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

9:18 a.m.

CHAIRMAN LASKEY: Good morning. I'd like to call this meeting to order. My name is Warren Laskey and pleased to chair this morning's session. Our topic this morning is a discussion of the pre-market application for the Revelation_ Tx Microcatheter with NavAblator^TM RF Ablation System P020039. Geretta, if you could please read the Conflict of Interest statement.

MS. WOOD: The following announcement addresses Conflict of Interest issues associated with this meeting and is made part of the record to preclude even the appearance of an impropriety. To determine if any conflict existed, the Agency reviewed the submitted agenda for this meeting and all financial interests reported by the Committee participants.

The Conflict of Interest statutes prohibit special Government employee from participating in matters that could affect their or their employers' financial interests. The Agency has determined, however, that the participation of certain members and consultants, the need for whose services outweighs the potential conflict of interest involved is in the best interest of the Government. Therefore, waivers have been granted for Drs. David Schwartzman and Albert Waldo for their interests in firms that could be affected by the Panel's recommendations.

Dr. Schwartzman's waiver involves consulting on a competitor's unrelated product for which he receives an annual fee of less than $10,001. Dr. Waldo's waiver involves consulting on a competitor's unrelated product for which he receives an annual fee of less than $10,001 and also consulting with a competitor on unrelated matters for which he receives an annual fee of less than $10,001. The waivers allow these individuals to participate fully in today's deliberations. Copies of these waivers may be obtained from the Agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building.

We would like to note for the record that the Agency took into consideration other matters involving Drs. Waldo, Schwartzman, Cynthia Tracy and F. Roosevelt Gilliam. Each of these panelists reported past or current interests involving firms at issue but in matters not related to today's agenda. The Agency has determined therefore that they may participate fully in all discussions. We would also like to note that Michael Morton, the Industry Representative for the Panel, has reported interests in firms at issue.

In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participate has a financial interest, the participant should excuse himself/herself from such involvement and the exclusion will be noted for the record. With respect to all other participants, we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement with any firm whose products they may wish to comment upon.

CHAIRMAN LASKEY: Thank you, Geretta. I would like to have the table up front introduce themselves beginning to my right please.

DR. ZUCKERMAN: Okay, Dr. Waldo, can you introduce yourself please?

DR. WALDO: Yes, I'm Dr. Albert Waldo from Case Western Reserve University.

DR. ZUCKERMAN: And please tell us if you can't hear any of these hearings and we'll adjust your phone.

DR. WALDO: Thank you. I hear very well. I guess I'm supposed to say I'm a Clinical Exphysiologist.

DR. ZUCKERMAN: Correct. I'm Bram Zuckerman, Director, Division of Cardiovascular Devices at Food and Drug Administration.

DR. SCHWARTZMAN: David Schwartzman, Electrophysiologist, University of Pittsburgh.

DR. NORMAND: I'm Sharon-Lise Normand, Associate Professor of Biostatistics, Department of Biostatistics, Harvard School of Public Health and also in the Harvard Medical School.

DR. GILLIAM: Roosevelt Gilliam. I'm at Duke University as a Clinical Exphysiologist.

DR. WHITE: I'm Chris White. I'm an Interventional Cardiologist from Ochsner Clinic in New Orleans.

MS. WOOD: I'm Geretta Wood, Executive Secretary.

CHAIRMAN LASKEY: Warren Laskey, I'm an Interventional Cardiologist at the National Naval Medical Center.

DR. MAISEL: William Maisel, I'm a Clinical Electrophysiologist at Bringham and Women's Hospital.

DR. TRACY: Cindy Tracy, Peds, Georgetown University Hospital.

DR. HUGHES: Allen Hughes, Associate Professor of Decision Sciences and Management Information Systems, George Mason University. I'm the Consumer Representative.

MR. MORTON: Michael Morton, I'm an employee of Sorin-COBE and I'm the Industry Representative.

CHAIRMAN LASKEY: Thank you, Members. I understand that there will be some additional folks showing up. There's been an accident outside so we'll have some more people coming in later. We can introduce them at that point. Geretta, if you could read the voting status statement please.

MS. WOOD: Pursuant to the authority granted under the Medical Devices Advisory Charter dated October 27, 1990 and as amended, August 18, 1999, I appoint the following individuals as voting members of the Circulatory System Devices Panel for this meeting on May 29, 2003: Sharon-Lise Normand, Ph.D.; Christopher J. White, M.D.; Alberto L. Waldo, M.D.; Francis R. Gilliam III, M.D.; David S. Schwartzman, M.D.; Thomas Ferguson, M.D.; William Maisel, M.D. M.P.H. For the record, these individuals are special Government employees and are consultants to this Panel under the Medical Devices Advisory Committee. They have undergone the customary conflict of interest review and have received the material to be considered at this meeting.

In addition, I appoint Warren K. Laskey, M.D. to act as temporary Chairperson for the duration of this meeting. This is signed by David W. Feigal Jr., M.D. M.Ph., Director, Center for Devices in Radiological Health and dated May 27, 2003.

CHAIRMAN LASKEY: Thank you. Before we commence with the topic of the day, there will be a brief presentation from the Office of Surveillance and Biometrics on the topic of "Diathermy Interactions with Implanted Leads and Implanted Systems with Leads". So if we can have Marian Kroen.

MS. KROEN: Good morning. I'm Marian Kroen with the Issues Management Staff. The Staff was charged with among other things facilitating centered discussions and evaluations of high profile public health issues through our center wide expert committee meetings. As part of a new initiative to periodically brief panels on important topics, I was asked to make this presentation to you on "Diathermy Interactions with Implanted Leads and Implanted Systems with Leads".

FDA learned of two adverse events. Both of these were interactions of diathermy therapy in patients with deep brain stimulators. Two patients had received shortwave diathermy, one patient treated after oral surgery for a pulled tooth and another patient was treated at the middle back region. Both patients went into a coma and subsequently died.

So what is deep brain stimulation? Basically there's a stimulator or pulse generator implanted around the clavicle and leads snake up through the neck and go to an electrode which is implanted in the brain. It's typically used to treat Parkinson's disease and other conditions.

What is diathermy? Diathermy can be thought of as an electromagnetic heating pad. But while a heating pad heats just the skin, this heats inside the body. Diathermy means deep heat but it can be used in both heating and non-heating modes. It's used for relief from pain with strains, sprains, bursitis and muscle spasms.

Here is a picture of a diathermy unit and a patient receiving diathermy. Diathermy units can come with one or two applicator heads which can be positioned over the patient in appropriate positions.

Who uses this? Diathermy is used by physical therapists, occupational therapists, sports trainers and others.

There are three kinds of diathermy: shortwave which is also called radio frequency, microwave and ultrasound.

When FDA learned of these two adverse events, an expert committee was formed. In parallel to this, Medtronic performed some testing. Medtronic was the manufacturer of the deep brain stimulators. They did some in vitro testing using their deep brain stimulator and the model of diathermy that was used in one of the adverse events.

The result was that there was a temperature rise of 55 degrees Centigrade ("C")at the DBS lead during a 15 minute diathermy exposure at maximum diathermy settings. As you all know, cardiac ablation happens around 50 to 55 degrees C. There was a temperature rise at 27 degrees C at the settings for the adverse event. Again this is a temperature rise from ambient temperature.

So the expert committee sat around and said "What kind of information do we need". And we decided that we really needed to find out how big this problem is and what's the scope. We needed more testing to determine this. Are all implants affected? Are all metal implants affected? Is the problem with only active implants and does the shape of the metal implants matter?

We had OST test active implants. OST is the Center's Office of Science and Technology. They do testing and stimulations in support of FDA's missions. They tested a spinal cord stimulator and a cardiac pacemaker system using a test set up which is similar to Medtronics. They found that the temperature rise at the lead electrode is high. The temperature rise occurs whether or not a pacemaker is connected. So really it's just the lead that's causing this temperature rise.

The cardiac lead in pacemaker systems showed the highest temperature rise of 48.8 degrees Centigrade. The spinal cord stimulator had the second highest temperature rise of 27 degrees C and the temperature rise was highest where a shallowly implanted lead was simulated. It's interesting to note that while Medtronics tested a 15 minute diathermy session, OST tested for a minute or two. They plotted the temperature rise and they tested until the rise flattened out. So this would encompass around 90 percent of the temperature rise. If you did 100 percent of the temperature rise, the temperature at the lead tip for the pacemaker would be around 54 degrees Centigrade.

OST also tested non-active implants to see whether they had an effect. We really found that there was minimal heating of one to four degrees C. We tested mainly orthopedic implants of various shapes and sizes to see if they had any effects. This included screws and plates and titanium and stainless steel rods. We concluded that diathermy interactions with dangerously high temperatures were limited to implantable systems with metallic leads and the implanted leads themselves.

The theory is that the implanted lead acts like an antenna to receive the radiated energy from the diathermy. The power is dissipated. It's collected all along the lead but it's only dissipated where there isn't any insulation which is the electrodes. So the current density and thus the temperature at the lead electrodes can be very high due to the small surface area of the electrodes.

Can all types of diathermy cause this interaction? Both the shortwave and microwave diathermy produce an electromagnetic field which can interact with implanted leads. But ultrasound diathermy would have a different mode of interaction. It has a mechanical vibration rather than an electromagnetic field.

So what did FDA do? We reached into our toolbox and took some actions. We recommend that the labeling on metallic leads and lead systems have a warning against patients with implanted leads. Pacemakers had this warning for eons. Labeling on diathermy equipment had a contraindication against use on patients with implanted leads.

The FDA issued a public health notification which was distributed both to implanting physicians and diathermy endusers. A journal article is in the process of being processed. An issue of "Patient Safety News" was devoted to this. "Patient Safety News" is a televised series that satellite broadcasts to hospitals and other healthcare facilities.

But there's an interesting question here. Why are there no injuries from pacemakers? We have testing that shows that the pace lead electrode temperatures are high, high enough to cause tissue damage. There had certainly been a long enough history of pacemakers and leads for this interaction to have shown up. There have been two reports of damage to the pacemaker but not to the heart with diathermy use.

So we have some possible reasons why we haven't seen any of this. First is the warning may be adequate but as we all know, who reads the labeling? Another thing might be that the blood flow in the heart carries away the heat. This is true but blood flow in the heart also carries away the heat in cardiac ablation and that seems to kill tissues so why would it not kill tissue in this instance.

Next is the distance. Maybe the distance is sufficient to stop damage but we believe the distance of the patient who got diathermy in his mid back is greater than the distance from a pacemaker to a diathermy unit.

What I think really happens is damage but nobody had really put two and two together. A patient might go to the doctor and present with loss of sensing or failure to capture and the physician might reposition the leads. The physician didn't know since this hadn't been publicized to ask the patient whether they had diathermy and the patient didn't volunteer that they had diathermy so they just repositioned the lead and went on.

Another theory is that the brain has no pain receptors and other parts of the body do. If a patient has pain receptors, they can tell the diathermy administrator to lower the settings or stop the diathermy unit. Your thoughts and comments are welcome. Thank you.

CHAIRMAN LASKEY: Thank you, Marian. I guess we will spend a couple of minutes responding. For openers, I don't know if any of those explanations make sense other than the conductive capacity of the heart within the ventricle but it's not a trivial exercise to reposition a lead. I think you would have heard more about that if this was really more than a curiosity. Do my colleagues have any thoughts?

DR. SCHWARTZMAN: Can you tell us more about the nature of the brain injuries that were associated with those deaths?

MS. KROEN: I'm not really that familiar with it. There was an autopsy and it showed edema and necrotic tissue around the lead electrodes but that's all I know. Some representatives from Medtronic are here and they don't know anything more about it either.

DR. TRACY: Was there disruption of the lead or is there thought that the diathermy is causing damage at the tip of the electrode?

MS. KROEN: There was no disruption of the lead. It is at the lead electrode.

DR. TRACY: So you might not know if an ablation took place at the tip of a pacemaker lead as long as it didn't damage things enough so that you lost capture. You might never know.

MS. KROEN: That's correct.

DR. TRACY: And not everybody feels pain with an ablation. Most do but everybody does. It depends on the location of the energy delivery so you might have something that you can't really detect.

MS. KROEN: That's true.

DR. GILLIAM: Short of the dysfunction of the pacemaker, I'm not sure how you'd even know that in pacing just because we ablate with the threshold safety margins we have. It probably could go undetected unless you really created a major lesion. So I wouldn't take shelter in the safety that we haven't had the complaints. It concerns me greatly actually.

CHAIRMAN LASKEY: Do you have any idea what the average number of diathermy sessions per individual is?

MS. KROEN: We don't know. We do that there are chronic pain institutions. They use diathermy quite a lot. I don't know how many pacemaker patients they have but they told me that they've been advocating that pacemakers patients not have this for a long time and had no evidence to back it up. They say they get thousands of hits a day on their site.

DR. GILLIAM: Did they do any work with the defibrillator patients at all? Was there any suggestion along that line? We have mentioned pacemaker but it would seem to me that the defibrillator may have a bigger antenna.

MS. KROEN: They may have a bigger antenna but I think the critical issue is how big the electrode is. I believe the electrode is bigger so it would have a larger area to dissipate the energy. Therefore it wouldn't be as intense and not as hot as the pacemaker lead electrode.

DR. GILLIAM: I guess there are multiple electrodes on a defibrillator. There are some very large ones but they are also smaller points. I'm just looking at a lot of the measurement data that we have. It's a bit more interactive in a defibrillator. We may be able to evaluate other parameters such R-waves or atrial waves and P-waves.

MS. KROEN: Paul, would you know if there are multiple electrodes would the temperature be dissipated among all of the electrodes?

