DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR DEVICES AND RADIOLOGIC HEALTH
CIRCULATORY SYSTEM DEVICES PANEL
Thursday, March 18, 2004
Hilton Gaithersburg Washington D.C., North
Salons A, B and C
620 Perry Parkway
Cynthia Tracy, M.D., Acting Chair
Geretta Wood, Executive Secretary
Salim Aziz, M.D.
Mitchell Krucoff, M.D.
William Maisel, M.D., MPH
Christopher J. White, M.D.
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.
L. Henry Edmunds, Jr., M.D.
Bram Zuckerman, M.D.
C O N T E N T S
Call to Order
Cynthia Tracy, M.D. 4
Conflict of Interest Statement
Geretta Wood 4
Julia Marders 8
Wolf Sapirstein, M.D., MPH, FACS 17
Kachi Enyinna 26
Open Public Hearing:
Randall Wolfe, M.D. 31
Robert W. Emery, M.D. 41
G. Phillip Schoettle, M.D. 50
Robert Frater, M.D., 55
Michael Mack, M.D., STS/AATS 66
Mark Slaughter, M.D. 76
Henry Frank Martin, M.D. 85
Bernard Hausen, M.D. 90
Prof. Uwe Klima 96
Open Committee Discussion 107
Cynthia Tracy, M.D. 262
1 P R O C E E D I N G S
2 Call to Order
3 DR. TRACY: Good morning, everybody. I
4 would like to call to order this meeting of the
5 Circulatory System Devices Panel. The topic today
6 is a discussion of type of data and study required
7 to effectively evaluate performance of aortic
8 anastomotic devices for marketing.
9 Conflict of Interest
10 MS. WOOD: The following announcement
11 addresses conflict of interest issues associated
12 with this meeting and is made a part of the record
13 to preclude even the appearance of an impropriety.
14 To determine if any conflict existed, the
15 agency reviewed the submitted agenda and all
16 financial interests reported by the committee
17 participants. The conflict of interest statutes
18 prohibit special government employees from
19 participating in matters that could affect their or
20 their employers' financial interests. However, the
21 agency has determined that participation of certain
22 members and consultants, the need for whose
23 services outweighs the potential conflict of
24 interest involved, is in the best interest of the
1 A waiver has been granted for Dr. Clyde
2 Yancy and a wavier was previously granted for Dr.
3 Judah Weinberger for their financial interests in
4 firms at issue that could potentially be affected
5 by the panel's recommendations. The waivers allow
6 these individuals to participate fully in today's
7 deliberations. Copies of these waivers may be
8 obtained from the agency's Freedom of Information
9 Office, Room 12A-15 of the Parklawn Building.
10 We would like to note for the record that
11 the agency took into consideration other matters
12 regarding Drs. Thomas Ferguson, Mitchell Krucoff,
13 Cynthia Tracy, Judah Weinberger and Clyde Yancy.
14 These panelists reported past or current interests
15 involving firms at issue but in matters that are
16 not related to today's agenda. The agency has
17 determined, therefore, that these individuals may
18 participate fully in the panel's deliberations.
19 In the event that the discussions involve
20 any other products or firms not already on the
21 agenda for which an FDA participant has a financial
22 interest, the participants should excuse him or
23 herself from such involvement and the exclusion
24 will be noted for the record.
25 With respect to all other participants, we
1 ask in the interest of fairness that all persons
2 making statements or presentations disclose any
3 current or previous financial involvement with any
4 firm whose products they may wish to comment upon.
5 DR. TRACY: Just before we get started, I
6 would just like to ask everybody to be sure that
7 you are speaking directly into the microphone,
8 including the speakers who will be coming up later
9 in the open public hearing. A transcript is being
10 made from these presentations today.
11 At this time I would like to ask the panel
12 members to introduce themselves.
14 MR. MORTON: I am Michael Morton. I am
15 the industry representative. I am an employee of
17 DR. WEINBERGER: Judah Weinberger,
18 Director of Interventional Cardiology at Columbia.
19 DR. YANCY: Clyde Yancy, Director of Heart
20 Failure and Transplantation at UT Southwestern, in
22 DR. WHITE: Chris White. I am the Chief
23 of Cardiology at Ochsner Clinic Foundation.
24 DR. HIRSHFELD: John Hirshfeld. I am an
interventional cardiologist in the University of
2 DR. KATO: Norman Kato, cardiovascular
3 surgeon, private practice, Encino, California.
4 MS. WOOD: Geretta Wood, Executive
6 DR. TRACY: I am Cindy Tracy,
7 electrophysiologist, George Washington University.
8 DR. EDMUNDS: I am Hank Edmunds, Professor
9 of Surgery at the University of Pennsylvania.
10 DR. FERGUSON: Tom Ferguson,
11 cardiovascular surgeon, Washington University St.
13 DR. KRUCOFF: Mitch Krucoff,
14 interventional cardiologist, Duke University
15 Medical Center; Director of Devices Trials, Duke
16 Clinical Research Institute.
17 DR. MAISEL: William Maisel,
18 electrophysiologist, Brigham & Women's Hospital.
19 DR. BLUMENSTEIN: Brent Blumenstein,
20 biostatistician, Seattle, Washington.
21 DR. BRIDGES: Charles Bridges, Chief
22 Cardiovascular Surgery Pennsylvania Hospital,
23 University of Pennsylvania.
24 MS. WELLS: Chrissy Wells. I am the
consumer representative on the panel.
1 DR. ZUCKERMAN: Bram Zuckerman, Director,
2 FDA, Division of Cardiovascular Devices.
3 DR. TRACY: Thank you. At this point we
4 will have the FDA presentation, and Julia Marders,
5 from the Office of Surveillance, will be the
6 opening speaker.
7 FDA Presentation
8 MS. MARDERS: Good morning.
10 My name is Julia Marders, and I am a nurse
11 analyst in the Division of Postmarket Surveillance,
12 Office of Surveillance and Biometrics.
14 I will present an analysis of adverse
15 event reports received by the FDA on aortic
16 anastomotic devices. My presentation will begin
17 with a brief description of the Medical Device
18 Reporting System, MDR, which is a system for
19 adverse events and product problems, and include a
20 discussion of its limitations.
21 Next, I will describe the database, search
22 methodology used to obtain the reports of aortic
23 anastomotic devices and provide a summary of
24 findings and analysis of conclusions. Then I will
finish the presentation with conclusions,
1 considerations and questions for the panel to
4 The Medical Device Reporting System is a
5 nationwide passive surveillance system which
6 includes both mandatory and voluntary reporting.
7 Since 1984 manufacturers and importers have been
8 required to submit reports to the FDA of
9 device-related deaths and serious injuries, as well
10 as events involving device malfunctions that may
11 cause or contribute to death or serious injury.
12 The Safe Medical Devices Act of 1990
13 introduced mandatory reporting of device-related
14 deaths and serious injuries by user facilities,
15 most notably hospitals and nursing homes.
16 Voluntary medical device adverse event and problem
17 product reports, most often submitted by healthcare
18 practitioners, consumers, patients or family
19 members, are received through the FDA's MedWatch
20 program. In general, approximately 95 percent of
21 the reports received by FDA are from manufacturers,
22 one percent from importers, and the remainder is
23 equally split between voluntary and user
1 The Medical Device System, which is the
2 MDR system as most people know it, while providing
3 signals of actual and potential device-related
4 problems, has some limitations. Under-reporting of
5 adverse events to hospitals, manufacturers and the
6 FDA by healthcare practitioners is a well-known and
7 recognized phenomenon. Thus, events reported
8 through MDR represent a subset of the total
9 occurrence of events.
10 In addition, manufacturers are not
11 required to submit denominator information such as
12 number of devices manufactured, distributed and
13 implanted. Thus, due to under-reporting and lack
14 of denominator data accurate incidence rates are
15 unable to be determined based on MDR data alone.
16 Furthermore, reports received may not be
17 representative and reflective of a variety of
18 reporting biases. Thus, for example, reporting may
19 vary by manufacturer or by the presence or absence
20 of publicity.
21 Although there is a regulatory requirement
22 for a minimum data set, event narrative
23 descriptions vary in completeness and complexity.
24 For example, one aortic anastomotic device report
indicates failure of the connector as the entire
1 event description and no further details are
2 provided. In addition, many reports do not contain
3 results of manufacturer failure analyses. Often
4 devices are not returned to the manufacturer for
5 evaluation because they are discarded or remain
6 implanted. Thus, root causes for reported events
7 are often unable to be determined.
9 Now I will describe the search methodology
10 used to obtain the data set of aortic anastomotic
11 device reports in this presentation and present the
12 findings. First I searched the database by product
13 code. All medical devices approved or cleared for
14 marketing have a unique three-letter identifier
15 called the product code. Next, I narrowed the
16 search by date. This search includes events of
17 aortic anastomotic devices that were reported from
18 May 24, 2001, the first marketing clearance date
19 for these devices, to March 1, 2004. I also
20 performed additional database queries by brand
21 names to validated that I had captured all aortic
22 anastomotic device reports that have been entered
23 into the database.
25 Now the findings, a total of 213 reports
1 are in the database and most reports were received
2 in 2003. The number of death reports is 23;
3 injury, 185; and malfunction, 5. The vast majority
4 of these reports, that is 203, came from
5 manufacturers, with 2 from user facilities and 8
8 Patient age was provided in 129 of the 213
9 reports and ranged from ages 35 to 83 years, with
10 most in the 50-65 age range. Slightly over half of
11 these events are noted in males, a quarter in
12 females and a quarter were gender unspecified. One
13 hundred and seventy-three events, or 81 percent,
14 occurred with patients in the U.S. and 14, which is
15 7 percent, with patients outside the U.S.; 26 event
16 reports, 12 percent, did not specify whether the
17 event is foreign or domestic.
19 Of the 23 death reports, 22 were from
20 manufacturers and one from a user facility. All
21 patient deaths occurred within 18 days of
22 implantation and 15 of the deaths occurred within 3
23 days of implantation. Interestingly, one patient
24 actually had both a dissection operatively and a
detachment postoperatively and is included in 2 of
1 the problem categories listed on this slide.
2 Additionally, another patient had both thrombus and
3 aortic detachment that was discovered on
4 postoperative day 2 when the patient coded.
5 Twelve reports indicate the problem of
6 occlusion or thrombus at the connector site. One
7 report describes the patient was noted to have a
8 predisposing hypercoagulable state, and 2 reports
9 indicate that patients had atrial fibrillation.
10 Aortic dissection associated with deployment or
11 after the connector is placed was noted in 7
12 reports. Device detachment, resulting in
13 hemorrhagic shock, occurred in 6 reports. None of
14 the devices associated with death were returned to
15 the manufacturer for evaluation and the
16 manufacturer has not been able to determine the
17 root cause of the events.
19 Now I will present an actual report to
20 illustrate these findings in a more clinically
21 relevant way. A patient was implanted with an
22 aortic anastomotic device during an off-pump
23 procedure. No difficulties were encountered with
24 loading or deployment of the device. Recovery was
good for approximately 40 hours when the patient
1 suddenly lost consciousness after a dramatic drop
2 in blood pressure. CPR was initiated and blood
3 appeared in the drains. At re-operation, the
4 aortic connector was detached from the aorta and
5 the patient died after 10 minutes. The autopsy
6 revealed the cause of death was hemorrhagic shock.
8 A total of 185 injuries were reported.
9 Stenosis and occlusion are overwhelmingly noted to
10 be the first and second most frequently reported
11 problems respectively. Although infrequently
12 reported, events involving device detachment have
13 also resulted in serious injury. Clinically, the
14 reported outcomes of stenosis and occlusion
15 resulted in life-threatening conditions resulting
16 in shortness of breath, chest pain, arrhythmias,
17 subsequent myocardial infarction and/or hemodynamic
18 instability requiring either surgical or
19 interventional treatment including catheterization
20 for PTCA and stenting.
21 The time from implantation to injury, as
22 noted in 37 of the 185 reports submitted, of the 30
23 noting stenosis or occlusion most, or about 60
24 percent, occurred within 90 days; 4 events occurred
within 4 days. The other 7 are
associated with a
1 variety of patient problems other than stenosis or
2 occlusion. Three reports of device detachment
3 occurred within one day of surgery, and another
4 event atypically occurred after 97 days, possibly
5 due to a fragile aorta and placement of the
6 connector on a pseudoaneurysm. Of all the
7 injuries, only 2 devices were returned to the
8 manufacturer for evaluation, both of which resulted
9 in manufacture evaluation indicating no device
10 failure detected.
