

 18   prospective and retrospective, non-randomized,

 

 19   multi-center clinical trial.

 

 20             The initial enrollment of the trial began

 

 21   in September of 1993 with a sample size of 64

 

 22   patients with an equal amount of device patients

 

 23   and control patients.  This was based on a 90

 

 24   percent power to detect a difference in the

 

 25   clinical outcome between control patients and

 

                                                                86

 

  1   device patients surviving to 30 days

 

  2   post-transplant.

 

  3             The assumption was that 20 percent of the

 

  4   control patients would survive to 30 days

 

  5   post-transplant, while 60 percent of the device

 

  6   patients would survive to 30 days post-transplant.

 

  7             The control arm of the clinical study

 

  8   consisted of 3 different groups.  It should be

 

  9   noted that all patients within these groups met the

 

 10   inclusion and exclusion criterias of the study.

 

 11   Thirty-two of the 35 control patients were

 

 12   identified through a retrospective examination of

 

 13   medical records.

 

 14             Twenty-two of the 32 control patients were

 

 15   identified from 1991 to 1993, before the first

 

 16   implant of the device in the trial had actually

 

 17   occurred.  Ten of the 32 patients, however, were

 

 18   identified from 3 of the primary implant centers

 

 19   from 1994 to 2002.  Three of the control patients

 

 20   were found concurrently through the trial and had

 

 21   refused treatment with the device.

 

 22             [Slide.]

 

 23             According to our regulations, a device

 

 24   that is intended to support or sustain human life

 

 25   is subject to premarket approval.  In order to

 

                                                                87

 

  1   demonstrate premarket approval, a sponsor must

 

  2   provide data that adequately demonstrates a

 

  3   reasonable assurance of safety and efficacy of the

 

  4   device.

 

  5             In order to demonstrate a reasonable

 

  6   assurance of safety and efficacy, the following

 

  7   relevant factors must be considered.  The patient

 

  8   population for which the device is intended, the

 

  9   conditions for use as suggested in the labeling,

 

 10   the probable benefit of the device versus the

 

 11   probable injury it may cause, as well as the

 

 12   reliability of the device.

 

 13             [Slide.]

 

 14             In regard to the preclinical or

 

 15   engineering topics that were involved in this

 

 16   application, these were some of the issues that

 

 17   were considered.  These topics were deemed adequate

 

 18   in determining reasonable assurance of safety of

 

 19   the device.

 

 20             [Slide.]

 

 21             These engineering topics, however, are

 

 22   still under review and we are actively working with

 

 23   the sponsor to resolve them in a matter of time.

 

 24   We do not believe that any of these issues will

 

 25   hinder the progress of our review.

 

                                                                88

 

  1             [Slide.]

 

  2             In conclusion, the results of the

 

  3   preclinical testing in conjunction with the outcome

 

  4   of the reliability  results from the clinical trial

 

  5   demonstrate a reasonable assurance of the device's

 

  6   safety.  Dr. Julie Swain, however, will be

 

  7   presenting the efficacy of this device.

 

  8             I now turn over the rest of the

 

  9   presentation to Dr. Lilly Yue, who will be

 

 10   presenting the statistical analysis.

 

 11                       Statistical Summary

 

 12             DR. YUE:  Good morning.  My name is Lilly

 

 13   Yue, statistician in the Division of Biostatistics,

 

 14   CDRH, FDA.

 

 15             [Slide.]

 

 16             In this presentation, I will briefly speak

 

 17   on study design and concentrate on statistical

 

 18   evaluation of effectiveness, and then give a

 

 19   statistical summary.

 

 20             [Slide.]

 

 21             This is a two-arm, non-randomized clinical

 

 22   trial conducted at five centers in the United

 

 23   States between January 1991 and September 2002.  Of

 

 24   the 35 control patients, 32 were identified by

 

 25   retrospective review, and the remaining 3 were

 

                                                                89

 

  1   found by prospective measurement during the study.

 

  2             Ninety-five patients received the total

 

  3   artificial heart.  Of these, 81 patients were

 

  4   defined out of the implant group for evaluation of

 

  5   effectiveness.

 

  6             The primary effectiveness endpoint was

 

  7   treatment success at 30 days post-transplant

 

  8   according to prespecified clinical criteria.

 

  9             Secondary effectiveness endpoints included

 

 10   survival to transplant and survival to 30 days

 

 11   post-transplant.

 

 12             Adverse events were evaluated for the

 

 13   determination of the safety of the device.

 

 14             I will focus on the statistical evaluation

 

 15   of effectiveness.

 

 16             [Slide.]