PAUL (tan jacket): I'm Paul Rejaric (PH) of OST. I did the in vitro studies. What we saw with the spinal cord stimulator which is four leads, four electrodes, is that it was distributed among the leads. The report that she gave you today was that the temperature was at the lead at the tip. Then as you went down, the temperature progressively actually got less.

We did find that the area is very definitely a problem. The reason why we saw higher pacemaker one was because the electrode itself was very small compared to the spine cord stimulator lead actually. So it's a surface area energy deposition problem.

CHAIRMAN LASKEY: Well, I'm not sure you're getting a lot of comments from the panel. I guess if you're interested in an inquiry, you go where the money is which is some of these centers where they do a lot of diathermy and do an observational study looking specifically for people with pacemakers which shouldn't be hard.

DR. GILLIAM: Is there a distance from the electrode that you have to be? What's the fall off? If you are within two or three inches, is there a greater amount than as opposed to 15 or 20 inches? Is there a safe distance to be away from it that you wouldn't get this even in a pacemaker lead?

PAUL (Tan coat): We did all of our tests very close. The only reason we did it close was because we were looking at the brain stimulator simulation where they put the thing very close. Ours basically were a half centimeter away which is very close. As we dropped, we dropped down to five centimeters at one point and virtually that knocked away all the heating.

It's a question of how well do you couple. How is the thing oriented? We find for example if you oriented a certain way, you actually get no coupling at all which is true with any antenna. It's all a problem of probability. Had the person turned the lead the other way he still would have gotten the heating but probably not even bother the patient. It's just one of those things.

MS. KROEN: I would also add here that again one of the patients had diathermy to the mid back and the electrodes were in the brain. That's quite a fair distance. So we really don't know what a safe distance would be.

CHAIRMAN LASKEY: All right. Thank you. We're puzzled as well. I would like to move on to the topic at hand and begin with the open public hearing. I would like to invite anyone from the audience who wishes to address the panel on today's topic to please approach the podium and identify yourself. If not, then I will close the open public hearing and we'll move to the sponsor's presentation. Geretta.

MS. WOOD: I would just like to remind the speakers to introduce themselves and state their conflict of interest before presenting.

MS. BALDWIN: Good morning, ladies and gentlemen, members of the Panel. My name is Marianne Baldwin. I'm with Cardima. I'm here today to present to you the results of a very complex study of a very complex disease, Atrial Fibrillation.

I'm going to start just by introducing the people who will be speaking for Cardima today on the Treatment Options of Atrial Fibrillation, Dr. Neal Kay, on the Preclinical Studies and the Protocol Development for the Clinical Studies, Dr. Hugh Calkins, the Results of the Clinical Studies, Dr. Abraham Kocheril, our primary investigator, and the Conclusions and wrap-up again by Dr. Kay.

Just as a brief background since 1993, Cardima has been developing, manufacturing and marketing catheter-based systems for the Electrophysiological field exclusively. The catheters include the PATHFINDER^TM family of mapping devices, the VENAPORT_, VUEPORT_ and NAVIPORT_ guiding catheters and the REVELATION_ family of mapping and ablation systems which represent the system under review today.

Currently the company is marketing its diagnostic and guiding catheters in the U.S.A., Canada and the European Union and Japan. The REVELATION_ family of mapping and ablation devices are marketed in Canada and the European Union.

This is a listing of our currently available products in the U.S. The initial strategic plan for Cardima was to introduce the mapping and guiding systems before the therapeutic devices to allow time for the physicians get comfortable with the smaller devices. 510k clearance was obtained for the diagnostic and guiding catheters being in 1995 with the VENAPORT_. Four years later this entire list was cleared.

These are illustrations of our catheters. This is the guiding catheter of the NAVIPORT_ an 8.0 French device. This device is used to deliver the REVELATION Tx the device we'll be studying today as part of the investigational system although this device has been cleared. It was the first deflectable guiding catheter on the market. The VUEPORT_, its predecessor on the right is practical to obtain good venagrams. We've had 20,000 of these units distributed.

This is the NavAblator^TM, Cardima's four millimeter hot tip developed specifically for this study to create lesions at the isthmus line.

This is the PATHFINDER^TM, examples of the catheters that have designed to access coronary sinus and its feeders. The 1.5 mini PATHFINDER^TM is useful for mapping small vessels such as Vein of Marshall. Our standard PATHFINDER^TM on the right, the 2.5 French, maps the coronary sinus and the feeders. Physicians have been using it lately for the placement of pacemaker leads. Cardima performed a clinical study for the PATHFINDER^TM to demonstrate its ability to map effectively. We have distributed 20,000 units of these devices.

The REVELATION_ Tx, the device shown on the left there, is a 3.7 French device to map and ablate. Its sister device for pulmonary veins is distributed in Europe and that's a 4 French. The AD-deflectable of the 3.7 French device is also distributed in Europe.

This is a description of the differences and the unique features of Cardima's core technology. The principal distinguishing feature is that it incorporates a guide wire technology that was developed for angioplasty permitting a variable stiffness of the guide wire that allows construction of a very flexible and compliant distal segment. It's a composite shaft that includes conductors tightly braided over the core wire and the structure is encapsulated in polyethylene before being coated with a hydrophilic lubricant.

The other really key feature is the fine wire coiled electrode located along the distal segment. We have eight electrodes on this catheter arranged in linear array allowing the catheter to maintain good contact with the wall of the beating heart. It also permits these small electrodes the catheter to ablate with greater current density and requiring less power.

Just to illustrate and compare the lesion shape of this catheter to a hot tip catheter, you'll see that there is similar depth but a significant difference in the width. Preservation of myocardium is a result of this and was one of the features of this design.

Here we have a segment of the linear array with three electrodes shown and four thermocouples in the middle. You'll see that the lesion formation overlaps the thermocouples. You have a continuous lesion there. We can ablate with between seven to 35 watts of power.

Last but not least, Cardima's indications for use are treatment of atrial fibrillation in patients with drug refractory paroxysmal atrial fibrillation by mapping, pacing and ablating with a set of lesions in the right atrium.

Now Dr. Kay will present to you Treatment Options on Atrial Fibrillation. Thank you.

DR. KAY: Thank you very much. My name is Neal Kay. I'm from the University of Alabama. I have no financial interests in Cardima. I'm not an investigator and I'm not a consultant. I'm being paid for my time today however.

Atrial fibrillation as we all know is a huge problem and there probably are as many treatments for this problem as there are patients who have it. This slide just shows the magnitude of the problem as it relates to age. As the population ages, you see that the absolute number of patients with Atrial fibrillation is increasing and both the incidence and prevalence go up with age. So this is becoming more and more of a problem

It's a complex rhythm and not every kind of atrial fibrillation is the same as every other kind. The ACC/AHA have recognized this and have proposed the following classification scheme. Patients may present with their first episode of detected atrial fibrillation in which case the subsequent course is not known. Then they may go on to have paroxysmal atrial fibrillation which are those that terminate spontaneously or have persistent atrial fibrillation, those episodes that require cardioversion or some drug intervention to terminate and do not terminate spontaneously.

Then finally there is this permanent atrial fibrillation which atrial fibrillation that is refractory to cardioversion or atrial fibrillation that in the opinion of the patient and their physician where it's just hopeless to try to maintain sinus rhythm. So it's a complex problem.

There are significant limitations with drug therapy. Shown here in this slide is just the results of one prospective trial looking at different doses of Sotalol vs. Placebo in the recurrence of atrial fibrillation. As you can see, there is a high risk of recurrence even on fairly large doses of Sotalol to maintain sinus rhythm. At six months, less than half the patients will have persistence of sinus rhythm.

Similarly the Canadian Trial of Atrial Fibrillation ("CTAF") look at different drugs after cardioversion as you can see. I think we all know that Amiodarone is more effective than other antiarrhythmic medications. It's more effective than Propafenone or Sotalol. But even so, many patients will have recurrence of atrial fibrillation.

The drugs that we use have limitations. The limitations are largely a function of the kind of underlying heart disease. These are the ACC/AHA guidelines. For example, patients who have congestive heart failure, it's recommended by the ACC/AHA that probably Amiodarone and Dofetilide may be the only drugs that may be used safety.

For patients with coronary artery disease, there's a significant limitation in the use of Type 1C drugs. So this scheme has been proposed that we perhaps start with Sotalol and then perhaps go to Amiodarone or Dofetilide and finally non-pharmacologic options can be considered.

Hypertension and specifically Left Ventricular Hypertrophy present problems with drugs that prolong the QT interval. When there is significant Left Ventricular Hypertrophy, we may be limited to the use of Amiodarone. For patients who don't have significant structural heart disease and just have hypertension, the Type 1C probably are reasonably safe to use followed by Amiodarone, Sotalol or Dofetilide.

This slide is presented to highlight the importance of the pulmonary veins in the initiation of atrial fibrillation. Here is our surface intercardioelectrograms from a patient with the spontaneous onset of atrial fibrillation that arises in this case from a pulmonary vein In this case it's the right superior pulmonary vein. You see an arrow here indicating the first onset of electrical activity in the pulmonary vein. It comes from the pulmonary vein to the left atrium and then subsequently to the right atrium. So the pulmonary veins appear to be important in the initiation at least of atrial fibrillation.

Arial fibrillation ablation is currently being practiced around the world in relatively limited of centers. The most common technique is shown here with a circular mapping catheter place at the ostium of the pulmonary vein to guide recognition of sites of early activation between the pulmonary vein and the left atrium. A radio-frequency catheter is placed just proximal to that mapping catheter and radio-frequency current is delivered with the goal being to electrically isolate the pulmonary vein.

This just shows an example of that. On the bottom tracings in green, you see pulmonary vein potentials and radio-frequency current applied. Finally after the onset of RF, we see that there is disappearance of pulmonary vein potentials. That is how this technique is being applied most commonly to ablate atrial fibrillation now throughout the world.

There are some significant limitations to pulmonary vein ablation that I want to highlight. One of these is the occurrence of this complication which is pulmonary vein stenosis. This is a right superior pulmonary vein. There is significant narrowing of the pulmonary vein just at the ostium as it joins the left atrium. This is an important and potentially devastating complication of ablation in the left atrium.

These results here published about a year ago are fairly typical of those that are seen in most of the large centers doing this procedure. You see that for patients who have paroxysmal atrial fibrillation about 70 percent of the patients can be rendered free of reoccurrence at least in the immediate term.

On the dotted line, you will see that the results for persistent atrial fibrillation really are not as good with only a small minority of the patients having control of their atrial fibrillation with a pulmonary vein isolation procedure if they have persistent AF.

AF ablation does have significant limitations. I just want to highlight them. Pulmonary vein isolation is only effective in about two-thirds of patients with paroxysmal AF. It's less effective in patients with more advanced forms of atrial fibrillation.

There are significant risks and these are likely to be operator dependent. Pulmonary vein stenosis has been reported to occur anywhere from one percent to eight percent and clearly seems to have a relationship to the experience of the operator. There's a risk of stroke that ranges from about one percent to four percent.

Pericardial tamponade again in the same range. Major bleeding complications can and do occur and they are related to the large amounts of anti-coagulation that are generally required to required to ablate on the left side. Importantly, it's limited to physicians who feel competent to perform transeptal catherization. Therefore it presently is not widely applicable.

I would like to highlight that there really have been no multi-center prospective trials of pulmonary vein isolation to this point. So it is likely that the complications that have been reported in single center series probably are actually higher when we finally start to do multi-center prospective trials.

This slide just shows the results of four surgical strategies to treat atrial fibrillation. Here in the top left is what is called the Cox-Maze operation which is bi-atrial operation involving incisions around the pulmonary veins and also incisions in the left atrium and importantly in the right atrium. The percentages are the percentage of patients who are rendered free of atrial fibrillation by the surgery. You can see that this probably is the best results that have been achieved which is with this bi-atrial surgery.

The other tracings show the results from different investigators looking at purely left atrial ablation procedures at surgery. As you can see, the results aren't as good. So it does appear that if you surgery ablate in both atriums the outcome appears to be better than in just the left atrium alone.

How does surgery work? It probably works by at least two different mechanisms. One of those is that it probably prevents pulmonary vein triggers by encircling the pulmonary veins but the incisions in the atrium probably interrupt macroreentry. We know that most atrial fibrillation is maintained by relatively large macroreentry circuits at least in our studies on the size of about nine centimeters. Surgery may help to prevent these by preventing the real estate that's available to maintain macroreentry circuits.

Now I've shown you some reasons why the pulmonary veins are probably very important. I've also shown you some reasons why operations involving both atria may be better than just the left atrium. Here are some results reported from Japan looking at their results with the Cox Maze 3 operation which again these first two columns are bi-atrial operations and just the right atrial Maze operation along. As you can see, the results are clearly better when two atria are operated on.

But there is at least a moderate degree of success with just a right atrial Maze operation done surgically. So it does raise the possibility that a right atrial operation may have some role to play in the prevention of atrial fibrillation.

As I've shown you the results at least from surgery, these results demonstrates that there may be a place for a right atrial ablation alone and I think Cardima will now presents some results showing that there is indeed a role for this approach in a prospective clinical trial. Thank you.

DR. CALKINS: Good morning. My name is Hugh Calkins and I would like to present some of the animal work that has been done with this system as well as the study design.

MS. WOOD: Dr. Calkins, please state your relationship to the sponsor.

DR. CALKINS: I've been involved with animal studies with Cardima about three or four years ago. I also was a clinical investigator in this clinical trial and I've been paid for my time for being here today. I'm not an equity owner in the company.

The Pre-Clinical Studies have been performed which have demonstrated that the Cardima REVELATION_ system is biocompatible. It's compliant with applicable ISO requirements. It also has been demonstrated that this is a reliable catheter be it again compliant with the mechanical and electrical performance requirements of the Massi guidelines.

Now a large number of animal studies have been performed with the Cardima REVELATION_ ablation system and they are summarized on this slide. The system has been worked at the Massachusetts General by David Kean, M.D.