12 Five reports indicated a device
13 malfunction. One report states the device was not
14 able to be used because the anchor tip was closed.
15 Two reports indicate the aortic plug was not seen
16 by the surgeon upon inspection of the device. Both
17 of these patients have not experienced any adverse
18 consequences. The fourth report indicates a device
19 malfunction resulting in an aortic laceration
20 requiring repair. It is not clear why the user
21 facility reported this event as a malfunction
22 rather than an injury, and no information was
23 included about the patient's outcome. Follow-up is
24 ongoing. The fifth report indicates failure of the
connector, with no other details, other than
1 indicating no consequences to the patient.
3 Conclusions--the reports of serious
4 adverse outcomes related to aortic anastomotic
5 device use raises a signal of a potential public
6 health problem. Some of these occurrences are
7 catastrophic, such as aortic dissection or device
8 detachment, and not expected. Others, for example,
9 occlusion or stenosis, may be expected depending on
10 the patient's underlying condition of adequacy of
11 antiplatelet therapy, or may reflect device-related
12 events, for example, stenosis at the connection
13 site or thrombosis potentially related to
14 bioincompatibility or poor hemodynamics. Lastly,
15 the reported information to date reflects
16 short-term experience. Long-term failure
17 information is also important.
19 Considerations--additionally, there are
20 two other important points to consider, first,
21 failure analyses of this adverse event data are
22 lacking or limited. The underlying root cause of
23 these events, particularly occlusion and stenosis,
24 is unknown. Multiple factors may be involved which
can make the evaluation of these events difficult.
1 Second, this adverse event data needs to be
2 factored into the risk/benefit profile for these
5 To conclude my presentation, I have the
6 following three questions for the panel to consider
7 that are based on adverse event report findings:
8 First is the question of collection of
9 long-term failure rate data. Should a longer
10 period of time for manufacturer collection of
11 device performance data post implantation be
12 required to fully understand aortic anastomotic
13 device failures?
14 Next, should studies comparing short- and
15 long-term patient outcomes between standard
16 suturing versus sutureless aortic anastomotic
17 devices to address risk/benefit issues be
19 Finally, should further study of
20 device-related events be considered?
21 I encourage the panel to consider these
22 questions before making final recommendations.
23 That concludes my part of the presentation and now
24 I will turn over to Wolf Sapirstein.
25 DR. SAPIRSTEIN:
Dr. Tracy, panel members,
1 good morning.
3 The people listed up there are members of
4 the Division of Cardiovascular Devices. My name
5 is Wolf Sapirstein. We are mandated by statute to
6 regulate cardiovascular devices, and are hopeful
7 that this panel will generate guidance for us in
8 undertaking this activity for these new and unique
9 devices used in treatment of coronary-artery
12 Vascular suturing was introduced by Carrel
13 in 1903 and has changed little over the next 100
14 years, except for the replacement of catgut with
15 synthetic suture. After about 30 years of attempts
16 by various investigators internationally, an
17 automatic device to effect vascular anastomoses,
18 the Symmetry Aortic Connector, was cleared in 2001
19 for commercial use by the agency. The drive for
20 development of these devices has undoubtedly been
21 coronary arterial bypass graft procedure which also
22 underscores the clinical importance of assuring
23 safety and effectiveness for these devices.
24 Incremental modifications to the coronary
arterial bypass graft procedure are seminal to the
1 acceleration in development of these devices. The
2 surgeons among us will have the forbearance, I
3 hope, while I undertake a thumbnail sketch of the
4 changes that have taken place in the performance of
5 the coronary arterial bypass grafting procedure.
7 The CABG procedure was the earliest
8 surgical therapy validated with a randomized,
9 controlled trial, the Coronary Arterial Surgery
10 Study. Autogenous venous conduits remain
11 extensively employed with anastomosis performed to
12 the aorta and the coronary artery distal to the
13 obstructive lesion. Induced ventricular
14 fibrillation and anoxic cardiac arrest with
15 hypothermic protection were initially used to
16 provide the quiet field demanded by the challenge
17 of suturing vessels 1-2 mm in diameter.
18 Cardioplegia inducing perfusion of the coronary bed
19 has since produced cardiac standstill with improved
20 myocardial preservation during the ischemic period
21 of conduit anastomosis.
22 Resistance of the internal thoracic
23 artery, ITA, to the atherosclerotic degeneration
24 that seemed inexorable with vein conduits has led
to is preferential employment since the late 1980s.
1 This also provides the advantage of eliminating
2 need for an aortic anastomosis. Patient survival
3 has since been shown in several studies to be
4 closely related to the effectiveness of
5 revascularization achieved for the anterior surface
6 of the heart and left ventricle. These
7 developments, patency of the ITA and anterior
8 cardiac revascularization justified introduction of
9 the minimal access direct CABG procedure in the
10 1990s to perform an isolated LIMA-LAD bypass. This
11 was shortly followed by beating heart and finally
12 off-pump CABG with elimination of extracorporeal
13 circulatory support entirely. Thus, were the ill
14 effects of cardiac arrest, extracorporeal
15 circulatory perfusion and aortic clamp manipulation
16 of the aorta obviated.
18 These modifications made to the CABG
19 procedure addressed its changing role in an era of
20 increasing catheter-mediated coronary treatment.
21 The MIDCAB is seen as reducing the morbidity of
22 incisional trauma, particularly in an increasingly
23 older patient cohort and patients with more
24 compromised coronary circulation not amenable to
percutaneous coronary interventions, and these
1 patients become candidates for operative
2 intervention. Dispensing with cardiopulmonary
3 bypass eliminated a potent activator of both the
4 systemic inflammatory response and the various
5 immunological cascades.
6 There is also increasing recognition of
7 the frequency with which neurocognitive
8 deterioration, apart from the more overt cerebral
9 ischemic events, occur with CABG procedures.
10 Extracorporeal cardiopulmonary bypass and
11 manipulation of the atherosclerotic aorta for
12 cardiopulmonary bypass perfusion, as well as
13 conduit anastomosis, have been indicted as
14 etiologic factors for these complications.
15 Anastomotic devices, by facilitating the various
16 modifications to the CABG that address morbidity,
17 can certainly play a major role in reducing this
20 Several studies during the development
21 stage of CABG evaluated the effectiveness of this
22 revascularization procedure measured as durability
23 of patency. While this slide presents a generally
24 accepted distillation of these study findings, it
should be noted that patency of CABG is dependent
1 on multifactorial elements that have likely been
2 affected by recent changes to the operation itself
3 that are still being evaluated, and by new measures
4 to inhibit the progression of coronary-artery
5 disease. This has to be considered when evaluating
6 anastomotic devices in a comparison to these
7 conduit patency rates.
9 Failure of the CABG conduit has been
10 attributed to several causes which are listed here.
11 They are broadly stratified by the period of their
12 most prominent effects: the perioperative failures;
13 6 months o 1 year, failure due to neointimal
14 hyperplasia; and the continuum from 6 months on are
15 both coronary-artery disease in the native vessel
16 and the conduit itself.
18 The advent of anastomotic devices carry a
19 promise for significant benefits in the performance
20 of the CABG procedure that go beyond simplifying
21 procedural mechanics for the benefit of the
22 operator. They have the potential for eliminating
23 many of the factors contributing to poor patient
24 outcome. It must be recognized that while the
precise benefit perceived for some of these recent
1 changes to the procedures, such as beating heart
2 and operations performed without cardiopulmonary
3 bypass, are as yet unresolved. The use of
4 non-suture constructed anastomoses will certainly
5 facilitate and increase the frequency of their use.
6 These are some of the benefits that seem intuitive
7 with anastomotic devices.
9 Well, Woody Allen has said every silver
10 lining has a dark cloud, and this is exactly true
11 with these anastomotic devices. Here are listed
12 some of the design characteristics that may
13 contribute to graft failure which do not obtain
14 with conventional sutured vascular connections.
16 In our evaluation of these devices for
17 clearance with a 510(k) notification, we have
18 required extensive preclinical data to support
19 limited clinical studies. The clinical material
20 was required to substantiate equivalence to
21 historical data for conduit patency, which was a
22 surrogate for correcting the deficiency in
23 myocardial perfusion.
24 We encountered some disagreement regarding
the study design, the duration of follow-up, and
1 the instruments for assessing effectiveness. While
2 general agreement exists regarding the use of
3 suture anastomosis as the gold standard to control
4 for patency, there is considerable advocacy to
5 employ measures of coronary perfusion for
6 assessment of patency. This is a reversal of the
7 original CABG use of patency as a surrogate for
9 With regard to duration of follow-up, the
10 initial concept was to take into consideration the
11 multifactorial causes of CABG failure by accepting
12 a relatively short period, such as 6-9 months, that
13 focuses on the adequacy of the anastomosis
14 constructed rather than the other factors in graft
15 failure. The changes made to the CABG procedure
16 itself and the introduction of measures aimed at
17 disease progression were not addressed. It was
18 also felt that a distinction could be made for
19 devices used on the proximal aortic or on the
20 distal coronary artery.
22 The problem encountered in designing a
23 study to evaluate these anastomosis devices goes
24 beyond the inherent problem of the multifactorial
causes of CABG failure. They
involve in general
1 the device-specific variables listed here that may
2 frustrate attempts at one-design-fits-all study
3 design for the devices.
5 From our initial experience with cleared
6 devices, we now have the belief that the rigor of a
7 randomized trials may be required unless there are
8 very mitigating circumstances to justify otherwise.
9 To this end, we would like input on an appropriate
10 template for study design that could be modified to
11 accommodate some of the variables intrinsic to
12 their use. This slide lists some of the
13 considerations we have encountered for designing a
14 study template and it is just put up for your
15 consideration as a straw man.
17 This slide represents a sample size
18 estimation for a one-armed study with the endpoint
19 for effectiveness based on the historical values
20 listed here for conduits performed with hand
21 suturing. For instance, a point estimate of 95
22 percent patency, with a lower confidence level
23 accepted as 5 percent, would require a sample size
24 of 150 patients for study. This is just placed
here for your consideration or evaluation for even
1 a one-armed study.
2 This completes my introduction to the
3 FDA's request for this panel's input in formulating
4 an appropriate regulatory approach for devices that
5 present the potential for critically affecting the
6 treatment of coronary arterial disease, which is
7 the wound stripe of modern society.
8 Kachi Enyinna, our lead engineer reviewer
9 for these devices, will now present or crystallize
10 some of the comments that I have made in the form
11 of questions that we would like this panel to
12 address in helping us wrestle with the regulation
13 of these devices. Thank you very much.
14 MR. ENYINNA: Good morning. My name is
15 Kachi Enyinna, biomedical engineer and lead
16 reviewer, Division of Cardiovascular Devices. I
17 will be presenting the questions we have come up
18 with and seeking some kind of guidance from panel
19 on how to evaluate clinical studies of these
20 devices. I would like to remind the panel members
21 to keep these questions in mind while I go over the
22 questions and to keep the questions in mind until
23 discussion time this afternoon allow members of the
24 medical community, as well as sponsors and industry
to speak before we discuss the questions.
2 Regarding trial design, the first
3 question, please comment on the choice of control
4 in the clinical trial required to evaluate vascular
5 anastomosis devices for CABG. The gold standard of
6 sutured CABG anastomoses has a well-documented
7 history of over thirty years.
9 Can historical data from sutured CABG
10 anastomosis device trials be used as the control in
11 the device studies?
13 Alternatively, are concurrently performed
14 CABG controls necessary given the multifactorial
15 causes of CABG failure, for example, technical
16 construction, extent and progression of native
17 vessel disease, condition of conduit and
18 progression of intima hyperplastic and atheromatous
19 degeneration, and the introduction of drugs for
20 mitigation of atherosclerotic disease?
22 If these trial designs are inadequate,
23 should randomized, controlled clinical trials be
1 With regard to device placement and device
2 design, please address the following: Given the
3 considerable differences between the proximal and
4 distal CABG anastomoses, what, if any, differences
5 in study criteria should be required?
7 Are there certain aspects of the clinical
8 study design, for example length of follow-up and
9 endpoints, that should be required for all devices
10 irrespective of device form and function? For
11 example, the U-clip performance closely duplicates
12 that of a suture, whereas the Symmetry has greater
13 similarity to a stent.