 

 17             On this slide, x axis denotes the year of

 

 18   implant, and the y axis represents the number of

 

 19   patients.  Red is for control and blue is for

 

 20   implanted group.

 

 21             The majority of control patients were

 

 22   collected in the early nineties, and the implant

 

 23   time is imbalanced between the two treatment

 

 24   groups.

 

 25             [Slide.]

 

                                                                90

 

  1             Also, there are multiple imbalances of

 

  2   baseline covariates, such as risk factors and

 

  3   previous intervention. In terms of some baseline

 

  4   covariates, implant patients were sicker than

 

  5   control patients, and with respect to some others,

 

  6   control patients were sicker.  For example, 60

 

  7   percent control patients had a prior cardiac

 

  8   surgery and 38 percent implant patients did.

 

  9             [Slide.]

 

 10             Given these imbalances, we can see that

 

 11   the two treatment groups are not comparable before

 

 12   the implant, so any direct treatment comparisons in

 

 13   the effectiveness endpoints are inappropriate, and

 

 14   all p-values from direct treatment comparisons are

 

 15   not interpretable.

 

 16             [Slide.]

 

 17             An immediate question to ask is, "Can we

 

 18   do treatment comparisons adjusting for these

 

 19   imbalanced covariates?"

 

 20             The sponsor performed traditional

 

 21   covariate analysis, such as logistic regression,

 

 22   and propensity score analysis on the proportion of

 

 23   patients with treatment success, survival to

 

 24   transplant, and survival to 30 days

 

 25   post-transplant, respectively.

 

                                                                91

 

  1             For the propensity score analysis, I would

 

  2   like to give you a little introduction.  Let's

 

  3   assume we have just one covariate to adjust for,

 

  4   for example, age.

 

  5             [Slide.]

 

  6             We can do this in two ways.  Number 1,

 

  7   matching patients with respect to age.  We randomly

 

  8   select one implant patient, then match him or her

 

  9   with a control patient with the closest age.

 

 10             We have a matched pair.  Now, we continue

 

 11   the process.  Finally, we can compare the two

 

 12   treatment groups based on the matched pairs.

 

 13             Number 2.  We can divide all patients into

 

 14   several age subclasses.  Within each subclass,

 

 15   patients age is relatively similar, and the

 

 16   distribution of age is relatively balanced between

 

 17   the two treatment groups, so that treatment groups

 

 18   are comparable with respect to age.

 

 19             Within each age subclass, we compare two

 

 20   treatments and obtain subclass-specific treatment

 

 21   difference, then estimate overall treatment

 

 22   difference by a weighted average.

 

 23             However, if patients from one treatment

 

 24   group are much older than those from the other

 

 25   group, that means the two treatment groups do not

 

                                                                92

 

  1   overlap with respect to age, then age factor could

 

  2   be confounded with the treatment, and no

 

  3   statistical methodology can be used appropriately

 

  4   to adjust for age in this extreme case.

 

  5             [Slide.]

 

  6             However, usually we have many covariates

 

  7   that should be adjusted simultaneously.  One way to

 

  8   do this is to perform propensity score analysis.

 

  9             The basic idea of propensity score is to

 

 10   replace the collection of covariates with one

 

 11   single number, which is called the propensity

 

 12   score.

 

 13             For example, giving a patient's age,

 

 14   duration of disease, and the status whether the

 

 15   patient had prior cardiac surgery, and so on, we

 

 16   can estimate a propensity score for the patient

 

 17   through a statistical model.

 

 18             Each patient just had one propensity

 

 19   score, just as the patient has just one age value.

 

 20             [Slide.]

 

 21             In definition, propensity score here is

 

 22   the conditional probability of receiving the total

 

 23   artificial heart, given a patient's observed

 

 24   baseline covariate values.

 

 25             [Slide.]

 

                                                                93

 

  1             When we use a propensity score to

 

  2   simultaneously  balance many covariates and reduce

 

  3   the bias in treatment comparison, we have to point

 

  4   out its limitations for use.

 

  5             Propensity score methods can only adjust

 

  6   for observed covariates and not for unobserved

 

  7   covariates.  It is always a limitation of

 

  8   non-randomized studies compared with randomized

 

  9   studies, when the randomization tends to balance

 

 10   the distribution of all covariates, observed and

 

 11   unobserved.

 

 12             Propensity score works better when there

 

 13   are many measured covariates, and it is seriously

 

 14   degraded when important variables affecting

 

 15   treatment selection have not been collected.

 

 16             [Slide.]

 

 17             In expectation, when the propensity scores

 

 18   are   balanced across the treatment and control

 

 19   groups, the distribution of all regional covariates

 

 20   are balanced across the treatment groups, so the

 

 21   propensity score is also called a balancing score.