Other animal studies have been performed by Dr. Mauricio Arruda at the University of Oklahoma in both a thigh muscle preparation and the dog preparation. Then additional studies were performed at the Mayo Clinical by Drs. Asirvatham and Doug Packer and also at Johns Hopkins. I would like to highlight the results of some of these for you today.

The study that was done by David Keane who was really first to work with this catheter was performed in an AF goat model where they demonstrated that AF could be induced at baseline and then delivered the four lesions that are part of this study design today and showed that AF was no longer inducible after delivery of these linear lesions. They also showed that these lesions were frequently transmural in nature.

The studies done at the Mayo Clinic were interesting. They look at 14 dogs. Their objection was to compare the linear microcatheter system with the linear standard ablation system where you have standard 4 millimeter 7 French electrodes lined up along a multipolar catheter. They created 30 lesions with the Cardima system and 36 linear lesions with the alternate catheter system.

The results are summarized here that the lesions were slightly narrower with the Cardima microcatheter system with a standard 7 French system. The lesions had similar depth. The volume was less with the Cardima system as you would expect given that it had smaller width. Importantly the presence of lesion formation was the same, 98 versus 95 percent. Lesion was transmural in a similar portion, 89 percent and 85 percent.

Dr. Arruda looked at the NavAblator system, the hot tip, flutter line catheter in six dogs and showed that this indeed was able to ablate the isthmus. Isthmus block was achieved in five of six dogs. The typical lesions were created with this 4 millimeter standard ablation catheter.

The study I'm most familiar with is the study we performed at Hopkins in 10 dogs. When I heard about this ablation system, I was somewhat skeptical that a small 3.7 French catheter could really create linear lesions and it could create deep lesions. I said "Why don't you let me do a little study in our own animal laboratory where we can compare head to head with standard ablation technology using a 4 millimeter catheter with a drag and pullback approach versus the Cardima system".

I set out about to do this again creating four linear lesions in the right atrium of these dogs. The dogs were survived one month and then they were sacrificed and the lesion characteristics were compared. This slide shows what we found with the standard 4 millimeter catheters using the sheath and pulling it back point by point trying to make linear lesions. What you can see on the left side of this screen is what happens on a good day where here's one lesion and here's a second lesion and here's a third lesion. We all would agree that this is a linear lesion as far as we can see.

But what we saw more commonly was this kind of thing where you have one lesion here, one lesion here, a third lesion here and then you had a clear gap between the lines. Then we also commonly saw what you see in the third which looks like a gunshot blast where you have one lesion here, one lesion here, one lesion here, one lesion here, nothing continuous or linear about it. Whether these lesions are in fact pro-rhythmic, anti-arrhythmic, who really knows.

Now you compare that with what we saw with the Cardima system and there was a striking difference. Here's a linear lesion created with a system that is thin and linear and this is on the free wall of the atrium. Here's a line along the septum going through the coronary sinus os going down to the inferior vena cava. Here's a flutter line from the tricuspid valve back to the inferior vena cava again a thin linear lesion.

Now when we looked at these lesions histologically, we found that they were narrower than standard RF lesions as we expect.

We found that there was no difference in the overall length from start to finish of these lesions. They tended to be longer than the drag and pull lesions but this was not statistically significant. They had a similar length than what we created with the standard catheter.

When you looked at depth, there was no difference. Perhaps the Cardima lesions were a little bit deeper. I think this was one of the most important findings of the study.

Then when we looked at things like lesion linearity, were the lesions delivered in a linear fashion, all the Cardima lesions were linear versus only 37 percent of the drag and pull lesions.

When we looked at continuity, were the lesions continuous, 72 percent of the Cardima lesions versus 37 percent with the standard drag and pull approach.

Finally when we looked at anchoring to anatomic structure to the SVC or the IVC, we had 92 percent anchoring versus 62 percent anchoring again very different result.

We concluded that atrial lesions created using the Cardima system are narrower, more continuous, more linear and more likely to anchored to an anatomic structure than those created with a standard ablation catheter using a drag and pull approach. These differences in lesion characteristics may facilitate cure of atrial fibrillation with a catheter based Maze procedure.

Now I want to go over to how the protocol came to be and how the study was designed. When the study was designed, the question was where to start. Do we do an left atrial procedure, a right atrial procedure? These discussions were really between Cardima and Massachusetts General who was very heavily involved at that point. David Keane was one of the early investigators with the system and the FDA.

It was considered that the optimal lesion set for atrial fibrillation treatment wasn't really known. We knew that Cox Maze procedure was effective when you did all the lesions. But if you take them away and peel them back, what was the least number of lesions you need to see effect?

We now know which we learned that safety and efficacy of pulmonary vein isolation was unknown then and it's unknown now. When this clinical trial started, that wasn't even on the horizons. Nobody even knew about the pulmonary veins at that point. The most important thing was that safety first above all do no harm and that right atrial ablation is likely to be lower risk than working the left atrium both in terms of thrombo-embolic complications and also the complications associated with a transeptal technique. For that reason, it may be more widely applicable in the AP community.

This is how the study was developed. The initial study design was performed six years ago in 1996 again between the MGH, FDA and Cardima. They were in discussions at that point. Phase Ia was just a mapping phase started in 1997 with the PATHFINDER^TM was similar to the REVELATION_ catheter without the thermocouples.

Phase IIa started in 1998 and that's the early ablation part of the study. In 1998 is when the Circulatory Systems Advisory Panel first met to get advice regarding how to design atrial fibrillation studies. Although the Cardima study was designed prior to this meeting, the study design was consistent with it. Phase IIb which is part of the phase included in the results presented today started in 1998. Mapping and ablation was performed again with the REVELATION_ catheter.

The second Advisory Systems Panel met in 2000 and additional discussions were made about design of atrial fibrillation trials. Then Phase III was started in 2000 and the main difference was the addition of the NavAblator^TM catheter, the standard deflectable hot tip catheter, to the protocol for Phase III.

When you look at the recommendations that were made by the FDA Advisory Panel at that time, they recommended a single arm non-randomized study. They recommended that patients serve as their own control, that patients should have failed at least two or more anti-arrhythmic drugs or amiodarone and that they should have had two episodes of atrial fibrillation over a three month period.

The Cardima study design was designed in consistency with each of these recommendations. But importantly instead of having two episodes of atrial fibrillation over three months, they had three episodes documented over one month to get in their study to be sure that this was a highly symptomatic patient population.

The Advisory Panel recommended that an appropriate endpoint would be a 50 to 75 percent reduction in the frequency of symptomatic atrial fibrillation episodes and that six months was a reasonable timepoint for evaluation of therapy effectiveness. Safety was obviously important as was quality of life. Again these were included in the study design of this protocol that you will hear about this morning.

The objective of the study was to determine if the REVELATION_ ablation system successfully and safely reduced the number of symptomatic atrial fibrillation episodes, that it was safe and that it improved the quality of life.

The inclusion criteria is summarized here. Again there was three or more symptomatic episodes of atrial fibrillation documented by an event monitor in the 30 days prior to enrollment and prior to the procedure. They had to be refractory to two or more anti-arrhythmics or refractory to amiodarone alone. They had to have no significant structural heart disease or marginal heart disease. The left atrial size had to be less than five centimeters. They had to have no echocardiographic evidence of an intra-atrial thrombus or an ASD or PFO on a trans-esophageal echocardiogram.

Some of the pertinent exclusion criteria are here: acute ablation failure within two months, MI within six weeks, CVA or TIA within six months, pregnancy or coagulopathy that was known.

This slide goes over the general scheme of the study. Patients were approached for entry into the study and form consent was obtained. They then were enrolled in a 30 day monitoring period where they were given an event monitor and were instructed to transmit their episodes of atrial fibrillation. They were not told they had to cross a certain bar to get in but they were just provided with this event monitor.

If they met the three episodes in two months, then the ablation procedure was performed. They had a TEE performed prior to this. Then after catheter ablation, they were followed prior to discharge in the one, three, six and twelve months. If they didn't meet the initial screening phase, if they did not have three episodes documented in one month, they could have another screening phase where they would have to have nine episodes documented over three months. So some patients were able to get in at that point in the study.

This slide summarized the lesion sets that were delivered. There were three lesion sets that were part of the protocol. One was the post-lateral line from the superior vena cava back to the inferior vena cava posteriorly and laterally. Second was a standard isthmus or flutter line. The third was the septal line from SVC to the IVC along the septum through the coronary sinus through the frame and the valley and down to the inferior vena cava.

Now in Phase IIb there was an optional anterior line that was subsequently delayed in Phase III both because it was delivering energy in the region in the sinus node and also because it was not felt to be that critical to the procedure. The whole idea was to have a procedure that could be accomplished in a short period of time. So we settled on the three A, B and C lesions summarized in this slide.

This just shows the patient follow-up that was performed. Again patients had a history, physical exam, 12-Lead EKG at baseline and at office visits at one, three, six and 12 months after ablation. The TEE was done prior to the ablation procedure. The echocardiogram and stress test was performed at baseline and three months.

An event monitor was performed weekly and with symptomatic episodes of atrial fibrillation at baseline and at months one, three and six. Patients completed quality of life questionnaires with both SF-36 and atrial fibrillation severity score at baseline, at three and six months following ablation. Then they had a telephone interview at 24 months.

The primary clinical endpoints as defined in the study protocol initially submitted to the FDA was the reduction in the frequency of spontaneous atrial fibrillation episodes and also to look at the incidence of adverse effects. The secondary endpoint was the quality of life on the on the SF-36 or the atrial fibrillation severity scale.

The acute procedural success criteria was a reduction in amplitude, fragmentation or widening of the local electrograms, appearance of split potentials and/or an increase in the pacing threshold following ablation. Then the long term endpoints or the primary endpoint was a 50 percent reduction in symptomatic atrial fibrillation episodes for patients with more than five episodes per month during the screening period. Then for patients who had less episodes of atrial fibrillation, only three or four episodes, they would have to demonstrate a 75 percent reduction to meet the effectiveness endpoint.

Clinical success was defined as reduction ain atrial fibrillation episodes while maintaining the same anti-arrhythmic drug regimen or a reduced dosage of the drug regimen during the follow-up.

The sample size was determined to be 80 evaluable subjects at six months based on an estimated patient success rate and statistical considerations. This sample size was specified again in protocol submitted to the FDA and approved prior to initiation of this study.

In summary, multiple animal studies have demonstrated safe creation of thin, transmural, linear lesions with a REVELATION^TM ablation system. This clinical study was designed in collaboration with the FDA. The clinical study incorporates a large number of measures of safety and efficacy. Importantly as the first atrial fibrillation clinical trial, this study is clearly charting new waters. Thank you for your attention. I would like now to introduce Dr. Abraham Kocheril who is the principal investigator of this study. He's from the Carle Clinic and he will be presenting the study results to you.

DR. KOCHERIL: Good morning, everyone. My name is Abe Kocheril. I was the principal investigator of the study. I'm being paid for my time and expenses. I do not have equity interest in the company. I will present to you the patient population and study results.

Twenty clinical sites took part in this trial. As you can see from this list, there was a mix of major university centers as well as community hospitals. I think that speaks to the generalizability of the results that you are about to see.

One hundred and twenty patients underwent radio-frequency ablation according to the protocol. There were eight patients who withdrew within the six months of follow-up. There are 18 patients who have not completed six months of follow-up which leaves 87 patients who have completed six months follow-up. These 87 are the subject of the effectiveness analysis. The 120 who underwent initial ablation are the subject of the safety analysis.

Looking at the eight patients who withdrew prior to six months, four of these were unable or unwilling to do follow-up. One of these patients had moved out of state. One had changed insurance carrier. Two others were simply unwilling to comply with the requirements of the study protocol.

The other four had undergone AV node ablation and/or pacemaker implantation. There is a tendency to look at these as failures. But I can tell you on one of these patients what happened was that this patient had a single recurrence of atrial fibrillation. After the ablation procedure, the family convinced the patient to seek a second opinion. The physician rendering the second opinion recommended AV node ablation and pacemaker implantation as the way to go. She did undergo this procedure. However in follow-up all the follow-up tracings, all the follow-up pacemaker checks showed either sinus rhythm or an atrial pace rhythm.

So we start with the 87 patients who completed six months follow-up. Five of these had indeterminate baseline episode data. One had indeterminate six month episode data leaving 81 who were evaluable for primary effectiveness endpoint.

The baseline characteristics are shown on this slide. The mean age was 57. Seventy-seven percent of the patient were male. Seventy-two percent had prior cardiovascular disease aside from their atrial fibrillation. Twenty-six percent had respiratory conditions. Twenty-six percent had endocrine conditions. Nineteen percent had neurologic disease. Eighteen percent had renal disease. So even though we tried exclude people who had significant structural heart disease, we nonetheless ended up with a group of patients who did have some structural heart disease and had co-morbidities.

Twenty-eight percent of the cohort had prior RF Ablation and I'll show more detail on that. Eight percent had had bypass surgery. Seven percent had DC cardioversion before. Six percent entered the study with a permanent pacemaker. Six percent more had percutaneous coronary interventions prior to entering the study.

Looking at the patients who had prior radio-frequency ablations, 22 patients had ablation of atrial flutter prior to entering the study. Nineteen additionally had ablation of super ventricular tachycardia or atrial tachycardia.

The baseline arrhythmia symptoms are fairly typical for an atrial fibrillation population. The most common symptom was palpitations occurring in 87 percent. The second most common was fatigue occurring in 59 percent. Shortness of Breath in 50 percent, lightheadedness in 37 percent. Nineteen percent had chest discomfort and other symptoms were also reported.