14 It is rarely possible to determine the
15 cause of conduit failure. Can you suggest criteria
16 to determine whether failure is device related?
18 Number three, do you believe that the
19 significant differences between an arterial conduit
20 and a venous conduit warrant distinct study
21 criteria and assessment for each? If so, please
22 identify these criteria and analyses.
24 Four, should the primary effectiveness
endpoint be graft patency alone, or include both
1 graft patency and myocardial perfusion?
2 Five, with regard to device safety, what
3 criteria, that is, acceptable adverse event rates
4 as compared to that for suture should be applied to
5 the evaluation of device safety as distinguished
6 from device effectiveness? For example, myocardial
7 infarction, reoperations, neurologic events,
8 incidence of aortic complications.
10 Regarding endpoint evaluation, number six,
11 with regard to appropriate patient follow-up, in
12 view of the possible persisting risk of failure of
13 some mechanical anastomosis sites, distinct from
14 the progression of native vessel disease, what
15 duration of follow-up is advisable for premarket
18 Should postmarket follow-up be required to
19 assess long-term device effectiveness? If so,
20 please define the appropriate length of follow-up
21 after primary patency evaluation.
23 The last question, can non-invasive
24 measuring instruments, for example,
echocardiography, ultrafast spiral CT, MRA, EBT,
1 etc., be used for primary assessment of graft
2 patency or is angiographic follow-up necessary? At
3 what time points should patency be assessed? Thank
5 DR. TRACY: Does that conclude the FDA
6 presentation? Does anybody on the panel have a
7 question for the FDA at this point?
8 [No response]
9 At this point, we will move on to the open
10 public hearing. Both the Food and Drug
11 Administration and the public believe in a
12 transparent process for information gathering and
13 decision-making. To ensure such transparency at
14 the open public hearing session of the advisory
15 committee meeting, FDA believes that it is
16 important to understand the context of an
17 individual's presentation. For this reason, FDA
18 encourages you, the open public hearing speaker, at
19 the beginning of your written or oral statement to
20 advise the committee of any financial relationship
21 that you may have with the sponsor, its product
22 and, if known, its direct competitors. For
23 example, this financial information may include the
24 sponsor's payment of your travel, lodging or other
expenses in connection with your attendance at this
1 meeting. Likewise, FDA encourages you at the
2 beginning of your statement to advise the committee
3 if you do not have any such financial
4 relationships. If you choose not to address this
5 issue of financial relationships at the beginning
6 of your statement it will not preclude you from
8 MS. WOOD: I have just a couple of
9 announcements for the open public speakers. We
10 have asked today, due to the number of speakers
11 that have requested time, that you limit your
12 remarks to five minutes each. I would also ask
13 that you provide me with either an electronic copy
14 or a hard copy of your presentation for the benefit
15 of the summary writer and the transcriptionist. If
16 you could see me at lunchtime, that would be great.
17 Thank you.
18 DR. TRACY: There are a number of speakers
19 and I will call them in order. The first is Dr.
20 Randall Wolfe, from University of Cincinnati.
21 Open Public Hearing
22 DR. WOLFE: Members of the panel, ladies
23 and gentlemen, good morning.
25 Thank you for honoring my request to speak
1 before you. My disclosure is that I was the
2 principal investigator on the multicenter U-clip
3 distal anastomotic trial. Those results were
4 presented at AATS two years ago. There is no
5 financial relationship.
6 I was a past consultant for Ethicon in
7 laboratory and clinical evaluation of proximal and
8 distal anastomotic devices, and in the past was a
9 consultant to Ventrica in helping set up their
10 clinical distal anastomotic connector trial.
11 I am currently on the steering committee
12 of the Prevent IV Core Gentech E2F Decoy trial
13 which uses synthetic DNA to prevent aortic coronary
14 venous graft atherosclerosis. That study is closed
15 with over 3,000 patients enrolled. I mention that
16 because I think we are going to be educated on true
17 graft patency of the results of that trial which
18 will be opened first quarter of next year.
20 My primary interest is that I have been
21 presenting summary of anastomotic devices at our
22 national meetings, both AATS, STS and ISMICS. In
23 the next five minutes I would like to summarize
24 some of the things that have been presented at
2 Overall, there are a lot of anastomotic
3 connector devices, and this shows a convenient way
4 to classify these into proximal and distal and
5 subsequently into automated versus manual. There
6 are 13 to 15 different devices in these different
7 categories but I find this a convenient way to look
8 at connectors.
10 There are different value propositions
11 with the connectors and they range from traditional
12 CABG all the way to total endoscopic CABG. I don't
13 have time to go over this in detail but only to
14 point out that there is a possibility of
15 eliminating the heart-lung machine by using certain
16 connectors and also reducing ischemic time. In the
17 endoscopic evaluation there is a potential to
18 reduce patient pain and trauma and to truly enable
19 endoscopic surgery.
21 This is a summary that you will probably
22 hear more about from other presenters, but vein
23 graft failures could be a bad vein; the vein could
24 be too long; it could be too short; there could be
poor run-off bed; or it could be a distal or
1 proximal anastomotic problem.
3 I think this is an important slide. This
4 is some of the science and this is based on some of
5 the work of the E2F Decoy trial but there is an
6 initial wave of inflammation in a venous graft.
7 There is injury. There is activation of smooth
8 muscle cells. There is migration proliferation and
9 intimal soil, if you will, for atherosclerotic
10 plaque and ultimately accelerated atherosclerosis.
11 However, this initial wave is in the first two
12 weeks after the venous graft has been harvested
13 from the leg and placed on the heart.
15 This is a summary of how I look at graft
16 failure. I divided it into three distinct
17 categories. The first is immediate, that is a
18 technical graft failure. These are all venous
19 grafts, by the way; it could be arterial as well.
20 Technical failure would be identified in the first
21 week. In other words, if one obtained a
22 postoperative coronary angiogram in a patient in
23 one week technical failures would be disclosed.
24 The next is intermediate, and this is what
relate to devices. This is usually in
1 six to eight weeks. So, a six-month angiographic
2 evaluation should pick up device failures.
3 The third is chronic and this relates to
4 accelerated atherosclerosis and this really takes
5 years. In the E2F Decoy trial we are looking at
6 one year but, in fact, it probably occurs over five
7 years. In my opinion, if the St. Jude device had
8 been evaluated at six months by angiography
9 stenoses and occlusions would have been discovered
10 that related to the device. In other words, the
11 intermediate category.
13 We now have second generation anastomotic
14 devices. They have proven to be more reliable than
15 hand sewn. There is a consistent orifice size.
16 They are easier to use. I think, importantly,
17 another change that has happened with the second
18 generation is a lack of vein manipulation. So,
19 these should be evaluated with six-month
20 angiographic equivalency and we should also look at
21 performance outcomes.
23 In summary, I believe the science supports
24 six months angiographic data for the intermediate
or device failure area. Proximal
1 devices have demonstrated excellent patency, which
2 will probably be discussed. And, second generation
3 distal devices demonstrate excellent patency. I
4 think we have to keep in mind as we think about
5 this is that unlike stents for coronary-artery
6 disease, these devices do not rearrange plaque
7 morphology. Thank you.
8 DR. TRACY: Thank you. Are there any
9 brief questions for Dr. Wolfe from the panel?
10 DR. EDMUNDS: What data do you have for
11 that last statement?
12 DR. WOLFE: Which part of it?
13 DR. EDMUNDS: The last statement, how do
14 you know that the device doesn't rearrange plaque?
15 I mean I don't know. I would just be interested in
16 your data.
17 DR. WOLFE: The last statement is
18 concerning distal devices. This is assuming that
19 the device is placed to a target site that is
20 relatively free of atherosclerotic debris. The
21 second bullet point is for the proximal devices.
22 The third bullet point is specifically for distal.
23 DR. EDMUNDS: That is what I am talking
24 about. Are you talking about magnetic coupling?
25 DR. WOLFE: Any
type. What I am trying to
1 relate is that stents and anastomotic devices are
2 not equal in that a stent is supposed to rearrange
3 plaque to open up a stenosis. For devices that we
4 are using that is not their purpose. We are not
5 rearranging the plaque. We are connecting,
6 hopefully, a fairly normal vein or artery to a
7 fairly normal coronary distal target.
8 DR. BRIDGES: I have a question about the
9 second bullet point. Can you also inform us what
10 data that is based on?
11 DR. WOLFE: I think that will be presented
12 by others, but I believe that the difference is
13 that stainless steel is stronger, and in a proximal
14 position where there is atherosclerotic disease a
15 stainless steel device can actually hold the aorta
16 open, whereas a nitinol device may not; it may
17 buckle and close. So, it is really the strength of
18 the material. The proximals are different from the
19 distals. In fact, the people that may need the
20 proximal devices the most are the ones who have the
21 worst aortas. They have disease in a situation
22 where it is maybe not safe to clamp the aorta.
23 DR. AZIZ: With the proximal devices, if
24 you do get narrowing, how do you propose that be
handled? Let's say in six months
you find that you
1 have osteal narrowing, how would you handle that?
2 DR. WOLFE: I don't know the answer to
4 DR. AZIZ: Can they be dilated in the cath
6 DR. WOLFE: I don't know the answer to
8 DR. AZIZ: And is there intimal
9 hyperplasia that you are seeing, if you do see it?
10 DR. WOLFE: I believe so with the second
11 generation devices. With the first generation
12 devices I think it was a more complicated situation
13 where the graft could actually embrocate over the
14 device. But in the second generation devices it
15 should be more related to disease in the aorta.
16 However, if a large lumen is maintained then there
17 shouldn't be significant stenosis. So, let's say
18 you get neointimal hyperplasia in every graft,
19 let's say you get a millimeter in every
20 graft--well, if you get a 1.5 mm opening, that is
21 significant. If you get a 3 mm opening that is
22 maintained, it won't be significant.
23 DR. AZIZ: Let me ask you one other thing,
24 with the proximal anastomotic devices, the angle
that the graft comes off is really at right angles
1 to the aorta. Right?
2 DR. WOLFE: In some of the products, that
3 is true.
4 DR. AZIZ: You mean there are ones where
5 you can have it coming off as a cobra head?
6 DR. WOLFE: That is correct.
7 DR. KRUCOFF: Have you actually retrieved
8 any of these devices and looked at them under a
9 microscope when they have failed?
10 DR. WOLFE: I have not--well, I have seen
11 the slides, I certainly have.
12 DR. KRUCOFF: Whose slides are those?
13 DR. WOLFE: St. Jude. I did go over those
14 at one point and, again, that is a first generation
15 device and I believe the mode of failure of that is
16 different from anything you might see in the
17 future. It is multifactorial but the occlusions
18 tend to be flush with the aorta. There is
19 neointimal hyperplasia; there is thrombus. First
20 of all, the angiogram does not look like a typical
21 angiogram that you might see with an occluded vein
22 graft; it is completely different. There is also
23 the possibility that the vein graft itself has
24 changed its position on the connector. In
addition, that was a connector that had a high
1 profile. There is also the possibility that there
2 could be a right angle kink right at the end of the
4 I think in summary, I give credit to the
5 pioneers for being the first ones out there. The
6 first eight patients who received a mitral valve
7 replacement all died. Fortunately, we still do
8 mitral valve replacements and maybe with the first
9 generation connectors we are seeing some of the
10 same things, some of the mistakes. I think many of
11 those have been changed by changes in device and
12 changes in material.
13 DR. YANCY: As you have worked through
14 your clinical trials with these devices, have there
15 been concomitant improvements or changes in medical
16 management because of anticipated problems with
17 these connectors vis-avis antiplatelet therapy,
18 anticoagulation, aspirin, etc.?
19 DR. WOLFE: We do have some data from the
20 E2F Decoy trial. The trial has not opened but we
21 have some demographic data. It has been shown that
22 when patients are followed more closely the chances
23 of them going home on antiplatelet agents are much
24 higher. Although most surgeons say that they send
their patients home on aspirin or some antiplatelet
1 agent, in fact, many patients do not go home on
2 that but in a careful study situation they do.
3 There is a study bias.
4 DR. YANCY: So, those anticipated events
5 that you thought would be predicted or captured at
6 six months, do you think they are product failures,
7 medical management failures or both?