 

 22             We can use the propensity scores as a

 

 23   diagnostic tool to measure treatment group

 

 24   comparability.

 

 25             If the two treatment groups overlap well

 

                                                                94

 

  1   enough in terms of propensity scores, we compare

 

  2   the two treatment groups adjusting for the

 

  3   propensity score, just as adjusting for age.

 

  4             [Slide.]

 

  5             In this study, we performed multiple

 

  6   imputations for 19 percent patients with missing

 

  7   baseline covariate values.  Otherwise, these

 

  8   patients would be excluded in propensity score

 

  9   modeling.

 

 10             We adjusted for all imbalanced and/or

 

 11   clinically important baseline covariates as well as

 

 12   the year of the implant, since over last 10 years,

 

 13   the medical management of heart failure has changed

 

 14   significantly.

 

 15             We found that the propensity score model

 

 16   accurately predicted the treatment group

 

 17   membership.  However, the two treatment groups did

 

 18   not overlap enough with respect to propensity

 

 19   scores to allow sensible  treatment comparison.

 

 20             [Slide.]

 

 21             Here are boxplots demonstrating the

 

 22   distributions of the propensity scores for the two

 

 23   treatment groups, respectively.  Let's look at the

 

 24   boxplot for the control group.

 

 25             The blue box here covers the middle 50

 

                                                                95

 

  1   percent of patients.  The lower and upper edges are

 

  2   respectively, the first quartile and the third

 

  3   quartile.  The median is the line here with red

 

  4   dot.  The width of the box is arbitrary and has no

 

  5   real meaning.

 

  6             "Whiskers" here, actually, we should have

 

  7   one here and a line here in the bottom, extend

 

  8   vertical lines from the center edge to extreme

 

  9   values.  All points that are more extreme than the

 

 10   "whiskers," if any are here, are potential

 

 11   outliers.  Now, here, the control group does not

 

 12   have outliers, but TAH group has a lot outliers.

 

 13             More than 75 implant patients do not have

 

 14   any overlap with the control patients.  Only 14

 

 15   outliers, about 17 percent implant patients, have

 

 16   some degree of overlap with control group.

 

 17             [Slide.]

 

 18             Let's check the treatment comparability in

 

 19   another way.  If we divide all patients into 5

 

 20   quintiles, in the fourth and fifth quintiles, 56

 

 21   percent implant patients do not have any control

 

 22   patients to compare with.

 

 23             At the same time, in the first quintile,

 

 24   66 percent of control patients do not have any

 

 25   implant patients to compare with.  In the third

 

                                                                96

 

  1   quintile, we have 23 implant patients, but only 1

 

  2   control patient.

 

  3             Only in the second quintile, the treatment

 

  4   comparison is meaningful with 15 percent implant

 

  5   patients and 31 percent control patients.

 

  6             So, what can we do with the 56 percent

 

  7   implant patients in the first and the fifth

 

  8   quintiles?  Throw them away?  No, we would have

 

  9   serious concerns with that no matter what study

 

 10   results you could get from the remaining patients.

 

 11             So, we have to conclude that the two

 

 12   treatment groups are not comparable at all, and

 

 13   then, any treatment comparisons adjusting for

 

 14   imbalanced covariates are problematic.

 

 15             [Slide.]

 

 16             How to proceed?  Now, we have to say any

 

 17   judgment of the performance of the device has to be

 

 18   based on the results from the total artificial

 

 19   heart group alone.

 

 20             [Slide.]

 

 21             In this group with 81 patients, the

 

 22   treatment success at just 30 days post-transplant

 

 23   has a point estimate of 69 percent and 95 percent

 

 24   confidence interval from 58 to 79 percent.

 

 25             Please note that it is inappropriate to

 

                                                                97

 

  1   consider the point estimate alone and ignore the

 

  2   variability associated with the point estimate

 

  3   since 7 survivors out of 10 patients will give you

 

  4   a different degree of evidence from 70 survivors

 

  5   out of 100 patients.

 

  6             The difference is reflected in the lower

 

  7   limit here, so we need to pay attention to lower

 

  8   limit of a confidence interval.

 

  9             [Slide.]

 

 10             Here are some results for 6-month survival

 

 11   from implant, 1-year survival from implant, and

 

 12   1-year conditional survival from transplant based

 

 13   on proportion.

 

 14             [Slide.]

 

 15             The mean time to transplant or death

 

 16   before transplant is 79 days and median is 47 days.

 

 17             The sponsor also performed Kaplan-Meier

 

 18   survival estimates prior to transplant.