The baseline symptomatic atrial fibrillation episodes per month are shown on this slide. Hugh Calkins already mentioned the FDA Advisory Panel recommendation of a minimum of two episodes per three months. That's the bar here for reference. You can see from the remainder of the bars that our study cohort easily exceeded that minimum. There was predominance of patients who had five to nine episodes per month. The range extended all the way up to 30 plus episodes per month. The mean was 10 episodes per month for the study cohort.

Congruent with the symptomatic episodes, there was a tendency to have a reduced quality of life. This is the SF-36 domains which is a nonspecific measure of quality of life. There was a trend towards a decrease in all of the categories but particularly striking where the decreases in role, physical and vitality and social functioning. As we'll show later, there are the three parameters that tended to increase the most after an ablation procedure.

Now that you know who the patient population is, we'll move to study results. Dr. Calkins already mentioned the prescribed studied lesions. These were the post-lateral line A, the Septal line B, the isthmus line C and as was mentioned line D was discouraged in Phase III of the protocol. As you can see on the right, line A was delivered in 90 percent of these patients, Line B in 93 percent, line C in 90 percent. Only 8.7 percent received line D. The majority of the procedures were done with lines A, B and C for a total of 83 percent.

The mean procedure time was 250 minutes. That mean fluoroscopy time was 47 minutes. There is a bit of a learning curve in this. At my center we were able to complete these procedures in about three hours with the fluoroscopy time of about 30 minutes.

Acute procedural success is generally considered a surrogate for clinical effectiveness. The initial plans to record specific measurement of acute procedural success became unwieldy. Just to illustrate that, what we are dealing with is an octopolar catheter. There are eight electrodes.

It usually takes more than one pass to complete each linear lesion so you are looking at potentially 16 electrodes at which to do these analyses of decrease in electrogram amplitude, fragmentation or widening of electrograms, possibly measuring pacing thresholds. There are three lines so that three times 16 is 48. If you do three measurements on each of those lesions, that's 144 measurements before you are done with the procedure. This really was unwieldy so these were not consistently recorded.

However in the EP community it is standard practice to look for reduction electrogram amplitude, appearances with potential, etc. as measurements of lesion adequacy. These were assessed in the procedure. The investigators were asked at the end of the procedure whether they had adequate lesion development. At the investigators' assessment, 110 out of 118 procedures were deemed successful for a rate of 93 percent.

Dr. Calkins also mentioned that the primary endpoint was a symptom reduction endpoint. There was at least 50 percent reduction in patients with five or more symptomatic AF episodes per month at baseline and 75 percent or greater in patients with three to four episodes per month at baseline. Using this definition, 69 of the 81 patients in the cohort or 85 percent were successful.

This is an important slide showing symptomatic episode reduction from baseline to six months. This shows each of the 81 patients individually. What you can appreciate is that there is a striking and almost uniform decrease in symptoms from baseline to six months.

Some of the other salient features are that the people with the most symptoms, the highest scores, are the ones who tended to benefit the most from the ablation procedure. The other thing that you will notice is that several of these subjects do approach zero episodes at the six month mark. As an investigator, these are very gratifying results. As a clinician what this translates into is patients feeling better and not calling in with symptoms.

This is a different way of analyzing the atrial fibrillation frequency reduction. This is looking at the entire cohort. What you see is at baseline there was 9.2 episodes per month on average. This was reduced by three months to 3.5 episodes per month. At six months, it was further reduced to 1.2 episodes per month. These are highly statistically significant decreases from baseline with P values in the range of .0001.

There were patients who were free of atrial fibrillation at six months. In fact 44 out of the 81 had no episodes at six months for a rate of 54 percent.

Congruent with this, there was a reduction in common arrhythmia symptoms. Graphed here are the various symptoms, palpitations, fatigue, shortness of breath, lightheadedness, chest pain and others. What you will see here is at least a 50 percent reduction in most of these parameters with the exception of chest pain and other.

A pre-specified secondary endpoint of the study was quality of life. This was addressed by two measures. One is the atrial fibrillation severity scales which is a measure specific to atrial fibrillation. The other is the less specific SF-36 domains.

In AFSS, we saw significant improvements in episode frequency, episode duration, episode severity and obviously the total AFSS score. These are not only statistically significant. What makes this clinically meaningful is that you'll see that there's at least a 10 point improvement in each of those categories in the AFSS.

The less specific SF-36 domains also showed improvement. We saw improvement in six out of the eight categories. Particularly striking here are the improvements in role physical, vitality and social functioning. As I mentioned earlier in my presentation, these were the ones that were more significantly depressed compared to U.S. norms at the beginning of the study.

The FDA definition of major complications for catheter ablation studies is an adverse event that occurs within seven days following an investigational procedure and is life-threatening, results in permanent impairment or damage to a body structure, requires significant intervention to prevent permanent impairment, requires hospitalization or an extended hospital stay, results in moderate transient impairment or damage to a body structure or requires intervention such as medication or cardioversion to prevent permanent impairment or damage to a body structure. This definition was used for our safety analysis.

Using this definition there were four major complications out of 123 procedures for a rate of 3.3 percent. There was one pericardial effusion requiring a pericardial window one week after ablation, one sinus note injury requiring pacemaker implantation, one stroke occurring four days after ablation and one AV fistula.

In the FDA review that will follow our presentation, one additional patient is coded as a complication who received a pacemaker within seven days. This patient had however pre-existing documented sinus node dysfunction. Pacemakers as an issue will be addressed later in my presentation. The main point of this slide is to show what the complications were. You can appreciate that this is very similar to what's reported in other studies of radio-frequency ablation.

Adverse events were tracked beyond the one week mark. In later follow-up, there were 30 patients reporting adverse events. Six patients had undergone AV ablation and pacemaker implantation. Two had undergone cardioversion. There were two infections reporting. One was an upper respiratory infection and the other a urinary tract infection. There was one report of sinus node dysfunction. There was one stroke that occurred over a year after the ablation procedure and 41 others who are detailed in the next slide.

This is an accounting of all the other adverse events that were reported. Without going into detail of these, most of these are unlikely to be due to the ablation procedure. The one that caught our attention was the one patient who had pulmonary hypertension. Looking into this further, this was a patient who had congestive heart failure and mitro regurgitation. Mild pulmonary hypertension was detected by an echocardiogram six months after the ablation procedure. The procedure as you know is entirely confined to the right atrium and we do not feel that this was related to the procedure.

Importantly there were no reports of mortality, cardiac perforation, arterial injury, stroke or thromboembolism. In fact, 75 percent of the patients reported no adverse events whatsoever.

Having presented the main data, the main clinical results, I would like now to move to some other issues for further clarification. The first is the catheters used in the lesion sets. The second is the issue of patients receiving pacemakers. The third is the protocol definition of clinical success and the use of antiarrhythmic drugs. The last is compliance with transtelephonic monitoring.

The first issue has mainly to do with the creation of the isthmus lesion or the flutter line. You will remember that some patients entered the study after having had a flutter line before. They had a prior flutter ablation. The anatomy of the isthmus became better appreciated through the conduct of the study.

Now we know that it's not a smooth structure, that it has peaks and valleys, the eustachian ridge. Because of this complex structure, investigators did not feel that they could always get an adequate lesion using the REVELATION_ Tx. Because of that, they resorted to standard four millimeter hot tip catheters to complete the flutter line.

Importantly the REVELATION_ Tx was used for all non-isthmus linear lesions. The REVELATION_ Tx was used for some of the isthmus lesions and at the investigators' discretion, they moved to a four millimeter hot-tip catheter to complete the flutter line.

Because of this issue, Cardima developed the NavAblator for use in Phase III specifically to complete flutter line. Despite its introductions, some investigators preferred to use the four millimeter hot tip catheter that they were accustomed to. Given the fact that flutter ablation is standard in the AP community, we did not feel that this variation in catheters used at this location materially affect the results of this study.

Next issue is pacemakers. The study protocol stated "Subjects electing to receive implantable pacemakers prior to six months follow-up will be considered failures." The intent of this statement was that pacemakers should not used as adjunctive therapy for atrial fibrillation. As I'll show you, most study subjects receiving pacemakers don't fall into this category. Also as mentioned at the beginning, patients with pacemakers were not excluded from entering this study.

A total of 20 pacemakers were implanted during follow-up. Seven of these were implanted after six months of follow-up and these do not affect six month efficacy outcomes. Thirteen occurred within six months. Of these, six were for pre-existing sick sinus syndrome or standard sinus bardycardia indication.

One was for sinus node injury. This was already reported in major complications. Another one was for AV block and not sustained ventricular tachycardia both of which were attributable to the antiarrhythmic medication that the patient was on. Five others underwent AV node ablation and pacemaker implementation.

One had moved away from the study center and at the second center underwent flutter ablation and then moved on to AV node ablation and pacemaker implementation. One had changed insurance coverage and the new physician recommended AV node ablation and pacemaker implementation. The other three were accounted as treatment failures in our analysis.

We looked especially carefully at patients receiving pacemakers within 10 days of the procedure. There were three such cases. There was no AV node injury in any of them. All three had known pre-existing sinus node dysfunction. Patient one had a heart rate of 49 when in sinus rhythm. Patient two had sinus pauses. Patient three had sinus bardycardia and three second pauses. We do not think that these were major complications. They had pacemakers implanted for standard indications.

Interestingly what we did see was that there was episode reduction in these patients. In patient two for instances there were 34 episodes per month at baseline. This was reduced to one episode at six months. Patient three had a reduction from six episodes per month to one episode at six months.

I would like to deal with the issue of the protocol definition of clinical success in antiarrhythmic drugs. Consistent with clinical practice, antiarrhythmic medications were adjusted by the investigator as clinically indicated. The primary clinical endpoint of the study was a reduction in systematic atrial fibrillation frequency independent of the antiarrhythmic drug use. As you've already seen this was achieved in 69 of 81 or 85 percent. We feel that this is a clinically meaningful endpoint since after addressing the stroke risk the major issue with atrial fibrillation are the symptoms.

The study population as you will remember was drug refractory. It was one of the entry criteria. What we found was that there were refractory to on average three medications. They had concomitant medical conditions as we've seen.

Clinical success as defined by the protocol was "a reduction in sAF episodes while being maintained on the same antiarrhythmic drug regimen or on a reduced dosage." As determined by the clinical site, 19 of the 69 successful patients had an increase in antiarrhythmic drugs. If you take these 19 out of 69 what you are left with is 50 of 81 or 62 percent achieving the clinical success per this definition.

We analyzed the antiarrhythmic drugs in more detail. Given the current information on the efficacy of antiarrhythmic drugs, it is difficult to determine a true increase in antiarrhythmic drug regimen.

Once again to go back to who we are dealing with, we started with the 81 evaluable patients. Fifty percent or at least 75 percent reduction of AF episodes were achieved in 69 or 85 percent. Of these, 50 achieved this with a decrease or no change in antiarrhythmic drugs for 62 percent.

Antiarrhythmic drugs are a complicated issue. We sought to understand this further by doing subsequent analyses. What we did was we first recognized amiodarone to be in a special category among antiarrhythmic medications.

Dr. Kay already showed you some efficacy data which distinguishes amiodarone from other drugs. We grouped class Ia, Ic and Class III into membrane active drugs. Then the final category was rate control drugs or no drugs.

What you can see through the study is that there is a trend towards decreasing antiarrhythmic medications. Eighteen patients started on amiodarone. Sixteen finished on amiodarone. These weren't necessarily the same patients. Nine patients were able to discontinue amiodarone. Seven patients started amiodarone through the study.

On membrane active drugs, there was a trend towards decrease going from 49 at the beginning to 44 at the end. There was an increase in patients who were either on no drugs or rate control drugs only going from 14 to 21 through the study.

We did a similar analysis for patients who were deemed to be successes based on a reduction of systematic atrial fibrillation episodes. In here, you'll see that a similar trend. Amiodarone was being taken by 15 patients at the beginning and 15 patients at the end. There was a decrease in membrane active drugs going from 42 at baseline to 35 at six months. The number of patients on either no drugs or rate control drugs only increased from 12 to 19.

This is the 21 patients who the study sites determined to be an increase. We looked carefully at these. When you look at the numbers, the use of antiarrhythmic drugs was virtually the same from baseline to six months. Only two of these patients started amiodarone.

Again looking at these 21 with the increased antiarrhythmic drugs as coded by the study center, 19 of these were successful based on the reduction in systematic atrial fibrillation episodes. Ten of these 19 had 100 percent reduction. In other words, they had no atrial fibrillation episodes at six months. The increases in some of these cases were restarting a drug which was refractory at baseline prior to entering the study. The other nine had the requisite at least 50 percent or 75 percent reduction in systematic atrial fibrillation episodes.

This is another analysis looking at reaching the primary endpoint versus either a decrease, no change or an increase in antiarrhythmic drugs. As you can see, there is no statistical difference. In all of these groups, there was 80 to 90 percent success rate.

So in summary the patients in the study were refractory to an average of three antiarrhythmic drugs prior to enrollment. Interpretation of antiarrhythmic drug change as you've seen is complex. Antiarrhythmic drug use decreased or remained the same in most of the patients. Among those with an increase in antiarrhythmic drugs, very few were placed on amiodarone. Even in those patients with an increase in antiarrhythmic drugs, the marked reduction in atrial fibrillation episodes is likely to be due to the procedure rather than due to the changes in medications.

The next issue is transtelephonic event monitoring. We had redundant mechanisms in the protocol for capturing episode data looking for assessment of treatment effect. The first was transtelephonic event monitoring to capture both spontaneous and scheduled transmissions. The second was office visits where arrhythmia events were reported and case report forms filed. This included a clinical history at baseline, one month, three months, six months and 12 months to document atrial fibrillation frequency and severity. EKGs were obtained at these office visits at baseline, one month, three months and six months. In addition, patients completed an AFSS questionnaire which provided additional documentation of atrial fibrillation frequency and severity.