8 DR. WOLFE: I expect they are product
9 failures and they probably would be in an extreme
10 environment such as a very atherosclerotic aorta,
11 but I am not sure. I am not sure.
12 DR. TRACY: I think we are going to have
13 to move on at this point. There is a number of
14 other speakers. Thanks very much. Dr. Robert
16 DR. EMERY: While we are setting up my
17 disc here, I am Robert Emery. I am in private
18 practice in Minneapolis-St. Paul, Minnesota. I am
19 not being sponsored by any companies but I have had
20 relationships in terms of research grants by St.
21 Jude Medical, ATS Medical, AtriCure, Congestive
22 Heart Failure Solutions. I have been on research
23 advisory boards for St. Jude Medical, Medtronic,
24 Myocor, Percardia, CardioGenesis, Inc.; data safety
monitoring boards for Cardioblate and for Myocor,
1 and I have received speaking fees for several of
2 the aforementioned companies.
4 I would like to address our early
5 experience in the Minneapolis-St. Paul area looking
6 at why vein grafts fail, the new issues with aortic
7 connectors. We have been through the etiology of
8 graft failures so I won't go into that, however,
9 there are several new issues that are introduced by
10 the currently used generation of connectors. There
11 can be overloading of the connector, that is, too
12 much vein graft placed below the prongs;
13 double-loading of the connector like putting on
14 your socks where you can invert the graft and load
15 that which inhibits flow through the graft. You
16 can skive the aortic punch and that make take out a
17 complete circle.
18 There are variations in operative
19 technique. For instance, performing your proximals
20 first, as most surgical trainees in the United
21 States perform distals first you are radically
22 changing the way we have been trained in our
23 everyday use in conduct of the operation. Grafts
24 can move. After the patient is closed the lungs
can push the grafts to various positions and this
1 can cause loss of the 90 degree angle, that has
2 been mentioned here, that is necessary for the
3 current generation St. Jude connector.
5 Let's look at some of these issues that we
6 have seen. Here is a surgical technical error at
7 the distal anastomosis that would lead to graft
8 failure if not completed. I don't think that could
9 be blamed on the connector but a connector was
12 This is the first case I performed in the
13 United States, the second one done in the United
14 States after FDA approval. You can see two
15 technical errors here that I learned over time and
16 if I had not changed my operative technique one
17 would have a consistent mode of failure that would
18 be uncorrected. That is, these grafts are placed
19 on top of the aorta instead of further down the
20 side toward the pulmonary artery, therefore,
21 maintaining a 90 degree angle. The grafts are also
22 reflected superiorly with some kinking at the
23 anastomotic site, not maintaining that 90 degree
24 angle. As I mentioned, these grafts can move. All
grafts should be tacked to keep that important 90
1 degree angle. If you lose that you can predict
2 some degree of graft failure.
4 There can be poor run-off, as shown on
5 this slide, to a patent vein graft but a poor
6 distal vessel.
8 Another example is shown here. The graft
9 can be too short, as mentioned. Again, it may be a
10 variation in operative technique.
12 Here a graft is tethered across the
13 pulmonary artery and you can see the narrowing
14 several centimeters distant from the connector
17 And a similar vein here wrapped around the
18 pulmonary artery more tightly than one would like
19 to see.
21 Improper placement of the graft is also
24 Here is a vein graft that was placed on
the right side of the aorta as we traditionally
1 place our saphenous vein grafts when we suture
2 them, rather on the anterior surface of the aorta,
3 riding over the right ventricular outflow tract
4 maintaining the 90 degree angle. You can see the
5 acute bend on the right side as this graft reflects
6 against the patient's pleural surface.
8 Aortic disease was mentioned and this can
9 be important. Here is an occluded connector in a
10 diffusely diseased aorta and you can see, as Dr.
11 Wolfe mentioned, the flush occlusion of the aorta.
13 A combination of factors--here is a small
14 vein graft and poor run-off.
16 And here is a very small vein graft that
17 has become atretic over time to a small distal
18 vessel, still patent through the connector but,
19 nonetheless, narrowed.
21 Then there is the unknown. Here is the
22 occluded connector again flush at the angiographic
25 Here is an occluded vein graft marked by
1 the stainless steel ring in the same patient. You
2 can see the connector graft slightly to the left
3 and one or two centimeters down in this example.
4 There are connector related issues that are key.
6 This is what was addressed a little bit in
7 the prior question, proximal anastomotic problems
8 in the face of appropriate graft and appropriate
9 distal connectors that need to be investigated.
11 Yet, there are technical issues. Here is
12 another proximal connector with a very good vein
13 graft and a large distal run-off system.
15 Improved and more extensive training may
16 obviate several of the modes of failure that we
17 have seen. We need to develop indications and
18 contraindications for the use of these devices,
19 particularly as they come out not just general,
20 overall approval. There are technical
21 considerations that need to be mentioned. Many
22 modes of failure are unstudied or unconfirmed.
23 Thus, prospective studies are warranted including
24 operative technical detail, both visual, such as
the photograph I showed you and verbal operative
1 reports, and improved mentoring may be necessary
2 even for devices that seem intuitively simple.
4 There are tips for success that I have
5 developed in my own practice based on my
8 What we do not want to do is throw the
9 baby out with the bath water because these
10 connecting devices offer us a great opportunity to
11 improve our service to our patients. Thank you.
12 DR. TRACY: Thank you. Are there any
13 brief questions from the panel members? I do want
14 to remind you that there are a lot of people who
15 want to present today.
16 DR. KRUCOFF: Just one question. The
17 angiograms you showed us, were they part of a study
18 protocol that required angiography or were these
19 clinical presentations of people who came back
21 DR. EMERY: These were clinical
22 presentations in approximately our first eight
23 months of use, from May, 2001 through the first
24 eight months, and we have seen very few since we
have modified our surgical techniques.
1 DR. KRUCOFF: And the denominator for
2 these eight months?
3 DR. EMERY: It was about 160, and these
4 are not all of them. These are representative
5 samples of technical errors that are correctable
6 with proper training and changing of your
7 techniques as you learn the process.
8 DR. AZIZ: When you say you tack the
9 grafts, are you putting many anchoring stitches or
10 what do you do?
11 DR. EMERY: Three or four generally on the
12 left side. I put one on the pulmonary artery and,
13 again, depending on the length of the graft,
14 because you are doing proximals first with this
15 device, I will connect it so that it won't move
16 with respiration. On the right side I connect it
17 down the body of the right ventricle as the graft
18 goes directly up from the aorta over the right
19 ventricle and down to the right coronary artery,
20 the posterior descending artery, just some 6-0
21 prolene suture tacking.
22 DR. AZIZ: We normally do the regular
23 suturing technique; usually you don't have to do
25 DR. EMERY: No,
1 DR. AZIZ: So, why do you think you need
2 to do it here?
3 DR. EMERY: Because I learned doing this
4 distal first and I think my measurement of the
5 length of the vein graft to the aorta is better on
6 a distal first process in my hands. So, sometimes
7 I would rather make my grafts too long than too
8 short because the shortness of the graft may be one
9 reason for disconnection of these connectors from
10 the aorta. As the pulmonary artery fills, if the
11 graft is too short you can pull these off. I have
12 pulled them off myself in the operating room by
13 tugging a little bit too hard and I had to put my
14 finger over the hole. So, a short graft can lead
15 to connector displacement from the aorta,
16 particularly as the patient moves or the heart
17 fills in the postoperative period.
18 DR. TRACY: Dr. Hirshfeld?
19 DR. HIRSHFELD: I would just like to say
20 as an angiographer who has probably taken pictures
21 of thousands of bypass grafts, I have heard a lot
22 about considerations that I was never aware of
23 before from your brief presentation, and I think it
24 calls for a sharing of information between
angiographers and surgeons about many of these
1 technical considerations that affect graft
2 performance. So, I would hope that out of this
3 will come that kind of sharing of information.
4 DR. EMERY: I have reviewed all the
5 angiographs of patients that failed in my hands.
6 DR. TRACY: I think we have to move on; we
7 have a number of speakers. I am sorry to cut this
8 short; it is very interesting. To remind you,
9 there will be more time this afternoon to discuss
10 things in detail. Dr. Schoettle?
11 DR. SCHOETTLE: Good morning. My name is
12 Dr. Phillip Schoettle. I am a thoracic and
13 cardiovascular surgeon in practice at Methodist
14 University Hospital in Memphis, Tennessee.
15 I am here this morning to discuss my
16 experience with the Symmetry proximal anastomotic
17 device. I would like to disclose at the outset
18 that I have no financial interest in this matter.
19 I paid my own way to Washington, and I am not
20 employed by anyone, nor intend to be employed by
21 anybody with a financial stake in this issue.
22 In September of 2001 I was trained in the
23 use of the Symmetry proximal anastomotic device,
24 along with two of my scrub assistants, by St. Jude
Medical. I was attracted to the device
1 the reasons mentioned previously which would allow
2 you to do a proximal anastomosis off the aorta
3 without the use of a partial occluding or
4 side-biting clamp with its attendant risk of
5 embolic debris.
6 I rapidly incorporated that device into my
7 practice and used it almost exclusively for the
8 next eleven months. Initially I was very pleased
9 with the results. I had minimal, if any, technical
10 issues with the device and was not aware of any
11 acute or subacute saphenous vein closures.
12 Unfortunately, at approximately ten months we began
13 to see almost a deluge of patients returning to the
14 cardiac catheterization laboratory with vein graft
15 occlusions or high grade stenoses invariably
16 occurring in the connector site.
17 This occurrence was totally incompatible
18 with my previous surgical experience. I reported
19 this to St. Jude Medical and I felt like it
20 warranted a distribution to the surgical community
21 and I began a review of my patients, resulting in
22 the paper that you see here. This paper was
23 entitled, "Use of an Anastomotic Device in Coronary
24 Bypass Surgery: A Word of Caution." It was
published in the January edition of the Journal of
1 Thoracic and Cardiovascular Surgery.
2 Without going into great detail, I would
3 like to summarize the results of that paper. It
4 was a review of two years of experience. The first
5 year was the year prior to my beginning to use the
6 device, while proximal saphenous vein connections
7 off the aorta were done in the conventional manner,
8 that is, hand-sewn with a partial occlusion clamp.
9 Beginning in September of 2001, for the next eleven
10 months, comprises the next group of patients where
11 almost exclusively all proximal anastomoses were
12 done with the St. Jude Symmetry anastomotic device.
13 I divided the group in group A and group
14 B. Group A was the first group, the prior year
15 with hand-sewn anastomoses. I reviewed all
16 patients who required repeated cardiac
17 catheterization after coronary artery bypass
18 surgery. What we found was that even though the
19 patients in group A had had a year longer of
20 exposure to my cardiology colleagues, less of those
21 required repeated cardiac catheterization, although
22 that number was not significant between the groups.
23 The number of grafts studied between group
24 A and group B was also similar. However, what we
did find was that the group A patients, those with
1 hand-sewn anastomoses, had an 80 percent patency
2 rate of vein grafts studied. Remember, these were
3 symptomatic, or at least theoretically symptomatic
4 patients. So, 80 percent of the grafts were widely
5 patent in the hand-sewn anastomoses, with no
6 significant stenoses, and 20 percent were
8 Unfortunately, in the group B patients,
9 those with the Symmetry proximal anastomotic
10 device, only 20 percent of the grafts studied were
11 patent. Fully 80 percent of the grafts were either
12 totally occluded or had high grade stenoses
13 uniformly occurring at the connector site. The
14 significance in p value in favor or patency of the
15 hand-sewn anastomoses was standardly evaluated with
16 a p value of 0.0001.
17 Based on my experience with the Symmetry
18 proximal anastomotic device and review of my own
19 patients, I have several observations and two
20 conclusions I would like to make. The use of the
21 Symmetry St. Jude proximal anastomotic device in
22 its current generation results in a significantly
23 higher saphenous vein closure and occlusion rate
24 when compared to hand-sewn anastomoses.
25 I do not believe that technical issues are
1 the major factor. I can show you arteriograms of
2 what appear to be perfectly laid out saphenous vein
3 grafts with a 90 degree angle off the aorta, with
4 no kinking, where the stenosis arises immediately
5 in the connector site off the aorta.
6 In two patients that I reoperated, I was
7 able to harvest the segment of aorta with the
8 connector and the saphenous vein. This was looked
9 at microscopically by the pathologists in my
10 hospital who reported basically a foreign body
11 reaction in the connector site with associated
12 neointimal hyperplasia.