 

 19             Here, death is an event and transplant is

 

 20   treated as censoring.  However, it is not clear to

 

 21   us if sicker patients received transplant sooner in

 

 22   this study.  If that is the case, the assumption

 

 23   underlying Kaplan-Meier estimates of independence

 

 24   of censoring and event is invalid.

 

 25             So, the Kaplan-Meier probability estimates

 

                                                                98

 

  1   are biased, and the survival probability estimates

 

  2   based on proportion demonstrated before are more

 

  3   appropriate.

 

  4             [Slide.]

 

  5             In summary, without appropriate control,

 

  6   it is difficult to perform statistical evaluation

 

  7   of the effectiveness of the device.

 

  8             For survival prior to transplant,

 

  9   Kaplan-Meier survival estimates are potentially

 

 10   seriously biased.

 

 11             Thank you .

 

 12                         Clinical Review

 

 13             DR. SWAIN:  Thank you.

 

 14             [Slide.]

 

 15             I am going to present the clinical review

 

 16   of the CardioWest TAH, and this review was done by

 

 17   both Eleana Pina, heart transplant cardiologist,

 

 18   and me as a cardiac surgeon.

 

 19             [Slide.]

 

 20             It is first important to note, as Dr. Yue

 

 21   just said, that we really don't find that it is

 

 22   statistically or scientifically valid to consider

 

 23   the control group, probably not clinically valid

 

 24   either, so we essentially have a single-arm study,

 

 25   and what are we to do to evaluate the study.

 

                                                                99

 

  1             It is first important to note that there

 

  2   are really no randomized control studies for

 

  3   bridge-to-transplant devices, and in devices that

 

  4   have been approved, when you look at the

 

  5   literature, there is really no comparable control

 

  6   groups in previous BTT studies that would withstand

 

  7   the rigorous analysis that the FDA does at this

 

  8   time on control groups.

 

  9             Also, in general, there is slow enrollment

 

 10   because of the relatively small number of patients

 

 11   requiring ventricular assist devices for a bridge

 

 12   to transplant, so the 10-year duration of this

 

 13   study is really not out of line with previous

 

 14   devices.

 

 15             What we decided to do, not relating to

 

 16   this device, but two years ago the agency decided

 

 17   to see if performance goals for left ventricular

 

 18   assist devices could be developed using both inside

 

 19   and outside consultants, it was looked at to see if

 

 20   the literature could help us develop performance

 

 21   goals.

 

 22             It is probably important for the panel and

 

 23   especially the audience to understand that this

 

 24   does not imply that in the future, the agency will

 

 25   accept proposals for single-arm

 

                                                               100

 

  1   bridge-to-transplant trials, and also it really

 

  2   doesn't imply that for bridge to recovery or

 

  3   destination therapy devices for ventricular assist

 

  4   devices, that the agency will accept performance

 

  5   goals in single-arm trials.

 

  6             [Slide.]

 

  7             So, what we did was look at the

 

  8   literature, and as a surgeon, I can say that with a

 

  9   decade of experience with these devices, it is a

 

 10   little disappointing to find a paucity of rigorous

 

 11   scientific studies in this area, but these are some

 

 12   of the inclusion criteria that we used, as well as

 

 13   the exclusion criteria for papers.  Again, this is

 

 14   left ventricular assist devices.

 

 15             [Slide.]

 

 16             Well, what did we come up with?  We came

 

 17   up with a performance goal, not an OPC, but a

 

 18   performance goal based on the literature for

 

 19   survival to transplant.

 

 20             That appeared to be the one area that we

 

 21   could find some number of papers with somewhat

 

 22   comparable results, and we came up with a goal of

 

 23   65 to 70 percent.  Again, this was started about

 

 24   two years ago and finished about a year ago, well

 

 25   before we saw any of the results of the CardioWest

 

                                                               101

 

  1   device.

 

  2              At the same time, Marv Konstam's group,

 

  3   Dr. Wong was looking at a meta-analysis of the

 

  4   literature and came up with a performance goal or

 

  5   an average of survival to 30 days post-transplant

 

  6   of 74 percent.

 

  7             [Slide.]

 

  8             Well, then, I looked at the literature on

 

  9   BiVAD survival to transplant, and you can see on

 

 10   the left is our goal of 65 to 70 percent.

 

 11             On the right, the CardioWest survival to

 

 12   transplant at 79 percent for the core group, and

 

 13   the higher bar, the success, which is defined as

 

 14   survival to 30 days post-transplant, NY Class I or

 

 15   II, ambulatory, which I have a question about the

 

 16   hemiplegic patient, of the definition of

 

 17   ambulatory, also not on a ventilator and not on

 

 18   dialysis.