Specifically on transtelephonic event monitoring the instructions given to the patients were to record and transmit each systematic episode. In order to make sure that we weren't missing episodes, we also asked them to transmit at least weekly. They were instructed to perform this at baseline, one month, three months and six months post ablation.

The patients were blinded to the required number of AF episodes to enroll in the study. The transtelephonic monitoring strips were reviewed by an independent cardiologist to verify a sufficient number of atrial fibrillation episodes for study eligibility.

There was an issue of more than one transmission coming in on the same day. We wanted to make sure that these were discrete episodes. We did the analysis in a couple of different ways. The first was to use a constraint of no more than one AF episodes per hour. Under this constraint, all the patients would have qualified and the six months success rate would be unaffected. If we restricted it further to say that they could only have one episode of atrial fibrillation on a given day using this severe constraint, 79 of 81 patients would still have qualified for the study. So I don't think there's a question of not having enough episodes.

The other side line is that most of these patients can tell when their atrial fibrillation episodes start and stop. So if they recorded more than once in a day, it was usually from discrete episodes.

This is a look at transtelephonic event monitoring looking at the number of transmissions per month first. At baseline there were 15.5 transmissions. At six months there were 3.9. The second line is shown graphically in the next slide but out of the transmissions at baseline, 57 percent actually represented atrial fibrillation. At six months of the transmissions, 29 percent showed atrial fibrillation. As we've already seen, there was a reduction in systematic atrial fibrillation episodes going from nine per month at baseline to 1.3 per month at the six month mark.

We noted that 22 subjects did not transmit recordings at six months. We looked at these as being non-compliant patients. However these were patients who were compliant with EKGs and office visits at six months follow-up as well as with AFSS and that data that will be shown shortly.

But first on the transmissions that were received, the left bars show the number of transmissions per month both at baseline and six months.

The right bars in the darker blue show the percent of transmissions that represented atrial fibrillation. We've already seen that there was a drop in transmission rate from baseline to six months. What's important is that of the transmissions that were received a still lower percent actually showed atrial fibrillation at the six month start which speaks to a treatment effect from the ablation procedure.

The other mechanisms for capturing the data in patients who weren't transmitting where the arrhythmia symptoms, the EKGs and the AFSS questionnaire. This was completed by most of the patients. Ninety-nine percent were compliant with office visits. Ninety-nine percent had EKGs done at six months. Ninety-one percent had completed the AFSS questionnaire.

We continued the analysis of patients who did transmit versus did not transmit. This is looking at the AFSS mean scores at six months looking at patients who did transmit in the light blue versus those who did not transmit. As you can see, there is no significant difference in the two groups with respect to quality of life.

This was analyzed further breaking the groups down into patients who transmitted three or more transmissions at six months versus less than that amount. Once again there was no significant difference in the AFSS domains between the two groups.

So to summarize the issues for clarification, the catheters and lesion sets deal primarily with the lesion at the flutter line. Flutter ablation does not cure atrial fibrillation. Since the adequacy of the flutter line was something that the investigator had to be comfortable with, different catheters were used. Since this is a standard procedure, we did not feel that the use of different catheters at this location affects the study results.

The second issue was pacemakers. We looked at this as first a safety issue and determined that there was only one patient who received a pacemaker as a complication. The others were generally for standard indications. Of course the patients who received pacemakers in conjunction with an AV node ablation were considered as failures by our study, but in the end we don't think that patients receiving pacemakers as part of this clinical generally affects either safety or effectiveness endpoints.

The third issue was the protocol definition of clinical success. This was part of the initial study protocol. However, we feel that the primary endpoint is meaningful by itself where the systematic AF reduction is meaningful clinically. Even looking at clinical success as defined, this was achieved in 62 percent of the patients. That's still pretty good.

We looked specifically at antiarrhythmic drug use and analyzed it in different ways because it is such a complex issue. What we saw at the end was there was a general trend towards drug reduction. Patients were on an average of 1.9 antiarrhythmic drugs at the beginning and ended with 1.5. For this study sample size, there was a statistically significant decrease. Once again we feel that the success that we are seeing is due to the ablation procedure and not due to the changes in antiarrhythmic drugs.

The last issue was transtelephonic monitoring compliance. The question here is whether we were unable to capture episode data. I think you've seen that because of the redundant mechanisms that were in place that the episode data was captured and there were significant improvements seen.

Just once again to highlight the results of the study, the primary endpoint which is the reduction of AF episodes at six months was achieved in 69 out of 81 or 85 percent. Patients having no atrial fibrillation at six months numbered 44 of 81 of 54 percent.

The protocol definition of clinical success which brings in antiarrhythmic drug use was achieved in 50 of 81 or 62 percent. Thirty-six of the patients had a reduction in antiarrhythmic drugs through the study, 44 percent. Six patients or seven percent were off antiarrhythmic drugs by the end of the study. You have already seen the quality of life data. There were significant improvements in several categories.

On safety, there was no mortality for zero percent. Major complications occurred in four for a rate of three percent.

So in summary I think what we are seeing here is that the use of REVELATION_ Tx to deliver right atrial linear relation to address drug refractory systematic atrial fibrillation is first safe and reasonably effective. Thank you very much.

CHAIRMAN LASKEY: Thank you. You had originally requested some additional time.

MS. WOOD: You originally indicated your presentation would take an hour. We've now exceeded that. I would urge you to try to wrap the presentation up as soon as possible.

DR. KAY: Thank you. If I could just have a couple of minutes to do that. Let me just say that as this study was being performed I must say I was skeptical that the right atrial ablation would be effective to treat atrial fibrillation. Let me point out that really this is the first multi-center clinical trial of catheter ablation for atrial fibrillation to have been completed.

If you look at it, they met all the prespecified criteria. I think the results show that this is reasonably successful. It's a moderate degree of success. It does improve quality of life. It had a surprisingly high reduction in the frequency of systematic episodes. It was accomplished with a very low risk of serious side effects. I think that's important.

Also if you look at the centers that did it, it was done at centers across the spectrum of medical centers and experience. If you look at the risk benefit ratio, remember this is a right sided procedure that's going to be I think widely applicable. It's a relatively simple procedure. The risks are extremely low especially when they decided to cut out that anterior line. It's going to be widely applicable to physicians who don't really feel comfortable going to the left side.

I think it's likely that this simple right atrial ablation procedure will be offered to patients who have highly symptomatic paroxysmal atrial fibrillation. Remember these are people that had failed a mean of 3 drugs. They had to fail at least two. It was done very safely. I think this will be done as a first line therapy for patients who fail drug therapy before considering any kind of a more invasive left sided procedure.

This is the indication the sponsor is requesting which is that it's indicated for the treatment of atrial fibrillation in patients that are drug refractory with paroxysmal atrial fibrillation. After looking at this and looking at it for the holes, it's very robust data. I think that the results in my firm opinion are that the results from this study provide reasonable assurance that this REVELATION_ Tx system is both safe and effective for this intended use. Thank you very much.

CHAIRMAN LASKEY: Thank you. An excellent presentation. Before we move to the short break, I wanted to ask the panel participants if there were any questions that you wanted to ask the sponsor before we get to the substantive issues.

As a non-electrophysiologist, I had one. These patients who are highly prone to atrial fibrillation by selection, what was the Cumidin program prior to selection for the study? Then how did you handle it post ablation?

DR. KOCHERIL: That is a good question but unfortunately that's one where I don't have numbers at my fingertips to give you. The general scheme though was that Cumidin was recommended for all these patients. The Cumidin was not discontinued at the end of the procedure. What we are looking at is a fairly short window of six months so it's too early to tell whether there would be sufficient remodeling or other factors to reduce stroke risk. The quick answer without giving you detailed numbers is that it wasn't changed.

CHAIRMAN LASKEY: I guess a corollary to that I can assume that since the TEE was done in close proximity to the procedures so you didn't identify any atrial thrombi.

DR. KOCHERIL: No. Exactly, if an atrial thrombus was identified, they wouldn't have entered for the ablation procedure.

DR. SCHWARTZMAN: Abe, before you sit down, a couple of questions. One is practical or procedural and the other is a clinical endpoint. With respect to the procedure, you mentioned the need for multiple passes to create a given lesion. The problem with multi-electrode ablation when we started was single pass. The difficulty obviously is topography, contact, etc. Can you elaborate on that? Indeed the procedure times speak I think apart from the duration of getting the flutter line done and those who had trouble to the difficulty in creating a linear lesion even with a multi-electrode catheter. Would you elaborate on this issue as to the number of passes? What are we actually doing with multi-electrode catheters to construct a linear lesion?

DR. KOCHERIL: Sure. A very good question. In order to ensure an adequate linear lesion, the eight electrodes or several of them at least need to be making contact with each pass. The ideal situation would be that the catheter would be placed against the one of the walls of the atrium and all eight electrodes would make contact. You would deliver an adequate linear lesion using that segment.

Even with that especially in the posterior lateral area, you would need to move the catheter at least once to make sure that the lesion reaches from SVC to IVC. So even in the perfect setting, there would generally be two passes to make sure that the linear lesion is contiguous and adequate.

In the other areas, there is more difficulty in creating that linear lesion. Typically at the septum you could get a pass where you enter the atrium from below and are able to advance the catheter straight and have it sit on the septum. More usually, not all eight will make contact during that pass. Once again it's up to the investigator to keep track of which electrodes are making contact and where the lesion has been delivered and then move the catheter up or down to make sure that once again you have anatomical obstacle to anatomical obstacle lesion formation.

What I do personally in many of these cases if the straight line approach doesn't work is to come back with a curve on the catheter to make a loop. The advantage there is that if you have a loop on the catheter you are able to push it up against the atrial wall and have better contact and more assured contact. That would lessen the procedure time because you'd have more adequate lesions with the electrodes that are making contact and there would be fewer passes required to make the lesion.

I think it is important to make adequate lesions. This was something that was stressed as your investigators were coming on board. I think we've had success as a result of accomplishing that.

DR. SCHWARTZMAN: I presume the vast majority of investigators were limited to fluoroscopy for guiding catheter position.

DR. KOCHERIL: Limited in?

DR. SCHWARTZMAN: In terms of their ability to visualize where they actually were relative to the endpoint.

DR. KOCHERIL: Yes, at the beginning of the study all there was was fluoroscopy. In later stages especially in Phase III, a number of the investigators had acquired non-contact mapping systems. A couple of the investigators were actually confirming lines using CARDO.

I think the most common scenario was that investigators got the SI system and they were confirming conduction block after each linear lesion using the SI system. But that was not a systematic part of the protocol. The only requirement was fluoroscopy which has its limitations.

In my own hands, I'd had been using Loca lessa just to see that there aren't significant gaps that we can track a linear lesion in a crude way using that system. But different techniques were employed later in the study but that was not part of the study design.

DR. SCHWARTZMAN: With respect to that, can you comment as you know the ability of local electrogram quality amplitude fractionation to act as an indicator of contiguity, continuity or whatever you want call linear lesions whether or not that is a relevant, desirable endpoint? That is the stated endpoint in this study based on your own experience of your collaborators with more traditional conduction based assessment of complete block. What are your thoughts as to what you get from this technology?

DR. KOCHERIL: Another very good question. As you know, this isn't the only catheter system to look at that as an endpoint. I was involved in the Thermocool catheter study recently. That was one of the stated ways of knowing that an adequate lesion was being delivered. So it is used and it is a standard of sorts within the AP community. However what you are raising as the question is how do you know that it's an adequate lesion after seeing that.

Quite frankly we don't know. I think the big picture here is that we got good results. We got symptomatic AF episode reduction and I think we were doing some good. The animal data is very helpful. It shows that adequate transmural contiguous lesions can be delivered with this catheter system employed in the fashion we were using it in the clinical trial. But aside from telling you that that was what was looked for, there's no good way to confirm that it confirms an adequate lesion.

DR. KAY: I just want to make two comments to that question. When the study was designed, the question was do you have to confirm these linear lesions. The point you bring up are right on target, David, in terms of it's hard to even make a flutter line in patients. That can take two or three hours particularly with standard catheters when you really have to confirm that it's complete and linear.

The strategy that was employed in this clinical trial was that we take this floppy catheter. We draw these linear lesions. We don't confirm anything. A part of the protocol was not confirming that the lesions were continuous or anchored or that there was electrical block or anything because as soon as you throw that in, it becomes an incredibly complex procedure. It adds another three or four hours of time to do the mapping. It marries the system to some complex mapping system.

The strategy was you take this catheter. You draw these lines based on these electrogram markers that yes, you've created a lesion. That's the only evidence that you have that the lesion has been created.

I think what was striking when you used this catheter is you burn with this catheter and the electrograms really disappear. You really have very good confidence that you've created a lesion. Are they all complete? No. Are they all transmural? No. Are they all anchored? No. But the strategy was you put these lines in in this position and here are the results. Does that help reduce the atrial fibrillation burden in atrial fibrillation? That's an important aspect of the protocol. It was a strategy rather than the science of linear lesion creation and these other things.

DR. WALDO: This is Al Waldo. Can I make a comment here too because I too was concerned about the definitions of block. I think anybody who looks at that would be very skeptical. The only one that is a definition of block is the double potentials what you called "split potentials."

It seemed to me that was what was used for a measure of procedural success. Now you are saying that it wasn't. Is that the idea that ?- signal only means self-conduction in most instances while amplitude signals don't mean there's no conduction? I was wondering why those endpoints were chosen but were not used.

DR. KAY: Al, your points are on target. Clearly the endpoint of the study was not creation of three linear lesions with demonstration of complete block across each one of these linear lesions. The objective of the study was to deliver lesions in certain trajectories using this catheter and to have some measure that you've created a lesion which a reduction of amplitude of the electrogram and so forth but not that you've definitely completed a transmural, complete linear lesion that has no gaps in it.