13 I would also point out that these
14 connector stenoses and occlusions are not
15 clinically insignificant. In this group at least
16 six patients have required early reoperation.
17 Thirty patients, over a year ago, required PCI
18 stents and angioplasty. There have been four
19 sudden deaths in these patients, two of which were
20 almost certainly related to myocardial infarction.
21 If I can have the liberty of making a
22 conclusion, I see no clinical indication for the
23 current generation of the St. Jude proximal
24 connector. The use of this connector or any other
vascular anastomotic devices must be evaluated by
1 scientifically controlled, prospective clinical
3 I do not believe that uneducated surgeons
4 and uninformed patients should be the testing
5 ground for these devices that have not proven to be
6 clinically safe or effective. I clearly am not
7 opposed to technological advances in coronary
8 bypass surgery. I have been an early proponent of
9 off-pump surgery and less invasive coronary
10 surgery. I do not want to throw the baby out with
11 the bath water. I do not believe, however, that
12 the cause of less invasive coronary artery bypass
13 surgery is furthered by the ill-advised use of
14 these unproven devices. Thank you. I would be
15 glad to answer questions if there is time.
16 DR. TRACY: Any brief questions? Dr.
18 DR. WHITE: Would you just clarify for me,
19 in the early part of your statement you said
20 something about follow-up at ten months. Was there
21 a ten-month interval that was special to you?
22 DR. SCHOETTLE: No, I believe it just
23 would have become apparent to me, you know, with
24 just the overwhelming evidence of patients. All of
sudden I was getting call after call from these
2 DR. WHITE: Would six months not have
3 identified these patients? Would a six-month
4 follow-up, do you think, not have been adequate?
5 DR. SCHOETTLE: I was asked that question
6 last night. I don't have that answer. My gut
7 feeling is that six months would probably be
8 appropriate but I don't know that answer based on
9 this review.
10 DR. BRIDGES: I have a question. In the
11 brief study that you gave us I didn't see the
12 mortality in the two groups. You said that there
13 were no sudden deaths in the hand-sewn group but
14 what was the overall mortality in the two groups
15 and are there any updates since this paper was
17 DR. SCHOETTLE: The operative mortality
18 was less than three percent but the overall
19 mortality long-term, I don't have that; there have
20 been no updates at this point although I intend to
21 do that.
22 DR. BRIDGES: But both groups--
23 DR. SCHOETTLE: They were very similar.
24 DR. BRIDGES: At least for the graft
connector patients, what would be medical therapy
1 for these?
2 DR. SCHOETTLE: It is in the paper, but
3 all patients were discharged on aspirin and all
4 patients were discharged on Plavix for two months.
5 DR. AZIZ: But when you had to reoperate
6 on them did you have to redo the whole anastomosis
7 or could you immobilize it and rehook it? How did
8 you do that?
9 DR. SCHOETTLE: In a couple of cases I was
10 able to continue to use that vein segment. Conduit
11 length was an issue. Several of the veins were
12 totally occluded and we just had to sacrifice those
14 DR. AZIZ: So, the orifice was like the
15 whole length?
16 DR. SCHOETTLE: That is correct.
17 DR. AZIZ: So, you probably had intimal
18 hyperplasia proximally and you had full flow and
19 then thrombus--
20 DR. SCHOETTLE: And then thrombus
21 distally, correct.
22 DR. AZIZ: Obviously most people don't,
23 and they probably should use some flow techniques
24 to measure flows.
25 DR. SCHOETTLE:
All patients in both
1 groups had mediastinal transit time flow evaluation
2 at the time of surgery, and 95 percent of cases
3 were done off-pump.
4 DR. AZIZ: When you did proximal
5 anastomoses with the device did you do any
6 sequential grafts--
7 DR. SCHOETTLE: No, they were sequential
8 grafts but I don't have that number available to
10 DR. TRACY: Thank you.
11 DR. SCHOETTLE: Thanks.
12 DR. TRACY: Dr. Frater?
13 DR. FRATER: Let me state immediately I am
14 the Medical Director of St. Jude and, obviously,
15 have that as a conflict of interest.
16 I have a few points to make. I had
17 expected ten minutes so I am going to try and make
18 them quickly. I think we can all agree that the
19 MDR system is a warning light that tells us nothing
20 about incidence and, unless we are very lucky,
21 doesn't give us much information on causality, hard
22 as we try to look into every single report that
23 comes in from the field. I shall not elaborate on
24 that. I suspect the FDA feels the same about the
MDR's utility as we do.
1 The question of comparing anastomotic
2 devices to historically published data for sutures
3 is an interesting one. The data that was obtained
4 in the past has been cardiac surgeons who were
5 trying to find out what they were doing 30 years
6 ago when they were making venous anastomoses. The
7 patients were younger. The vessels were better.
8 The extra conditions, such as diabetes, were far
9 less common and it was a different group of
10 patients. Those patients have long since been
11 captured by the interventionalists and the cardiac
12 surgeon today faces a very different patient.
13 We need to know what the patency rates are
14 today with the current set of patients. We also
15 need to know what the difference may be between
16 off-pump and on-pump. There was a paper presented
17 just a few weeks ago at the ACC, the so-called Prog
18 IV Trial, a randomized comparison between off-pump
19 and on-pump surgery with angiography at one year.
20 The patency rate of the cases performed on-pump at
21 one year was 59 percent; the patency rate of those
22 performed off-pump was 49 percent.
23 There is a paper being published in The
24 New England Journal of Medicine by Kahn. It came
out of Britain. Again, a
randomized study of
1 on-pump and off-pump anastomoses studied at three
2 months by angiography, which was performed in 80
3 percent of the patients in the trial. The patency
4 rate of the on-pump cases was exemplary. At three
5 months they had a 98 percent patency rate but the
6 off-pump cases had an 88 percent patency rate at
7 three months.
8 I present this material, which is clearly
9 important in trying to assess what will be the
10 target of patency that we will be looking at in
11 future trials, and a recognition that times have
12 changed and circumstances are clearly very
14 We have done a meta-analysis of some 7,000
15 patients in which angiograms were done between 6
16 and 12 months. We chose that 6- and 12-month
17 period for the obvious reason that you have already
18 heard today, that after 12 months atherosclerosis
19 dominates the failure of vein grafts. The mean
20 occlusion rate was 16 percent in this meta-analysis
21 between 6 and 12 months of sewn anastomoses. But
22 the range was from 9.5 to 26.5. There is an
23 immense diversity from different institutions and
24 we can speculate forever, certainly not in five
minutes, as to what the reasons for those
1 differences are. I am sure that the surgeons in
2 the Prog IV study have not suddenly become
3 incompetent; there are factors that we need to look
5 The question of the extent to which
6 clinical utility data is considered to be
7 necessary, I think we have already begun to deal
8 with this. Six months seems to be a period of time
9 that people are reaching, and that is not
10 unreasonable considering that stents are a Class
11 III device which may or may not be identical to
12 anastomotic devices--that is debatable--are being
13 evaluated with MACE and target vessel interventions
14 at six months.
15 Certainly, it is reasonable to state that
16 it should not be more than 12 months for the
17 obvious reason that by then atherosclerotic disease
18 dominates. There is intimal damage and technical
19 factors in the first week, neointimal hyperplasia
20 for the next few months, blending finally into
21 atherosclerotic disease.
22 Now, it is essential that the FDA provide
23 clarity on the type of clearance that we need. If
24 the clinical data requirement reaches the point
which would normally be required for a PMA, then it
1 should be a PMA. If you are required to produce
2 the data for a PMA, then the process should be done
3 under a PMA process. Thank you very much.
4 DR. TRACY: Any questions?
5 DR. WEINBERGER: I have a question for
6 you. You said that you were concerned that the
7 follow-up should be at six months because you
8 thought that atherosclerosis dominates the
9 subsequent natural history of graft failure. I am
10 concerned because we have a pretty good idea that
11 there is distinct biological heterogeneity in
12 different vascular beds in terms of the kinetics of
13 responses to manipulation. For instance, we know
14 that for coronary interventions basically at six to
15 nine months the process is over. But if you look
16 in the periphery, like the iliacs, the usual time
17 is three years. Do you have any data to suggest
18 that the process to response to injury in vein
19 grafts is over at nine months?
20 DR. FRATER: Well, if you look at the data
21 from peripheral vascular intervention where it is
22 far easier to follow the patients, it seems fairly
23 definite that while there is an acute phase, which
24 is partly technical and partly because of the
damage we do to the vein by the various things we
1 do when we take it out and manipulated it, it
2 starts in the first week. The neointimal
3 hyperplasia seems to blend at 12 months in these
4 peripheral vascular studies with the
5 atherosclerotic process. It would seem reasonable,
6 if there is an atherosclerotic process taking place
7 in veins after 12 months, not to attribute that to
8 how we handled the vein at the time of the initial
10 DR. WEINBERGER: Just one follow-up, if
11 there is any kinking in the vein and you have a
12 jet, that jet wouldn't lead to an accelerated
13 atherosclerotic process later on as well?
14 DR. FRATER: It would happen far quicker
15 than that. Usually, if you leave a kink in a vein
16 there is a consequence that is soon and definite.
17 DR. WEINBERGER: Data?
18 DR. FRATER: Data? Clinical experience.
19 I am a cardiac surgeon.
20 DR. TRACY: Dr. Bridges?
21 DR. BRIDGES: Yes, there are two points.
22 One is to echo Dr. Weinberger's point that I don't
23 think we know exactly. There is nothing to suggest
24 that there can't be an interaction between
mechanical factors and atherosclerosis that extends
1 beyond one year. To say that you can divide these
2 into two discrete processes that are technical,
3 device related and then atherosclerosis I think is
4 unsubstantiated and you can't really defend that.
5 Furthermore, I am sure we are going to
6 hear from Dr. Mack but his own data that was
7 presented at the STS meeting, just in January,
8 showed, at least in his series which I am sure he
9 will comment on, that it was not until you got out
10 beyond one year that you started to see a
11 difference in MACE, that is, you know,
12 cardiovascular events. So, that, in and of itself,
13 also suggests that the idea of only looking at a
14 one-year or six-month time period is clearly going
15 to result in us missing failures.
16 DR. FRATER: The obvious issue is how long
17 would you like it to be? Clearly, it becomes
18 extraordinarily difficult if you are suggesting
19 that we should wait five years, or something like
20 that. Dr. Mack can speak for himself but I believe
21 that in diabetes there was a difference and there
22 may well be factors like that that make a
24 DR. BRIDGES: My point is not to suggest
how long we need to look, I am simply objecting to
1 the concept that we can definitively or
2 declaratively state at this point, based on what
3 evidence we have, that we know that six months or
4 nine months is an acceptable time frame in order to
5 exclude device-related issues.
6 DR. TRACY: Dr. Yancy, and then if there
7 is time Dr. Maisel.
8 DR. YANCY: Just a very short yes/no
9 question. I have not seen the referred to NEJM
10 article comparing on-pump versus off-pump surgery.
11 Were connectors used in the off-pump cases?
12 DR. FRATER: This was absolutely a study
13 of on-pump versus off-pump vein patency.
14 DR. TRACY: Dr. Maisel?
15 DR. MAISEL: You have eloquently stated
16 that times have changed and that historical
17 controls are just that, historical, and you stated
18 data that the patency rates vary greatly from
19 institution to institution. In many respects that
20 is a strong argument for randomized trials but you
21 didn't come out and state that. Are you a
22 proponent of randomized clinical trials to assess
23 these devices?
24 DR. FRATER: You know, in the best of all
possible worlds, yes. I am
speaking as a cardiac
1 surgeon now.
2 DR. TRACY: Thank you.
3 DR. EDMUNDS: Did I hear you say that the
4 one-year patency rate for off-pump proximal veins,
5 the occlusion rate was 9-26 percent?
6 DR. FRATER: In the meta-analysis that we
7 did between 6 and 12 months, there was a range from
8 9.5 to 26.5 percent in this meta-analysis of some
9 7,000 cases. In the Prog IV study--
10 DR. EDMUNDS: For hand-sewn?
11 DR. FRATER: Hand-sewn anastomoses, in the
12 Prog IV study the patency rate was 59 percent
13 patent at one year on pump, 49 percent patent
14 off-pump. It was just presented at the ACC.