 

 19             You can see that even with the high bar of

 

 20   success, it was comparable to our performance goal.

 

 21   When you look at four papers I chose for the

 

 22   literature that looked at biventricular assist

 

 23   devices, you can see the results of survival in

 

 24   those four studies, and again, the CardioWest

 

 25   device is certainly comparable or better in those

 

                                                               102

 

  1   studies.

 

  2             It is important to note that the papers

 

  3   that I could find that both looked at LVAD devices

 

  4   and BiVAD, BiVADs were always less good results,

 

  5   and clinically, it is our opinion that BiVAD

 

  6   patients are certainly sicker.  So, we can see that

 

  7   when you look in relation to the literature,  the

 

  8   CardioWest device compares favorably.

 

  9             [Slide.]

 

 10             Well, what happens with left ventricular

 

 11   implantation with right ventricular failure?  It is

 

 12   a diagnostic dilemma, first of all, to define who

 

 13   on the operating table at the time of implantation

 

 14   needs a biventricular device.

 

 15             We know that when left ventricular devices

 

 16   are put in, the right ventricular failure rate is

 

 17   about 10 to 30 percent.  In fact, an article in the

 

 18   Wong literature, 24 percent of the patients

 

 19   required RVADs.

 

 20             A very interesting article just came out

 

 21   of Mehmet Oz's group, with their large experience,

 

 22   saying that they had a 7 percent right ventricular

 

 23   failure rate requiring biventricular assist.

 

 24             What is usually done?  Well, these days,

 

 25   medical therapy, inotropes, volume load, and, in

 

                                                               103

 

  1   particular now, off-label use of nitric oxide is

 

  2   the treatment of choice for right ventricular

 

  3   failure with left ventricular assist device in

 

  4   place.

 

  5             Then, there are short-term pumps that can

 

  6   be used, and then one approved percutaneous

 

  7   external biventricular device.  So, that is the

 

  8   current treatment, and the diagnostic dilemma for

 

  9   the panel to consider is who needs a biventricular

 

 10   device at the time of bridge.

 

 11             [Slide.]

 

 12             Let's look now at the data for CardioWest,

 

 13   and the study was approved in 1993, and the agency

 

 14   agreed at that time that there was not clinical

 

 15   equipoise for a randomized study, and FDA approved

 

 16   the use of the control group that was subsequently

 

 17   utilized for these patients.

 

 18             [Slide.]

 

 19             As you can see, we are mainly considering

 

 20   the 81 core patients who met all the inclusion and

 

 21   exclusion criteria for the device.

 

 22             [Slide.]

 

 23             These are the data that you have seen.

 

 24   For the core patients, 79 percent survival, all

 

 25   device patients, which is 75 percent survival.  The

 

                                                               104

 

  1   out-of-U.S. patients, 59 percent survival.

 

  2   Success, again the bar is higher, 69 percent, again

 

  3   looking at our performance goal of 65 to 70 percent

 

  4   on the righthand side.

 

  5             So, this device really appears to again

 

  6   look reasonably equivalent or better to the other

 

  7   devices.

 

  8             [Slide.]

 

  9             What about adverse events?  Well, it is

 

 10   really difficult in the literature to develop a

 

 11   performance goal for adverse events, because mainly

 

 12   the definitions are completely different among the

 

 13   studies, and some of the studies don't even list

 

 14   definitions.

 

 15             The rates differ for devices, and the

 

 16   rates for the same device changes over time, so we

 

 17   think that this clearly has to rely on clinical

 

 18   judgment.

 

 19             [Slide.]

 

 20             When you look at the adverse event rates

 

 21   that have been presented for the CardioWest device,

 

 22   they are really comparable especially to the

 

 23   meta-analysis by Wong in Konstam's group, where

 

 24   they found a bleeding incidence at 28 percent, and

 

 25   I have to second the idea that right ventricular

 

                                                               105

 

  1   assist patients are certainly sicker.

 

  2             Even if they have a normal PT and a PTT,

 

  3   they really can't recruit the coagulation agents,

 

  4   and they have a higher incidence of bleeding no

 

  5   matter what you are doing, be it liver transplant

 

  6   or any other operation.

 

  7             So, Wong found a 28 percent incidence of

 

  8   bleeding with lots of definitions of that,

 

  9   infection of 22 percent, which compares favorably

 

 10   here, and a thromboembolism rate of 8 percent,

 

 11   which again compares favorably to this device.

 

 12             [Slide.]

 

 13             There are two areas of discussion that we

 

 14   would like the panel to look at in particu