So your question is a good one and I think that speaks to this as a strategy. The strategy was not to have complete contiguous lesions but to deliver lesions. As far as split potentials, many times you see split electrograms. Sometimes you just see a reduction of the electrogram amplitude but I think you can always be quite confident that you would able to deliver a lesion that had some measure of the fact that a lesion was created at a given location.

DR. WALDO: I think that says delivery of the lesion is there but the notion that you obtained block is not demonstrated in my opinion. Now the other thing I would ask is since this was a totally empiric approach, would you review again with us or would someone review with us again why you selected these sites to do it? Then as a correlate to that, do you know that you needed all those sites? It's so empiric in the first place but a little more about that would help me.

DR. KAY: Again a very good point. When the study was designed, the question was where should these lesions be. Do we have any notion? The only data that at least I'm aware of is the Cox Maze procedure they have flutter line and then a cava line. About the time the study was designed there was a number of reports.

You probably remember Michelle Hassenger's paper that was on the cover of "The Journal of Cardiovascular AP" showing atrial fibrillation terminating during a burn with a right atrial lesions that was again a inner cava line and that had a lateral line compartmentalized in the atrium in a horizontal direction. Everyone had a slightly different lesion set.

So the question is which of these is the most critical lesion. We don't know. It's like in the Cox Maze procedure. We don't know which one of those lesions is critical. There has now been some work subtracting some of them and the results seem to decrease the more you subtract. Once this type of technology gets approved and available, investigators can do substudies looking at peeling lesions away and seeing if you just do a inner cava line and a flutter line do you get the same results as if you put in all three lines.

So I guess the answer is we don't know. This was the best guess that the folks that designed the lesions, Jeremy Rusk and David Keane, looking at the literature available in 1996. But it is an empiric approach.

DR. WALDO: I think that's important, Hugh, because as you well know with beginning to find out about mechanism of atrial fibrillation most of the time it looks like there's some kind of drive and most often it's on the left side. The original Maze operation as everyone well knows was designed to treat multiple ?- which don't appear to be very often in the mechanism of atrial fibrillation. The striking data that was pointed to of the Maze procedure shows that less is as good or better. The interesting thing about the Maze operation is that it doesn't make any distinction between paroxysmal and persistent or even permanent atrial fibrillation. So all these empiric approaches I find that it's hard to know how to evaluate them. Where and why to put lesions is an issue still.

DR. KAY: Yes, I think the other point which you make is the surgeons have a difficult time because once you do a thorocotomy you have one time in to get the job done. You're not going to bring patients back for redo procedures. One of the things that the electrophysiology community has an opportunity to now as new tools become available is to start looking at this more critically and trying to figure out what lesion sets are better in a procedure that you do as an out-patient procedure from the leg rather than from a thorocotomy.

But all your questions are terrific. I think as a scientist studying reentry, this empiric approach probably goes against your general grain in its fact that we aren't proving this and proving that. But this is a clinical strategy that we are looking at today.

DR. WALDO: As a clinician, it does too I might tell you. There are some other questions as well if I can follow up. We heard no mention at all about a weight control. Of course, weight control is a very important part of symptoms and you are dealing with symptomatic atrial fibrillation. Are there any data you can give us about how carefully this is looked at and what impact this may have had on assessment of the whole picture here?

DR. KAY: Certainly in weight control and drugs, patients were managed by the clinical investigators, how they usually would be managed so that rate control could go up or could go down during the course of the study. As we looked at the fine studies if they switched from Cardizim to Verapamil or Cardizim to a beta-blocker, it's hard to say again what's up or down or a lateral move.

If Digoxin is stopped, is that a decrease? If Digoxin is added, is that an increase? If you switch from Verapamil and a beta-blocker to just the beta-blocker at a higher dose, is that an increase? Is that a decrease? I think what we basically came away with is that it's very hard to say that they are up or they are down. So that's how we ended up with these broad groups of amiodarone going up on drug therapy. Amiodarone stoppings are going down on drug therapy. I think everyone would agree on that.

Then having these broader classes of membrane active drugs versus rate control drugs in most patients had a general decrease in drug therapy whether they were going off amiodarone or off membrane active drugs or the same or decrease rate control drugs. But it is very difficult as you get into these dose questions. Then you have concomitant hypertension and other issues that a patient is being treated for.

DR. WALDO: Let me be more specific. Were there any criteria for maintaining adequate rate control? For instances if you decided that when your patients returned, you wanted their rate less than 80 for any number of reasons, they might not be symptomatic. Whereas if you didn't have such good control and may recur at a ventricular rate of 120, they would be symptomatic. It might be that you could have missed lots of episodes because someone was taking good care of the rate control. Do you have any data or anything you can enlighten us on this aspect of the study?

DR. KAY: Well the study design was that antiarrhythmic drugs or drugs related to atrial fibrillation would be kept unchanged. So that was the study design. The study design was not to do the ablation and then double rate control drugs and see how they do. The study design was that antiarrhythmic therapy or therapy for atrial fibrillation was maintained during the course of the study.

The protocol did not specify decrease or increase in any way. What we see in all studies is physicians manage patients as they think is best for the patients. So we saw some motion as I mentioned in calcium blockers, beta blockers. Then how do you interpret that? It becomes very difficult as you look at each patient through a microscope. But overall I believe looking at the data that I've seen that there was overall no change in the rate control strategies that were used.

DR. WALDO: Thank you.

DR. SCHWARTZMAN: I wonder if I might follow up on where Dr. Waldo was going which was characterizing burden. Having been involved with studies like this, I can empathize that it's extremely difficult short of an implantable full encompassing holder recorder to get to the nitty-gritty here.

Nevertheless I think what's well established is that based on the context of the study is support involved, psychosocial issues patients naturally characterize lower burdens after intervention than before. I'm not criticizing the design of the study. Without a placebo or parallel arm, I don't think that was possible. But nevertheless I need a better understanding of burden not in relation to symptoms as much as what was happening electrocardiographically.

So two specific questions. One is it's unclear to me that patients were complying post ablation with these weekly mandatory transtelephonic monitoring issues. Can you give me a little better flavor whether they were complying? I just don't understand. If so, what the data showed asymptomatic, just routine weekly recordings which as I understand it was part of the protocol?

The second question is there was a high rate of electro 12 lead EKGs at follow-up. Can you give us any data as to what the atrial rhythm was at those times?

DR. KAY: As far as the study, it was designed on this whole issue about how do you measure atrial fibrillation burden. It was very challenging other than this continuous 24 hour Holter which would make your skin fall off or put in the Reveals (PH) but they can't even detect atrial fibrillation with much accuracy. So it's a very difficult question.

The way this study was designed was to do it the best they could possibly think. So what did they do? One had an outside company in charge of it. So it wasn't something that was in-house. It was an outside event monitoring who was in charge.

Each patient when they went home was asked to give a day that was the best day that they were supposed to transmit. So they may say Monday. They may say Wednesday. They may say Friday. Every week they are supposed to transmit whether they are in sinus rhythm or not. They are just supposed to transmit their rhythm.

What we found is over time, that compliance with that decreased. By six months, I think 20 of the patients transmitted never even though they were supposed to transmit every week. They transmitted none. What the event monitoring company was to do not if they don't transmit just do nothing but really go after the data. They would call the patient. If the patient didn't transmit on Monday which is their day, they would call the patient on Tuesday and say "Could you please try to transmit". They would get on the telephone or leave a message.

When they got through, the patients would say "Well I'm feeling just fine and I'm not having fib so I didn't transmit." They really struggled to try to get very good compliance but what they found is what we see. If a patient has the procedure today, six months later they are feeling and if someone is calling them and telling them to start transmitting these episodes for no benefit to them, then they just have a hard time motivating themselves to do it.

I think that accounts for the fact that there was less than perfect compliance which speaks to this notion that we need some implantable monitor to do this in future to make it an easier job. I think a heroic effort was taken. All this is documented. The event monitor documents. We called X patient on this day and the patient said that they are doing fine and they didn't want to transmit. These are in the logs from this event monitoring company.

As far as the EKGs, these patients did show up to their six month appointment and said that their quality of life was better. They are doing fine. They aren't having an atrial fibrillation and the EKGs were obtained which showed almost uniformly sinus rhythm. There were sample patients that showed atrial fibrillation at that six month follow-up point but that was five to ten percent, a very small number. Most of them were feeling good. They are having no atrial fibrillation. They are in sinus rhythm on 1280 EKG at the six month follow-up point.

I was encouraged by the fact that these weren't patients ?- The skeptic would say that all these patients were on amiodarone at the end. Well, that's obviously not the case. The skeptic would say that these patients have no benefits so they just disappear and didn't complete the study. They did complete the study. They showed up to their visit. They told people they were feeling fine and not having any atrial fibrillation and the EKG largely showed no atrial fibrillation.

Those that did transmit a lower frequency were atrial fibrillation so there's lots of things speaking to that. But it also speaks to the fact that it's very hard to get this perfect compliance in patients largely who are active, patients that have lives and jobs to do and things like that.

DR. SCHWARTZMAN: This issue of reduction of symptoms associated with atypical AF management floating around whether that's due to psychosocial factors or better rate controls Dr. Waldo mentions or even anesthetization for lack of a better term based on collateral damage to cardiac enervation, can you comment on whether there was some undercurrent observed in your cohort along the lines of a reduction in per event symptom burden or is it just too much of a vague science?

DR. KAY: I think it is a vague science. I think one of the questions which the community has been thinking about this is why is this catheter that's just doing a right atrial approach seeming to have such really striking benefit to these patients. The mechanisms that have been proposed are either reentry rotors that are in the right atrium. Even though most in the left atrium, there are some in the right atrium.

There's a recent paper presented by Ashian Chen looking at rotors near the SVC. So the fact that two of the lines are anchored to the SVC may have taken care of rotors going around the SVC or perhaps even triggers in the SVC. Sonny Jackman likes the notion of autonomic function and maybe somehow this is affecting the autonomics.

We don't have a clear understanding of exactly what is the mechanism of benefit and of the lines which is the critical line. I think we feel quite confident that it's not the flutter line because we know you do flutter ablations and what do you see in the follow-up but atrial fibrillation. But all we know is this lesion and you get these results.

As far as the per episode symptoms at least in my own group of patients, it seemed like when they had atrial fibrillation these patients tend to know it and they tend to transmit the episodes. There was really fairly good correlation. In the beginning prior to the ablation, only half the episodes show atrial fibrillation even when the patient thought they had atrial fibrillation.

But as time went by, even less showed atrial fibrillation. So these patients were highly tuned into their heart rhythm, would transmit and would show premature atrial beats. I remember in many of our patients we would get these transmissions come through and they show APBs. The patient may have thought they were in atrial fibrillation but they were in sinus with APBs. I think it's a very tough problem to wrap your hand around well.

DR. SCHWARTZMAN: You don't think patients were rendered asymptomatic or minimally symptomatic in regard to their AF by the ablation procedure.

DR. KAY: Yes, I do not get the impression that patients were now in atrial fibrillation and their symptoms were gone. They would show up at the six months follow-up appointment saying that they felt fine. You do an EKG. Even though they felt fine thought they were in sinus or in atrial fibrillation. I've seen that with some of our pulmonary vein patients but I did not see that with our center's patients that were in the study.

DR. SCHWARTZMAN: I just have one housekeeping question. I just want to understand this issue of increase. The way you presented it made it seem like you left to the nurse coordinators at each center to characterize what that meant. That makes me nervous. So for example on slide 100, you show 69 patients with successes, 42 on a membrane active drug. Seven of those patients went to amiodarone at six months.

Yet for those characterizing increase, only two of that group of patients from membrane active went to amiodarone. I would think all seven would be characterized as an increase of change from Type I or III or nonamio to amio. I just want to make sure that I understand how this was done.

DR. KAY: Let me give my version of this and then Abe will give his. On the form, it said in the follow-up is the patient on more or less of the same antiarrhythmic drugs. Did the patient have an increase? That was really independent of the company. That was the investigator that coordinated to check that box or not. To agree that it's completely out of any post hoc analysis, we can't go back in now and change it. They checked the box the way they thought the box was supposed to be checked. That's what we have.

What we found as you get in and look at it some of the times they would say it's an increase because you went from amiodarone to flecainide and a beta-blocker. So the number increased but you went off of amiodarone. So that's why we looked at it again as far as these broad groups which we thought were as an electrophysiologist when I looked at the data this is something that everyone could understand. You start the amio. That's more. You stop the amio. That's less. There were some patients that went from Sotolal 80 BID to Sotolal 160 once a day. It seems funny to give it once a day but is that an increase or a decrease. The dose is up. The frequency is down. It just becomes difficult. Let me give it back to Abe.

DR. KOCHERIL: Dave, you were astute to pick up who was filling out that information. It was supposed to be determined by the study site and very often it did flow to the study coordinator to do that. As Hugh illustrated, this is very complex in determining exactly what happened. That's why we redid the analysis breaking it up into amiodarone, membrane active drugs and rate control drugs because that's easier for the EP community to understand. I think that basically looks good from our standpoint.

The issues were just as Hugh outlined what was coded as increases. If there was an increase in some fashion, there was one patient who was on both flecainide and amiodarone at baseline. Flecainide dose was increased. The amiodarone was discontinued. Yet that patient was counted as an increase.

These issues of doses going up and down even in some of the patients who received amiodarone and would legitimately be considered by all of us in EP to be increase, a couple of those patients were actual refractory to amiodarone at baseline and then did well on it afterwards.

So yes, it was an increase in medications but that speaks to a treatment effect if it was refractory at baseline and they responded later. But it is a complex issue. We analyzed it in multiple different ways because it is such a complex issue. In our final analyses breaking it up into amiodarone versus membrane active versus rate control, the overall trend is to decrease medications during the study.

CHAIRMAN LASKEY: And the panelists will have additional time for query.