15 DR. TRACY: We do have to move on, I am
16 sorry. Dr. Mack?
17 DR. MACK: My name is Michael Mack and I
18 am a cardiac surgeon in Dallas. By way of
19 disclosure, I am not sponsored by anybody today. I
20 paid my own way here. I have served as a
21 consultant in the past at St. Jude, also to
22 Cardica. I have received research grant support
23 from St. Jude regarding anastomotic devices, and I
24 am also on the scientific advisory board for
Medtronic and Guidant, both of which have equity
1 interests in anastomotic device companies.
2 The thrust of my presentation was to
3 discuss saphenous vein graft patency and not our
4 own St. Jude device paper, which has been
5 presented, in view of the fact that I only have
6 five minutes but I will try and get this done in
7 four and just spend the last minute discussing
10 Specifically, what I would like to discuss
11 is I thought until I looked at all this that I knew
12 what the gold standard for saphenous vein graft
13 patency was. Eeverybody throws around numbers but
14 until I did a meta-analysis of the literature I
15 really didn't know, and this is specifically to
16 address the trial design question number one of the
17 FDA, the gold standard of sutured anastomoses had a
18 well-documented history of over the past thirty
19 years, and I would like to go over those thirty
20 years right now--
22 --or the why of saphenous graft failure in
23 five minutes.
25 Since between 1979 and 2001 there have
1 been thirty studies published, analyzing a total of
2 28,081 grafts.
4 Factors impacting the studies are
5 angiogram survivors. You lose early graft
6 occlusions resulting in death so, therefore, you
7 are automatically losing patency in any angiogram
8 series because you can only angiogram survivors.
9 Studies are impacted by the completeness of
10 follow-up, the percent of patients actually
11 undergoing angiograms, and whether the study was
12 done as a surveillance study or done for cause.
14 If we look at a meta-analysis of all
15 studies that looked at 30 days or less, there has
16 been a total of 11,000 grafts looked at. If you
17 just skip to the number at the bottom right, the
18 patency rate at 30 days in these 7 studies,
19 comprising 11,000 grafts, is 87.8 percent.
21 If you now look at 3 to 6 months and look
22 at the 10 studies published here, with a total of
23 2,290 grafts, at 3 to 6 months 84 percent is the
24 saphenous vein graft patency.
1 If we now go to 12 months and look at the
2 13 studies comprising almost 12,000 grafts, the
3 patency rate is 82.7 percent in the literature.
5 Lastly, if we look at 2 to 5 years, with a
6 total of 3,100 grafts in these 3 studies, the
7 patency rate between 2 and 5 years is 74.3 percent.
9 If we summarize all this, there is a
10 significant attrition in the literature of about 12
11 percent of vein grafts in the first 30 days.
12 Between 30 days and 3-6 months another 3 percent of
13 grafts are lost at that point. If we go between
14 3-6 months to 12 months another 1.5 percent of
15 grafts are lost. Then there is a slightly greater
16 attrition from 2 to 5 years. If you look at
17 overall graft patency of all 28,000 grafts done at
18 any time, it is 84 percent.
20 Variables known to impact graft patency
21 include age, gender, diabetes and how well the
22 diabetes is controlled, obesity, which vessel is
23 bypassed, the LAD, the circ. or the right, the
24 target vessel size, the presence of distal disease,
the size of the vein graft, harvest injury, whether
1 an endarterectomy was done, what the graft flow at
2 time of implant was, individual versus sequential
3 vein grafts, how much myocardium was supplied, what
4 the ventricular function of the patient was,
5 whether lipid management was tightly controlled,
6 whether antiplatelets were used, surgeon
9 Variables that are not known how they
10 impact on graft patency has been alluded to.
11 Whether it is done on a beating heart or an
12 arrested heart. That recently has been called into
13 question. And whether anastomotic connectors
14 impact positively or negatively on the saphenous
15 vein graft patency.
17 I think how you design your study you can
18 get 100 percent patency at ten years if you do an
19 LAD, do it as a sequential and a three millimeter
20 target with no distal disease, use a small vein,
21 have a large run-off in a thin male that does not
22 have insulin-dependent diabetes and normal ejection
23 fraction, does not have a hypercoagulable state, is
24 on antiplatelet agents and is well controlled with
statins. On the other hand, you
can do the
1 converse of all those and you will end up with a 10
2 percent patency rate in less than 30 days.
4 In conclusion, I think that many variables
5 other than anastomotic connectors impact graft
6 patency. Angiography is the only reliable method
7 to determine patency. Meta-analysis reveals an
8 overall saphenous vein graft patency of 80-85
9 percent. There is no significant difference from
10 3-6 months versus 12-16 months or, for that matter,
11 even between 30 days and the latter two endpoints.
12 An angiographically normal graft at the earlier
13 study times is often likely to develop occlusion on
14 later follow-up and, in my opinion, a 6-month
15 angiographic endpoint is adequate to evaluate graft
16 patency with anastomotic devices.
17 Real quickly regarding our experience, it
18 has been published on the St. Jude device. We did
19 find that there were events that happened after six
20 months. These were clinical events. The study
21 that we performed was very similar to Dr.
22 Schoettle's. We took a one-year experience with
23 the St. Jude device and compared it to one-year
24 previously with a similar cohort of patients and
found that there was a higher incidence of clinical
1 events in the St. Jude patients. However, these
2 were all limited to diabetics. We looked at the
3 non-diabetic population and we looked at all
4 possible variable by logistic regression diabetes
5 was the only thing that sorted out. The
6 confounding variable in all this is that all these
7 procedures were also done on a beating heart.
8 Thank you.
9 DR. TRACY: Thank you. Panel, you have 4
10 minutes and 36 seconds to ask questions. Any
11 questions? Dr. Weinberger?
12 DR. WEINBERGER: In surgical literature
13 everyone seems to focus on patency. Are you
14 interested at all in morphology, like quantitative
15 angiography looking at 30 percent stenosis, 40
16 percent stenosis? Is that information valid to
18 DR. MACK: Absolutely. Because I do think
19 that that is a precursor of potential total
21 DR. WEINBERGER: And if that is the case,
22 are your angiographic colleagues who have looked at
23 these connectors able to assess the morphology
24 right around the metallic connector adequately?
25 DR. MACK: I
think the answer is yes.
1 DR. KRUCOFF: Not being as familiar with
2 the surgical literature, in this list you sort of
3 ended with do you think there is sufficient data to
4 create a real propensity score in planning a trial?
5 DR. MACK: Yes.
6 DR. KRUCOFF: To actually create risk
7 categories that could be sufficiently evaluated in
8 new populations?
9 DR. MACK: Yes, I do. I think that
10 everything I listed there--one study or another has
11 listed those factors implicating graft patency and,
12 yes, I think you can develop a propensity score.
13 DR. TRACY: Dr. Edmunds?
14 DR. EDMUNDS: Mike, you said that all of
15 these were off-pump bypasses.
16 DR. MACK: In our St. Jude experience,
18 DR. EDMUNDS: Were they mostly right
20 DR. MACK: First of all, we did not have
21 any connectors placed to the LAD so they all went
22 to diagonal circumflexes or right, and which vessel
23 it went to, in our experience, did not sort out as
24 a factor.
25 DR. EDMUNDS:
1 DR. MACK: All non-LADs.
2 DR. EDMUNDS: And these were surveillance
3 angiograms, not for symptoms?
4 DR. MACK: No, the surveillance was
5 clinical events only. The only angiograms--
6 DR. EDMUNDS: So, you have bias towards
7 symptomatic patients.
8 DR. MACK: The endpoint was not
9 angiography. The endpoint of our study was
10 clinical events, major adverse events at now two
11 years of follow-up. We did not do a specific study
12 angiogramming the patients. The only angiograms we
13 had was in patients that were done for cause.
14 DR. EDMUNDS: The 28,000 patients were
15 from 30 studies, weren't they?
16 DR MACK: Okay, I am mixing up your
17 question then. Ask again, Hank.
18 DR. EDMUNDS: Well, the cohort of 28,081
19 angiograms was from 30 papers--
20 DR MACK: Right.
21 DR. EDMUNDS: --and were those
22 surveillance angiograms or for symptoms?
23 DR. MACK: I am sorry, I thought you were
24 talking about our own experience with the
1 DR. EDMUNDS: No, I am sorry, Mike.
2 DR. EDMUNDS: No, all of those were
3 surveillance. Any that was done for cause and I
4 did not include in that. All those were
5 surveillance studies. Similarly, there were a
6 couple of other studies that looked at just
7 saphenous vein graft, the LAD, I did not include
8 those because those were abnormally high.
9 DR. TRACY: Dr. Bridges, did you have a
11 DR. BRIDGES: My question really is that
12 given that in the results you presented recently
13 there was a difference in major adverse
14 cardiovascular events, I guess in the manuscript
15 that I have seen a draft of it was limited to
16 diabetic patients. However, those were non-insulin
17 dependent diabetics, I believe, and I was wondering
18 if you had a hypothesis as to why non-insulin
19 dependent diabetics would be different than insulin
20 dependent diabetics. Given that, should we be then
21 separating diabetics from everyone else in terms of
22 determining the applicability of these devices?
23 DR. MACK: That is an excellent question,
24 and we were a little bit surprised to find that
that was the case also because from the stent
1 experience you would expect it would be more so in
2 insulin dependent diabetics but such was not the
3 case. We have hypothesized that perhaps it was due
4 to the fact that with non-insulin dependent
5 diabetic oral agents the blood sugar is not as
6 tightly controlled, but we have no proof; it is
7 total hypothesis.
8 I also think that the way that we look at
9 diabetes now, today, is totally blurring the line
10 between insulin dependence and non-insulin
11 dependent diabetics. I think we have a lot of
12 metabolic syndrome patients who are actually Type 2
13 diabetics but are insulin dependent and we are
14 actually categorizing them as insulin dependent
15 when, in fact, they really should not be.
16 DR. TRACY: Thank you. Dr. Slaughter?
17 DR. SLAUGHTER: Thank you. I was asked to
18 speak today on behalf of Converge, and I am a U.S.
19 investigator for their ongoing trial for distal
20 anastomotic studies and they did pay my travel here
21 but I have no other financial relationship with
24 To date so far we have heard predominantly
about proximal anastomotic devices and what I would
1 like to do is to tell you a little bit about a
2 current and ongoing look at distal anastomotic
5 Certainly, this comes up in many issues
6 and I don't think we need to belabor the fact but
7 perhaps at the end I will comment briefly on some
8 of the other questions asked, but there is still no
9 question, and it is really sort of one of the
10 unspoken issues for any outcome for the patient,
11 and that is, you know, the quality of anastomosis
12 and the overall revascularization and long-term
13 patency. Certainly surgeon skill is very
14 important. There are also the other issues of the
15 anatomy, disease state, access and visibility that
16 would affect these things. But all these things
17 are very important in determining not only acute
18 but long-term graft patency and the overall outcome
19 for the patient.
21 This has been brought up now several times
22 and I think is very important. This is just
23 another way of presenting it. It is looking at
24 sort of the time scale injury. That is, as was
brought up by Dr. Weinberger as well, there is no
1 question that there is good information and good
2 data as to the initiation of the injury,
3 inflammation and then subsequently intimal
4 hyperplasia. As a rule of thumb, the idea is that
5 within, say, six to eight weeks the injury has
6 stopped. I don't think anybody in their right mind
7 would argue that there is not heterogeneity and
8 certainly there are differences within patients.
9 Certainly that would show up as stenosis and
10 changes in morphology, as you mentioned.
11 But the idea is there is reasonably good
12 science and information to suggest that within
13 about 60 days a vascular anastomosis has healed,
14 particularly within the coronary-arterial tree.
15 So, beyond that time, if there are graft failures,
16 the question is what are they due to, and it is
17 generally due to ongoing atherosclerosis, intrinsic
18 patient factors and/or perhaps a lack of medical
19 therapy such as antiplatelet agents, aspirin and/or
21 So, you know, if hand-sewn anastomosis is
22 so perfect, why are we here today? The issue is
23 they are not perfect and there certainly is room
24 for improvement. Certainly, by hand sewing in a
bad distal vessel it is calcified in a diabetic.
1 They have lateral calcification. By piercing them
2 with needles--we all had that experience, you end
3 up with plaque rupture. You have hemorrhage within
4 the media. The idea is this is a traumatic event.
6 The other is reliability. The issue is
7 how can you do it day to day, 20,000 a year. The
8 idea is you want to make it as reliable as
9 possible and it needs to be reproducible between
10 different surgeons at different institutions.
11 The other is it must be reversible. The
12 idea is if you don't like it you have to cut the
13 suture, take it out and redo it. You want to be
14 able to do the same thing, perhaps in a less
15 traumatic fashion, with a coupler device.
16 The other is it must be easy to use. The
17 idea is if anybody walks up to the podium and is
18 giving you a talk, they basically should all be
19 able to have the same results without any
20 significant extensive training.
21 The other is I think we do need to realize
22 there are differences between proximals and
23 distals. I don't think we need to spend a lot of
24 time on this today but the main two differences are
the flow dynamics which clearly are different at
1 the proximal and distal ends, as well as tissue
2 characteristics. On the tissue characteristics, on
3 the right it is either going to the aorta or vein
4 aorta or artery depending on which you conduit you
5 use. Certainly for distal anastomoses what you are
6 looking at is vein to coronary artery or an
7 arterial conduit to an artery but it is a very
8 different scenario.
9 Also, with flow dynamics there is no
10 question that the size or the shape of the opening
11 or the angle of the take-off is very important, the
12 pressure differential, as well as the vessel
13 diameter throughout the length.
15 I think one other issue which hasn't been
16 brought up today which does need to be mentioned,
17 at least just to bring it up, is actually the type
18 of material. I think this sort of goes into the
19 heterogeneity or perhaps ongoing injury to intimal
20 hyperplasia. These are not new materials. They
21 have all been used before. They have all been used
22 in intravascular scenarios and the idea is there is
23 good evidence to suggest, whether it is nitinol,
24 stainless steel, titanium, that they are
compatible, and I don't think that we can sort of
1 imply or say that they are intrinsically the source
2 of perhaps later stenoses or some ongoing failures
3 beyond the eight-week time period. Certainly there
4 is the heterogeneity of healing in some patients
5 but it is a relatively small number. It is like a
6 cheloid. Some patients get cheloids but not all.
7 The answer is you see it as it progresses. If you
8 follow them and you look for it you can identify
9 who those patients are.
11 I would like to just show you a histologic
12 series which I think is interesting in helps people
13 visualize. Really the sort of best description I
14 think for the Converge distal anastomotic device is
15 that of sort of a compression clip. The idea is it
16 is two frames which are expandable. In the upper
17 right it sort of gives you the diagrammatic picture
18 of a graft into the artery. The important thing
19 here is that you now are able to mechanically
20 manipulate flow dynamics as well as other
21 engineering aspects so you get a perfect 30 degree
22 take-off; you get perfect dynamics. You won't get
23 turbulence at the site of the anastomosis.
24 The left side shows the bypass graft,
which is CABG going down to the circumflex artery.
1 I think the important thing here is this was done
2 at 90 days but, once again, the idea is it is
3 completely endothelialized so the idea is if you
4 get an angiogram at six months and you have a
5 normal lumen you have no narrowing. The idea is
6 are you going to have ongoing intimal hyperplasia
7 that would be an unexpected finding? I think the
8 answer is no.
9 DR. TRACY: If you could finish up in the
10 next few sentences.
11 DR. SLAUGHTER: Sure, I can finish up in
12 about 30 seconds.
14 The idea is you see very clearly that it
15 is a well healed anastomosis and you have the
18 This has already been brought up. The
19 idea is are historical controls acceptable? I
20 think the answer is yes.
22 There is no question there is lots of
23 existing data. We have also lots of information to
24 suggest not only at seven days but at years out
that you can evaluate intimal hyperplasia.
2 Certainly angiography--we know the causes
3 of failure, early failure and what we need to do is
4 differentiate between a device failure and ongoing
7 I will just show--
8 DR. TRACY: I am sorry, we are just going
9 to have to cut this off if we are going to have
10 time for questions from the panel.
11 DR. SLAUGHTER: I apologize.
12 DR. TRACY: We have three minutes left for
13 questions from the panel. Anybody? Dr. Hirshfeld?
14 DR. HIRSHFELD: I would just point out
15 that in the coronary stent experience if we used a
16 two-month follow-up we never would have discovered
18 DR. TRACY: Dr. Krucoff, did you have a
20 DR. KRUCOFF: I would just also say that
21 in the stent experience I think if we started with
22 historical controls based on lung literature, we
23 would have left a lot of important information out.
24 DR. SLAUGHTER: I think the one difference
though, and this has come up I think in other
1 discussions with the FDA panel, is that although it
2 uses a similar material and it is stent-like, it is
3 not a stent. The idea is it is just the edges that
4 are present along the edges of the coronary artery.
5 It is not compressed plaque and the idea is it is
6 very different. It is really sort of a compression
7 clip that applies the vein graft to the distal
8 coronary artery.
9 DR. TRACY: Dr. White?
10 DR. WHITE: I think there is no evidence
11 for that, and I think everything that we have heard
12 today sounds like it is a stent, although a stent
13 in a graft. So, the question would be if you don't
14 believe it is a stent, then you should show us data
15 that suggested that intimal hyperplasia within the
16 tube is not the primary cause of these closures.
17 DR. SLAUGHTER: Sure.
18 DR. YANCY: And because of that, I think
19 it is even more important to state that historical
20 controls would be really problematic I think.
21 DR. TRACY: Any other comments from the
23 [No response]
24 Thank you. Is Mr. Lotti here?
25 [No response]
1 We will move on then to Dr. Martin.
2 DR. MARTIN: Good morning. As so many
3 members of the panel have already suggested,
4 including Dr. White, I can make my comments brief.
5 My name is Dr. Frank Martin. I am Chairman of the
6 Department of Cardiology at Methodist Care in
7 Memphis, one of the largest private hospitals in
8 the country. I have no financial ties with any
9 anastomotic device companies or, for that matter,
10 any stent companies.
11 My historical experience, I trained with
12 John Simpson back in 1985, '86, and have
13 relationships with many of the members of this
14 panel. I trained with people who are icons today,
15 like Paul Yak, Paul Tierstien, Dean Keriakus, Met
16 Selman, Morris Bookbinder, Rock Califf, Eric Topal
17 and did interventional cardiology until
18 approximately four years ago and made a life style
19 change, and now I do only diagnostic caths and do
20 my chairmanship. Also as discussed earlier, in the
21 late 1980s, with Dr. Chris White, we did brachy
22 therapy because some of the early DCA slides showed
23 needle intimal hyperproliferation similar to
25 That having been said, I, as many of you
1 all, have honed this sixth sense of skill with
2 cardiology over the last 15 years of practice. Dr.
3 Phil Schoettle, who has already presented here
4 today, and I have worked collaboratively for the
5 last 20 years. We basically make a good team
6 because he knows what I do and I know what he does.
7 Our group opened one of the first
8 outpatient cardiac cath labs and it was a labor of
9 trust on his part. Both of us have a sixth sense
10 about when patient is dissected and needs to go to
11 surgery urgently, and have always had that sort of
12 feel. Obviously, in the early days of intervention
13 with PRCA lots of patients went to CABG and, of
14 course, more and more patients went to CABG at that
15 time than do now.
16 So, imagine my chagrin in September of
17 2002 when I cath'd an ER nurse friend of mine and
18 found one occluded and two stented Symmetry aortic
19 connectors, the first patient I had ever seen.
20 When Dr. Schoettle referred to September, 2002 that
21 was the watershed moment. I walked out of that
22 cath, called him and said, Dr. Schoettle, I don't
23 know what this device is but it is a stent and it
24 will act like a stent and it will always be a
stent. I said, what is it? What do we know about
1 it? And, he basically told me his experience over
2 the last ten months.
3 At that point I found an interventional
4 colleague of mine and said, what are these devices?
5 He said he had been stenting them since April; he
6 didn't know much about them. I did an Internet
7 search and found out they were made in nitinol, and
8 realized at that point in time that no
9 cardiologists were involved in either the research,
10 the design, the implementation or the roll-out of
11 this device basically because all the
12 cardiologists, interventional cardiologists,
13 especially know the problems associated with that.
14 It took me about 45 days, almost two
15 months, with multiple interventional colleagues of
16 mine and surgeons in Memphis to have it withdrawn
17 from all the shelves of all three hospitals in
18 Memphis, Tennessee, and that was in the latter part
19 of fall of 2002.
20 As patients have returned to the clinic,
21 dozens, and dozens, and dozens have been found to
22 have virtually total and/or subtotal occlusions of
23 these devices. The first contact I had with St.
24 Jude was in December, 2002 after I had gone to TCT,
in September I believe, and HA in November, telling
1 them about the problem with these devices and why
2 they acted like stents. Finally, they walked in on
3 me while I was cath'ing a 70 year-old ob/gyn who
4 had two patently occluded Symmetry aortic
5 connectors. Basically, I said this is the problem.
6 You don't understand anginal syndrome because most
7 of these patients won't come back with chest pain
8 for multiple reasons--denervation of the heart;
9 more LV dysfunction problem. You don't understand
10 the role of clopidogrel or Plavix in these patients
11 because most of them go to surgery without Plavix
12 on board, and you don't understand the fact that in
13 stent pathology, which we obviously cath a lot, you
14 can have one patent graft, for instance the LIMA
15 which most of these patients get, and the other two
16 can subtotally occlude slowly and their only
17 symptom is LV dysfunction.
18 We, as cardiologists, as members of this
19 panel, diagnose ischemia. We send these patients
20 to a surgeon for treatment and continue to reattach
21 and stent these folks. They will come back for
22 years with their LIMAs. An anecdotal experience of
23 one surgeon in Jonesboro, Arkansas, close by
24 Memphis, asked two of the cardiologists in his
community, "so what's up with this Symmetry aortic
1 connector?" And the cardiologists response was,
2 "are you having any problems?" And he said, "well,
3 I don't know." And they said, "well, don't worry
4 about it."
5 He wasn't satisfied with that, came to
6 Memphis, we cath'd him and his two connectors were
7 occluded and his lumen was patent. After
8 intervention he told me that as an oral surgeon he
9 uses nitinol every day to induce scar tissue
10 formation and keep bridge reconstruction in place.
11 The fact that you auger a hole in the aorta, hold a
12 finger over it beginning the platelet clotting
13 cascade, implant a metallic device with hooks
14 without the benefit of loading doses of Plavix or
15 predictable absorption is inconceivable.
16 The idea of a connector makes sense for
17 improvement of stroke risk, however, I feel the
18 present device should be withdrawn and should have
19 been withdrawn years ago. Basically, I think the
20 cardiologists need to be involved in any future
21 trials or designs and I think to do otherwise is a
22 violation of our sacred oath to our patients.
23 Thank you.
24 DR. TRACY: Thank you. Any questions from
the panel? Comments?
1 [No response]
2 Thank you. Dr. Hausen?
3 DR. HAUSEN: By way of introduction, my
4 name is Bernard Hausen. I am the present CEO of
5 Cardica. My background, I am a cardiac surgeon by
6 training. My financial conflicts are inherent with
7 my position, otherwise I have none other.
9 I want to use this opportunity to show you
10 new generation of products that we are developing
11 beyond the pioneers in this field that we have been
12 discussing so far.
14 We have two products in the pipeline. One
15 is a distal anastomosis system.
17 It is called C-Port and it is based on the
18 principle of simulating interrupted stitch distal
19 anastomosis by applying a set of eight implantable
20 clips, all simultaneously, and performing
21 arteriotomy with the push of one button. This type
22 of a system results in a minimal amount of metal
23 exposure. It is a applied in distal anastomosis
24 and it is in clinical evaluation as we speak.
1 This is just a video showing how it works.
2 This is a 1.5 mm LAD. You insert the anvil; pull
3 it out and you are basically done; place one stitch
4 to close the anvil insertion hole. This is a 1 mm
5 diagonal cadaver heart and shows how it works.
6 This technology, we believe, will enable
7 beating heart surgery as it is quick and does not
8 require any temporary ischemia of the myocardium
9 during placement.
11 We have a second device which is called
12 PAS-Port. It stands for proximal anastomosis
13 system, and it is a second generation proximal
14 system. We have the advantage of being a company
15 that is going to be able to take advantage of the
16 knowledge from the predecessors, predicate devices.