DR. WALDO: This is Al Waldo. Can I ask a couple of other questions now?

CHAIRMAN LASKEY: Al, how about one question and then we'll break.

DR. WALDO: One point and then one question. Very brief. Hugh was saying that the A2 front isthmus lesion doesn't affect atrial fibrillation. The data from Walling lab showed 15 percent of cases were affected that way. They speculated a rotor that involves the flutter isthmus. So I think there are data that that lesion may have some effect.

The thing I wanted to get to is what were the overall number of patients on amiodarone. The reason I ask that is it gets back to one of my earlier points about rate control. Amiodarone is an excellent AV nodal blocking agent. I still have some concerns about symptoms and rate control and how you assess burden. I would like to know really what percentage of the 81 patients that you were dealing that were analyzable were on amiodarone.

DR. KAY: Eighteen patients started on amiodarone and 16 ended up on amiodarone. That decreased. As far as the flutter line, I think your point is correct although I don't think that's major effect of this protocol. Certainly for pulmonary vein isolation, very few centers do a flutter line as part of that procedure. Clearly all we know is that these three lines together decrease symptomatic atrial fibrillation episodes. The line that we have actually confirmed conduction block is the flutter line.

DR. WALDO: But you are making that point though and that is requisite in most instances that you need antiarrhythmic drug therapy. Isn't that right?

DR. KAY: Yes, most the patients. Only six patients were off all drugs and 21 patients were off any membrane active or amiodarone drugs. That would be 59 patients were still on antiarrhythmic agents at the end of the study.

CHAIRMAN LASKEY: I'll take advantage of this pregnant pause to break and let's meet back in 15 minutes. I have 11:20 a.m. so 11:35 a.m. we'll have the FDA presentation. Thank you. Off the record.

(Whereupon, the foregoing matter went off the record at 11:22 a.m. and went back on the record at 11:42 a.m.)

CHAIRMAN LASKEY: Back on the record. Thank you all. Prior to continuing with the FDA presentation, Ms. Wood had an announcement.

MS. WOOD: I need to make a point of clarification. Dr. Thomas Ferguson was read in the voting status statement in error this morning. He is not participating in today's meeting. Also I would like to once again remind the speakers that when they come up to the podium to speak although you have previously introduced yourselves for the benefit of the transcriptionist, please identify yourself once again before you begin speaking. Thank you.

CHAIRMAN LASKEY: And now may we have the FDA. Thank you. Welcome.

MS. DEMIAN: Thank you. Good morning. My name is Cindy Demian and I'm the FDA Lead Reviewer for the Cardima REVELATION_ Tx Microcatheter with NavAblator^TM RF Ablation System submitted under PMA P020039. The FDA presentation will provide an overview of the following: identify the FDA review team members; provide a brief overview of the Agency's interactions with Cardima; provide a brief summary of the description of the products; discuss the non-clinical evaluation and summarize the major outstanding device performance issues to date; provide a summary of the clinical and statistical evaluation and conclusions; and identify the FDA questions for the panel.

The FDA team was comprised of myself, our medical officer, Dr. Lesley Ewing, who will present the FDA clinical review summary, our statistician, Dr. Heng Li, who will present the FDA statistical review summary, Dr. Nick Jensen, who performed the animal review, James Cheng, who performed the electrical engineering review, Lisa Kennell, who performed the sterilization review and Barbara Crowl, who performed the bioresearch monitoring review.

The sponsor proposed the following indications in their PMA application as well as in their panel pack. It is worth noting that this indication mentions creating a set of continuous linear lesions along the lateral and septal walls and along the isthmus in the right atrium. However the sponsor has chosen to use the shorter version which you see here which mentions ablating with a set of continuous linear lesions in the right atrium.

I would like to briefly discuss the history of the Agency's interactions with Cardima. In December 1997, the sponsor's feasibility study was approved which included 10 patients.

Then in July 1998, FDA Advisory Committee made recommendations for atrial fibrillation clinical study designs. In particular they recommended a single arm study where the patient serves as their own control. This Advisory Committee recommended a 75 percent decrease in the frequency of symptomatic episodes or cure as considered as clinically significant endpoints in the treatment of atrial fibrillation.

In August 1998, the sponsor submitted their first progress report on their first five patients. However there were problems with creating the isthmus line with the REVELATION_ Tx. The FDA recommended that Cardima could either pursue a licensing agreement with another company for use of an approved standard four millimeter ablation catheter or Cardima could design their own catheter in order to complete the procedure. Cardima opted to design their own standard four millimeter catheter. In addition the Agency informed Cardima that the use of non-investigational device would be considered a clinical failure.

A few months later, in December 1998, the sponsor submitted their progress report from the next five patients, now a total of 10. However despite the FDA's concerns, there was still wide spread use of non-investigational catheters in Cardima's study.

Then two years later in May 2000, Cardima was granted approval to begin their pivotal trial, Phase III. In addition the NavAblator^TM 4mm catheter was added to their IDE. At that time, the Agency's thinking was that 80 patients would be treated with the new catheter and there would be sufficient acute and chronic effectiveness results within a narrow enough confidence interval to properly evaluate the NavAblator^TM. Again FDA continued to communicate to the sponsor that the use of non-investigational catheters were considered failures.

Then in June 2000, the sponsor met with the Agency where the company agreed and stated in their meeting minutes that they do not want to pool the feasibility with the pivotal.

A year later in May 2001 Cardima submitted a progress report on their Phase IIb patients. FDA informed Cardima of our concerns which were primarily based on the feasibility results. The first concern was that there was patient non-compliance with transtelephonic monitoring. The second concern was that there was varying definitions of acute success.

The basic components of the RF Ablation System are the catheters, the accessories and a compatible RF generator. In particular, the catheters include the REVELATION_ Tx Micro Ablation Catheter and the NavAblator^TM 4mm Ablation Catheter. The accessories consist of the Naviport^TM Guiding Catheter, the REVELATION^TM Tx Select Switch Box and the REVELATION^TM Tx Cables. Investigational generators were used in this clinical study. However it is worth noting that a compatible generator is needed in order to properly function with the Cardima catheters.

Specifically the REVELATION_ Tx Microcatheter is a 3.7 French single use catheter. It is steerable and has non-deflectable distal tip coil. It's flexible and has non-electrically active tip. It has eight electrodes and eight thermocouple temperature sensors on the distal end of the catheter. The ablation electrodes are 6 mm in length and it has a 90/10 platinum iridium tip. Radio frequency is applied to each individual electrode to produce thin linear RF ablation lines. This catheter is used with a deflectable guiding catheter, the Naviport^TM, to properly position the distal tip.

The NavAblator^TM is an 8 French 4mm ablation, single use, catheter. It has an electrically active deflectable tip. It has four electrodes including one embedded in its tip just proximal to the thermocouple. This catheter is intended to create spot lesions from its tip and is designed to be used without a guiding catheter. This catheter has a control mechanism in the handle that activates a pull wire to steer and deflect the tip which can be locked in place when the desired curve or position has been achieved.

The Naviport^TM Guiding Catheter is used to aid in the positioning of the REVELATION^TM Tx. This guiding catheter is sterile, single use catheter that has a dual lumen, a device lumen and closed pull wire lumen. This device has already been cleared through the 510(k) process for the same intended use. However it is integral to the REVELATION_ Ablation System and so it is described here as an accessory.

This catheter has a deflecting mechanism that allows for its tip to be straighten while be inserted into and while it's deflected into the heart. It also features a friction locking mechanism that permits the distal tip to retain its deflected shape once the catheter is in position for ablation.

FDA's preclinical review goals were safety and reliability. The safety was to ensure that the device has been appropriately designed and tested and that the safety features have been qualified for use. The second preclinical review goal was reliability which was to ensure that the device design and manufacturer provides assurance of consistency with performance of the catheters.

The Catheter Preclinical Qualifications included biocompatibility testing of catheter materials, reliability of the catheter design, mechanical and electrical testing of the catheter performance and qualification of the sterilization procedures.

The Preclinical Conclusions are as follows. Preclinical testing performed by the sponsor is designed appropriately and met the pre-specified pass/fail criteria. In addition, testing shows that the device can be reliably manufactured to meet the product specifications.

It is important to note however that there were several device failures for the REVELATION_ Tx catheter that occurred in this clinical trial that were not predicted by the clinical testing. We are working with the sponsor to determine how these failures will be mitigated. I would like to mention that the customer experience reports do not discuss all of the issues in detail. However there are a few points that are worth noting.

For example, there were five complaints in which the Coagulum formed on electrodes. There were seven complaints of rough burrs or in which there was delamination of THV which is a coating to improve the conductivity of the electrodes.

In addition, there were 21 electrical failures which included thermocouples and electrodes. Specific examples of electrical failure complaints included such comments as "the thermocouples never registered temperature". There were complaints of electrodes in specified positions not ablating. There were noisy ECG signals. There were complaints of electrode noise or cross talk.

In addition, there were also three complaints regarding failure of the cable which is an accessory to the Cardima system. These cable failures are not included in this table. This concludes the preclinical evaluation. Dr. Lesley Ewing will now present the clinical summary portion of this presentation.

DR. EWING: Good morning. I am Lesley Ewing and I do not have any financial conflicts to report. As you've heard the Indications for Use statement that Cardima proposed for the clinical trial included treatment of the patient with paroxysmal atrial fibrillation and using a set of linear lesions along the lateral and septal walls and along the isthmus. This is the indication upon which the clinical trial was based.

As you also have heard, the Cardima catheter system is composed of two catheters. The REVELATION_ Tx which in the proposed indication for use in the PMA states that it's "intended for the creation of continuous liner lesions for the purpose of interrupting arrhythmia pathways." The NavAblator^TM was to be used when the REVELATION_ Tx was not used to complete the isthmus lesion.

It was a single arm, non-randomized study which began in 1997. There were three study phases. The submission of the PMA includes data from Phases IIb and III. The major difference between the latter two stages was the introduction of a new catheter to the device system.

The inclusion and exclusion criteria have been presented by the sponsor.

The patients underwent a baseline pre-ablation monitoring period during which 30 days prior to the ablation they were to achieve a minimum of three episodes of symptomatic paroxysmal atrial fibrillation to be eligible for ablation. It is the FDA's information that the patients were aware that a minimum number of episodes were required. Re-screening was allowed and for re-screening the patients if they did not meet the initial baseline screening period they were required to have a total of nine episodes over 90 days, so three episodes per 30 days.

The ablation procedure has been described and as specified by the protocol, it includes three linear lesions: the posterolateral, the posteroseptal and the tricuspid isthmus. The anterior lesion was optional. According to the protocol, all lesions were to be attempted first by the REVELATION_ Tx catheter. If the tricuspid isthmus lesion was not successful, then the NavAblator^TM catheter could be used.

The follow-up was at one, three, six, 12 and 24 months. The 24 month follow-up was by telephone. There are quality of life questionnaires at three and six months. Those were compared to the baseline quality of life questionnaires. The patients were given a transtelephonic monitor which was used up to six months. As has been previously discussed, the transmissions were to be compulsory weekly during the first, third and sixth month even if the patient was without symptoms.

The primary effectiveness endpoint was reduction in frequency of symptomatic episodes during the sixth month for a thirty time period post ablation procedure compared to the baseline period while on the same medications or reduced dosage. Because the endpoint was to be measured during this discrete sixth month post ablation period, the Agency thinks that it is important to try to standardize medication use.

If the patients had five or more episodes in the baseline period, they were called successful if they had 50 percent less episodes during that sixth month post ablation. If they had three or four episodes, they were required to have 75 percent decrease. That worked out to be that if they had three episodes in their baseline period, they were required to have zero episodes to be called a success. The episodes were to be counted by a measurement of if it was truly atrial fibrillation on their transtelephonic monitor.

The secondary endpoint was improvement in the quality of life.

The procedural success endpoint was as also has been discussed the demonstration of one of the following at lines of ablation during sinus rhythm: reduction in amplitude; fragmentation or widening of local electrograms; split potentials; or increase in pacing threshold. Prior to Phase III, the pivotal trial, the increase in pacing threshold was required by the study protocol but that requirement was dropped in Phase III to measure the increase in pacing threshold.

The safety endpoint was incidence of complications. Major complications were for 7 days and adverse events in the 24 months post ablation.

There were 120 patients that had the procedure performed and there were 116 that had verified data. The demographics I will not go over because the sponsor has gone over these.

The mean procedure time was 250 minutes plus or minus 123 minutes. The range was 100 to 755 minutes. Mean fluoroscopy time was 47 plus or minus 46 minutes with a range of two minutes to 265 minutes.

As you will hear for the rest of my slides, there are numbers from the FDA review which are slightly different from what the sponsor has reported. I will describe why those numbers are different. But also in an attempt to be transparent in the Appendix to the FDA clinical review, the data upon which the review team made these decisions is presented and available to verify the numbers one way or the other.

As I said there was 120 patients that had the linear ablation procedure and 116 patients had verified data which Cardima presented to the FDA. There were five patients that when the independent cardiologist reviewed the transtelephonic monitoring, there was a dispute upon the number of atrial fibrillation episodes. Those patients are not included in the effectiveness cohort per the FDA.

So there were 111 patients that had unambiguous number of baseline episodes. Out of that 111 patients, there were 21 patients that had less than six months follow-up. Two were deemed lost to follow-up and the Agency's number of effectiveness cohort is 88. The patients that left the study or went on to have a further procedure to treat atrial fibrillation are included in this number of 88.

So the denominator for safety for the adverse events is 116 patients and the effectiveness cohort is 88 patients.

The number of atrial fibrillation episodes at baseline is shown in this table and has been discussed. It also has been mentioned that there is a range of accuracy of whether the patient actually was having atrial fibrillation when they transmitted symptomatic episodes. On this graph you can see number of patients on the vertical axis and the horizontal axis is the percentage of the transmissions that actually were symptomatic atrial fibrillation.