18 So we were able in our design to spend a
19 lot of time on key improvements from things we have
20 learned from the other devices. We have focused on
21 trying to minimize or completely eliminate
22 endothelial trauma of the graft during loading. We
23 wanted to minimize blood-exposed non-endothelial
24 tissue, i.e., metal exposure. We wanted to
maximize the orifice area and reduce the incidence
1 of kinking by a low profile.
3 We did that by basically having nothing
4 touch the endothelium of the vein during loading or
5 deployment. This is a cross-section of the
8 We have a minimal amount of metal exposed
9 with the stainless steel device. It is the same
10 material as is being used for coronary stents. And
11 we wanted to maximize the orifice, especially for
12 small vein grafts, and have a very low profile
15 For all this we have done a clinical
16 trial. We have had a lab cardiologist review our
17 data by QCA and determine what is the amount of
18 narrowing of the implant versus the graft body.
19 They first looked at some hand-sewns that were done
20 concurrently in those patients and, as you can see,
21 the average narrowing of a hand-sewn is about 5
22 percent at discharge and about 18 percent at 6
23 months. This is in agreement with all the
24 published literature.
1 Then we asked them to look at the PAS-Port
2 data. What you find, and you can hardly see this,
3 this is a minus 7 percent narrowing, i.e., the
4 grafts at the anastomosis are larger than they are
5 in the graft body and that is by design. That is
6 how the implant has been designed.
7 Now at 6 months, the most important
8 figure, the average narrowing is 3 percent compared
9 to 18 percent in hand-sewns. I propose that if a
10 device had a problem at discharge or at 6 months
11 you would be seeing that in this quantitative
12 analysis. If you don't see it because the
13 injurious event was at the time of surgery, you are
14 very unlikely to see it going forward besides the
15 normal decay of a vein graft, as alluded to by the
16 previous speakers.
18 So, Cardica's regulatory position is we
19 are applying for 510(k) clearance based on
20 prospective multicenter non-randomized trials, and
21 our primary study endpoint for this distal device
22 is vessel patency at discharge and 6 months, and
23 for the proximal device performing a vessel patency
24 study at 6 months with QCA. Thank you very much
for your attention.
1 DR. TRACY: Thank you. Any questions?
2 Dr. White?
3 DR. WHITE: I just noticed that on the
4 last slide you said you were going to do MRI
5 follow-up on these metal grafts. How are you going
6 to do that?
7 DR. HAUSEN: We have done that on the
8 proximal anastomotic device. Basically, with the
9 gadolinium contrast injection you can see--the only
10 thing CTs and MRIs allow you to do is determine is
11 the graft patent or not. You cannot evaluate the
12 degree of stenosis at the implant. So, a preferred
13 method is a quantitative angiography.
14 DR. WHITE: Do you have experience with
16 DR. HAUSEN: We have done five MRIs in the
17 patients in this study.
18 DR. WHITE: And also CT?
19 DR. HAUSEN: And CT too and MDCT.
20 DR. WHITE: And there is no difference in
21 your hands?
22 DR. WHITE: I like the MDCT much better.
23 I think the image is much clearer. The 3-D
24 reconstructions are very impressive.
25 DR. AZIZ: And
how does that correlate
1 with angiograms?
2 DR. HAUSEN: It depends on what your
3 outcome variable is. If you want to just know if
4 the graft is patent or not, there is a very, very
5 good correlation. That has been shown in the
6 literature. If you need more than that, if you
7 need to know is there a degree of narrowing, that
8 will not suffice.
9 DR. AZIZ: If you do distal anastomosis if
10 you have bleeding, how can you control that? Can
11 you put a regular stitch over that?
12 DR. HAUSEN: Yes, you can. It is the same
13 as a steel device. It is very firm. The pull-out
14 force of this device is very high because stainless
15 steel is three times stiffer than nitinol. So,
16 what you do, you just place the first string around
17 the anastomosis and slowly tighten it. That brings
18 the aorta closer to the implant--
19 DR. AZIZ: If you do distal anastomosis if
20 you have bleeding, can you do regular stitches?
21 DR. HAUSEN: Yes.
22 DR. AZIZ: You have obviously shown a vein
23 graft. If you had an arterial graft can you use
24 your same distal anastomotic site for that?
25 DR. HAUSEN:
This generation of device,
1 no; the next generation, yes.
2 DR. TRACY: Dr. Bridges?
3 DR. BRIDGES: You showed differences in
4 percent stenosis of the proximal anastomoses at
5 discharge and at 6 months.
6 DR. HAUSEN: Yes.
7 DR. BRIDGES: What about occlusion or
8 patency at the same time points?
9 DR. HAUSEN: We have 87.9 percent patency
10 rate so we had 6 occlusions in 50 implants, which
11 is 100 percent in agreement with the historical
12 data from the meta-analysis you saw and we did too.
13 DR. BRIDGES: So, how would you interpret
14 the fact that in spite of having a higher orifice
15 area or diameter you have the same patency at the
16 6-month time point?
17 DR. HAUSEN: That is wonderful proof that
18 it has nothing to do with the connector. It is
19 probably your distal run-off or any of the other
20 200 factors that Dr. Mack said.
21 DR. AZIZ: If you had a very thick
22 proximal ascending aorta--
23 DR. HAUSEN: Yes?
24 DR. AZIZ: --sometimes you do a hand-sewn
vein graft that dunks in and obviously you don't
1 want that.
2 DR. HAUSEN: Yes.
3 DR. AZIZ: Does your anastomosis
4 technique--where would that fit in? Would the vein
5 also dunk in?
6 DR HAUSEN: It is inverted over the
7 implant so it is in the lumen but, because it is a
8 stainless steel implant, it props the anastomosis
9 open and you will not have lumen reduction, if that
10 is where you are heading towards. And we have
11 shown that, minus 7 percent widening of the
12 anastomosis is evidence that that is exactly what
13 the implant does and it accommodates the varying
14 thickness of the aortic wall because it is like a
15 paper clip. It can adjust to varying thicknesses.
16 DR. AZIZ: And the angle at which it comes
17 off proximally, is that oblique or head-on?
18 DR. HAUSEN: It is theoretically 90
19 degrees. We asked our core lab to evaluate that
20 too. There are hardly any at 90 degrees. They
21 vary from 10-70 degree take-offs. Because the
22 hinge point is so small, only 1.5 mm, the vein can
23 come off almost at any angle it wants to.
24 DR. AZIZ: So, could you take the proximal
25 along through the transverse sinus and pull it
2 DR. HAUSEN: You could, yes.
3 DR. AZIZ: You could?
4 DR. HAUSEN: Yes.
5 DR. EDMUNDS: What is the size of the shoe
6 inside the vessel?
7 DR. HAUSEN: The shoe inside the vessel?
8 DR. EDMUNDS: Against which you are
9 putting the clamps down. The part of the device
10 that goes inside the vessel, what are the
11 dimensions of that shoe of the device?
12 DR. HAUSEN: There is really nothing
13 inside the vessel. The vein is pulled through the
14 implant and inverted so there is no metal inside,
15 except for the prongs that penetrated the vein and
16 then go outward. I would be more than happy to sit
17 down afterwards and show you maybe some work. I am
18 kind of limited by the time here.
19 DR. AZIZ: Can you do a sequential of this
20 for the distal anastomosis?
21 DR. HAUSEN: No. Well, you could if you
22 did your side by side by hand, absolutely.
23 DR. TRACY: Thank you very much. Prof.
25 PROF. KLIMA:
Ladies and gentlemen,
1 members of the panel, my name is Uwe Klima.
3 I am a full professor at Hanover Medical
4 School for Cardiac Surgery. The financial
5 disclosure I have to make is that Ventrica paid for
6 my trip here and my lodging, and Ventrica provided
7 us with an unrestricted grant for preclinical
8 testing of an anastomotic device three years ago.
10 I expected a talking time of ten minutes
11 and I will try to cut that down to five minutes.
12 Basically, what I want to talk about is mechanisms
13 of how wound healing takes place after an
14 anastomosis; give you some of our clinical
15 experience with hand-sewn anastomosis, especially
16 with our MIDCAB series; more update or experience
17 with our anastomotic devices; and I will have a
18 little discussion of appropriate methods and
19 follow-up time frames for CABG surgery.
21 As background, we all know that hand-sewn
22 anastomoses now are more or less on the market for
23 more than four decades. Everything is pretty much
24 well tested and evaluated. We have a pretty clear
understanding of what happens to an anastomosis.
1 What happens is a healing response--at what time
2 frame this will be stable. So, I will go into
3 details with my next slide.
5 There are several publications out now
6 which tell us exactly what happens after an
7 anastomosis has been performed. We know there is a
8 lot of trauma coming after surgery. Cell
9 proliferation is coming out. And the most
10 important message that comes out of this
11 publication, for example, is that the repair
12 process is about to be completed two months after
15 We wanted to know what is happening with
16 anastomotic devices. Is it the same response? Can
17 we expect the same thing to happen? Filsoufi
18 published, from Boston. He tested the Ventrica
19 device and what is happening after implantation two
20 months, three months and six months after surgery,
21 and we could see that there is a single layer of
22 endothelium covering after two months, three months
23 and after six months and there was no sign of any
24 inflammatory response at the site of the
2 As a control group--I would like to speak
3 a little bit about our MIDCAB experience in
4 Hanover. We have now enrolled more than 700
5 patients. Out of the first 500 patients we did
6 angiographic follow-up in 6-7 percent of these
7 patients. The first group, which was the big one
8 with 297 patients, had a pre-discharge angiogram.
9 What was pretty interesting was that in about 6
10 percent of these patients we had a highly
11 significant problem at the site of the anastomosis,
12 as you can see here. As a Swedish colleague
13 presented his data with the same problem four years
14 ago at the ASCTS and recommended just to wait
15 because this is part of the healing response, we
16 changed our politics, which you will see on the
17 next slide, and just let the whole situation be as
18 it was; waited for 3-6 months, reevaluated these
19 patients and saw that the degree of stenosis or the
20 number of intimal hyperplasia went down without any
21 intervention from 6 percent to 1 percent.
23 There is another example here and, as you
24 can see again, there is a highly significant
stenosis here at the pre-discharge angiogram;
1 perfect anastomosis 3-6 months later when we
2 reevaluated the patients. What we learned here is
3 that the healing response is still evolving in the
4 earlier time frame. We changed our angiographic
5 follow-up from pre-discharge to a 6-month follow-up
6 so the remaining 203 patients were evaluated 6
7 months after surgery instead of having a
8 pre-discharge angiographic follow-up.
10 So, what is the Hanover experience now
11 with anastomotic devices? Just to give you a quick
12 overview, Hanover does approximately 2,000 open
13 heart procedures per year. It is a large teaching
14 institution. We are affiliated with several
15 research centers so we are exposed to new
16 technologies and clinical trials. The studies I
17 have performed were with Ventrica, St. Jude and
18 Converge. In addition, I have a little experience
19 also with Cardica and Coalescent, however, I just
20 want to present you the data where I have
21 angiographic follow-up.
23 St. Jude--we had a prospective, randomized
24 trial with 11 patients where every patient received
two proximal anastomoses. One
was hand-sewn and
1 the other one was an automatic anastomosis. This
2 is the strongest study design that you can create.
3 The data that we saw was that in 11 patients who
4 were enrolled in the study and came back after 6
5 months and there were 10 postoperative angiograms
6 showing that only 3 grafts were patent. We had 6
7 occlusions and 1 highly significant stenosis at the
8 site of the anastomosis, with a consequent PTCA and
9 stent after graft. Even though the patients were
10 asymptomatic, the study was stopped because the
11 data did not look the way we wanted to have it.
13 Even though all patients were
14 asymptomatic, due to several reasons that we can
15 discuss, I think independent of the cause a
16 prospective six-month angiographic evaluation was
17 sufficient in our study to detect performance
18 issues of the device.
20 Ventrica was part of a multicenter trial
21 that we did with two other centers. We enrolled
22 100 patients, 48 came from Hanover--
23 DR. TRACY: Can I ask you to start
24 wrapping up?
25 PROF. KLIMA:
Yes. The most important
1 information I want to give you here is that in the
2 first 48 patients we had a pre-discharge angiogram
3 and a 6-month angiogram to study efficacy of the
4 device and performance of the anastomosis after 6
7 The last study we did was with Converge.
8 It was also a multicenter trial. We had 8 weeks of