There is a wide range of this. The percentage of symptomatic transmissions that were ultimately diagnosed to be atrial fibrillation ranged from 12.9 to 100 percent. It is also unknown whether each transmission represented a discrete atrial fibrillation episode.

I do have available for the panel's review if requested the actual logs from the symptomatic transmissions. The sponsor submitted detailed information on 89 patients. There were 62 of these patients that transmitted multiple times per day on at least one day of the 30 days. Five patients had atrial fibrillation transmissions within five minutes. Seven patients had atrial fibrillation transmission at least twice within 30 minutes. Twenty-six of these patients had transmissions twice between one and three hours.

In terms of the ablation procedure that was actually performed, not all of the patients had the same lesion set performed. The REVELATION_ Tx catheter, the microcatheter, was used for all septal and lateral linear lesions. Some patients had only a non-investigational catheter used for the tricuspid isthmus lesion. The non-investigational catheters used were from five different manufacturers and included a cooled tip catheter.

As you can see from this chart, the majority of patients, 82 percent, did have the standard protocol defined linear lesion set. But there are a number who did not.

This chart shows the breakdown of the different catheters that were used for this procedure. In Phase IIb, the NavAblator^TM was not available to the investigators but you can see that of the patients who received a tricuspid isthmus lesion, 21 out of the 33 had the REVELATION_ Tx catheter used first. But 12 had a non-investigational catheter used. So the investigational catheter plus a non-investigational device was used.

In Phase III when the NavAblator^TM was available, the REVELATION_ Tx was used first for the tricuspid isthmus lesion in only eight patients. And the NavAblator^TM was tried first in 57. But eight patients continued to have an non-investigational catheter used instead of the investigational device system. As I said, this is just the description of the patients that had a tricuspid isthmus lesion. The other lesions were attempted with the REVELATION_ Tx.

So the protocol was that the REVELATION_ Tx was to be used for all lesions and the NavAblator^TM was to be used if the REVELATION_ Tx was unsuccessful to the tricuspid isthmus lesion. The NavAblator^TM was used in 59 out of the 116 patients. In 57 out of 59 of those times, it was the first catheter to be used at the isthmus. In 10 of those patients, a non-investigational catheter was also required in addition to first attempting the lesion with the NavAblator^TM.

In measurement of the successful biodirectional conduction block at the tricuspid isthmus using the NavAblator^TM catheter, there were 42 out of 59 patients who were successful with the first use of the NavAblator^TM. In comparison, the non-investigational catheters used alone at the tricuspid isthmus were successful 100 percent of the time achieved bidirectional conduction block.

And as has been previously discussed, the sponsor states that there is not adequate information available to measure acute procedural success.

In addition to this being an endpoint for the study, the other issue that has been raised is when does the electrophysiologist know that the lesion has been successfully completed. We do not have information to tell the FDA upon what decision point the individual investigator used to stop their lesion. So some investigators may have used a different criteria than others and the procedure for the use of the investigational catheter might be slightly different per procedure.

In the transtelephonic monitoring in the sixth month post procedure, information submitted by the sponsor included detailed information on 83 patients. Out of these 83 patients, there are 22 patients who had no transmissions. We do know that at least two of these patients either lost the monitor or had a non-functioning monitor. There are 31 additional patients that transmitted between one and three times. So 53 out of the patients upon which we have information had poor compliance with monitoring in the six month post procedure.

To chart the number of patients that actually did achieve the primary effectiveness endpoint, there were 88 patients in the effectiveness cohort. As previously stated, 70 of those patients had the lesions performed only with Cardima catheters primarily.

There were 18 patients that had the Cardima catheter plus a non-investigational catheter used. In other words, the non-investigational catheter was used primarily for the isthmus lesion. The lesion were only done with Cardima catheters in 70 patients and 58 of those ?- Actually this slide after going over these slides so many times. The lesions done with primarily Cardima catheters and non-investigational catheters secondarily. So there were 58 patients who had only Cardima catheters, 12 patients that used both but only after using the Cardima catheter first.

Out of the patients who only used Cardima catheters, there were 24 who reached primary effectiveness, 34 who did not. Out of the patients who had Cardima catheters first and did non-investigational catheters secondarily, there were seven who reached the primary effectiveness and five who did not. Then of the patients who had the non-investigational catheter used primarily for the tricuspid lesion, there were 11 who reached primary effectiveness and seven who did not.

To look at this in a slightly different way, there were 42 out of the whole 88 patients who were in the effectiveness cohort who reached the primary effectiveness endpoint and that is the reduction and symptomatic episodes that meets the prespecified endpoint during the sixth month post ablation compared to baseline while all on the same medications or reduced dosage.

Out of those 42 patients that meet the primary effectiveness endpoint, 24 reached the primary effectiveness endpoint using only the investigational device system. Out of that 42, there were 18 patients who required a non-investigational catheter to complete the ablation procedure. The breakdown of that 18 patients is seven used the non-investigational catheter secondarily after using a Cardima catheter and 11 who used the non-investigational device primarily.

Of the patients that were deemed to have failed the primary endpoint, eight had AV node ablation prior to the sixth month. Five had a pacemaker prior to the sixth month which is a prespecified failure according to the protocol. One withdrew due to failure to improve. Nine did not have sufficient episode reduction. Twenty-one did have episode reduction but with a new antiarrhythmic drug or increased dose. This also includes two patients who also had a pacemaker implantation. Two patients had both an increase in antiarrhythmic drug and an insufficient reduction in episodes.

There were 43 out of the 88 patient effectiveness cohort that were reported no events or transmitted no events in the sixth month post ablation. Two of this 43 have had an AV node ablation, one prior to the sixth month and one after the sixth month assessment. One out of the 43 had a surgical MAZE procedure after the sixth month assessment time period. One out of this group had an atrial defibrillator after the sixth month assessment period. Seven of these patients had amiodarone added prior to the sixth month assessment.

In terms of antiarrhythmic drugs that the patients were taking in the sixth month, there were data presented on 82 patients. This data is included as I had mentioned before in the appendix to the clinical review. There were 14 patients that were on no antiarrhythmic drugs or no Class I or Class III. Six of these patients also were on the same medication regime that they were at baseline. So eight patients had changed.

There were 26 patients that had an increase in their antiarrhythmic drug dosage or an adding of an antiarrhythmic drug. Amiodarone was added in six and removed in 10.

There were 10 patients out of the effectiveness cohort of 88 that had an AV node ablation procedure the linear percutaneous ablation procedure. Two patients went on to have a surgical MAZE after the linear percutaneous ablation procedure.

The secondary effectiveness endpoint for the study was improvement in quality of life as measured by the Short Form-36 and the Atrium Fibrillation Severity Scale compared to baseline.

When the patients were compared to their baseline and in the sixth month assessment period, there were greater than 50 percent that had a clinically significant improvement only in the vitality domain of the SF-36 but in two domains of the AFSS. This information does include all of the patients that had disputed baseline numbers and it includes some patients who had an AV node ablation.

The patients who had major complications within seven days of the ablation procedure includes patients that have been described already by the sponsor. But the Agency has included one additional patient as the sponsor had mentioned. This patient is the fourth patient down that had sinus node dysfunction and required pacing within three days of the procedure.

This patient is included in the Agency's assessment of adverse events because the patient required emergent or semi-emergent transcutaneous or transvenous pacing. The documentation says both. Because of the urgency of the treatment required and also it was not a planned pacemaker even if the patient had sinus node dysfunction, it was somewhat unexpected to the investigators that the patient needed to have a pacemaker implanted.

According to the strict definition of major complications as has been presented by the sponsor, the two other patients that had pacemaker implantation in the week post ablation procedure really would fit into the definition. This is an issue that we are hoping that the panel will discuss because it is in this study without a control group. It can be difficult in this patient population to know if these are truly adverse events or not.

So five out of 116 patients or 4.3 percent with a confidence interval of 1.7 and 9.4 percent of the patients had one or more major complications. Also a possible safety concern is the 20 patients that a requirement for a pacemaker implantation in the follow-up period of the study which ranged from one day to 1.5 years after ablation. There were seven patients that had pacemaker implanted within six months of the ablation procedure without an AV node ablation.

Dr. Li will present the statistical summary and then I will come back to the podium and present the summary statements of the presentation.

DR. LI: As Dr. Ewing has mentioned, the study is a single arm, non-randomized, multi-center study with subjects serving as their own controls which means that the treatment effect is primarily evaluated by comparing the post treatment measurements to its corresponding pretreatment measurements.

Unlike many of the investigational protocols for IDEs, this particular one didn't come with an explicitly stated rule that specifies what the study results have to be in order for the investigational device system to be approved.

But data still are looking for interpretation in relation to safety and effectiveness. So I will make some remarks that may be relevant to such interpretation. My comments will consist of some general remarks about this kind of study which I will refer to as single arm, pre/post study in general and how they might apply to this particular study. It will include an issue that is very special to this study under discussion.

Every design comes with a standard list of pitfalls that are routinely contemplated. This study is no exception. By pitfalls I mean the existence of factors that are not separable from the treatment effect. That is to say when study outcomes may be considered to be explained by the treatment effect not being zero, it can also be explained by those factors.

For this particular kind of study mainly single arm, pre/post study, there are some general pitfalls. I will go over those items one by one. History means that some events may happen between pretreatment measurement and post-treatment measurement, events such as the approval of a new drug.

Second there may be changes within subjects themselves. One thing that inevitably changes is the age so that's probably how the name "maturation" is given to this kind of pitfall.

Third is placebo effect which is the effect of the perception of treatment instead of the treatment itself. Those endpoints that are subjective self-evaluations may be considered to be particular vulnerable to placebo effect but few endpoints are known to be immune to the influence of such an artifact.

Fourth is selection. The way that patients are selected along may influence the distribution of the post-treatment measurement in relation to its corresponding pretreatment measurement.

Here are some of these pitfalls that may apply to this particular study. Placebo effect. The secondary endpoint is clearly a subjective self-evaluation. And to some extent, the primary endpoint may also be considered as subjective to the extent that it's self administered.

Maturation. The subjects may experience some changes in the interim between the pretreatment measurement and the post-treatment measurement in terms of motivation to report events and as has been mentioned there is possibility that a single episode may mistakenly be recorded as a multiple episode. So the ability to avoid such mistakes may also change as the subjects are getting familiar with this procedure. A so-called learning curve.

Selection. As we know, patients are selected according to their AF frequency during a 30 day screening period. As we know, some patients may typically have three or more episodes per a 30 day period. Some people may just be caught in a bad month and for the second kind of patients, their frequency of episodes may go down post-treatment. Some may refer to this as a "regression to the mean" effect.

The above issues are general to the kind of study which is called single arm, pre/post. Here is a very special issue to the current study under discussion. As we know, the study procedures were performed using non-investigational devices on a sizable number of study subjects. Those subjects are not randomly selected which means that the patients who are treated only with investigational devices do not form a random subgroup of all the study subjects.

How this statement is supportive of the statistical aspect of the conclusion which Dr. Ewing will present shortly is logically fairly straight forward. So instead of belaboring the point in strictly statistical terms, I'm going to use a metaphor.

Imagine a picture being cut into two pieces and one of the pieces is lost. You are presented with the remaining picture. The question is do we know or do we have any information of what the original whole picture would look like when the lost piece was found. You would immediately ask what kind of picture and how were the pieces cut. In general, the question has no answer.

Suppose someone comes and presents you with a piece from another picture and the question is would this additional piece help you figure out what the original picture looked like. The answer in general would be perhaps "No, it wouldn't be that helpful."

In this metaphor, the missing piece refers to the outcomes that would have observed for those patients treated with non-investigational devices had they been treated with investigational devices only. The remaining piece of this picture refers to the data on those patients who were actually treated with the investigational devices only. The piece from another picture refers to those patients who were actually treated with non-investigational devices. With that, I'll turn the podium back to Dr. Ewing.

DR. EWING: Some summation statements that the FDA would like to leave you with were that:

1. The ablation procedure, the lesion made and catheters used, was not the same for all the patients in the study. (a) There were a variety of investigational and non-investigational catheters used for one of the lesions. (b) A small number, 38 out of 116 or 33 percent, of the patients had the ablation procedure performed using the investigational device system as specified by the protocol. That is all of the lesions to be performed first with REVELATION_ Tx catheter. (c) Some patients did not have the lesion set performed as specified by the protocol.

2. As a result of the unblinded study design, there could be a bias toward reporting AF episodes at baseline and against reporting in the sixth month.

3. Acute procedural success cannot be assessed for the procedures in the study due to incomplete reporting of the various acute procedural endpoints.

4. Some patients had further procedures to treat paroxysmal atrial fibrillation after the linear ablation procedure.

5. Twenty patients had a pacemaker implanted one day to 1.5 years post procedure.

6. There was poor compliance with "compulsory" transtelephonic monitoring during the sixth month.

7. More than 50 percent of the patients had clinically significant improvement in two AFSS domains. The quality of life results include data from patients with the ambiguous number of baseline episodes and some patients who had an AV node ablation procedure.

So in conclusion from the clinical and statistical perspective, it is not clear if data can support any conclusion about the safety and effectiveness of the investigational device system.

CHAIRMAN LASKEY: Thank you, FDA. We have some time again from the panel to query the FDA on some elements of their presentation.

DR. NORMAND: Can you see me? I don't know if you can see my hand going up. It's really important for me to understand some of the numbers in the very first table because some mention has been made of that fact that a sizable number of the participants had non-investigational devices. I just look at the numbers and I keep getting different numbers. Can you just go over it one more time, the table where you're saying that X number of patients had non-investigational devices just to help me understand it a little bit better.

DR. EWING: Okay. This table?

DR. NORMAND: Yes, that table.

DR. EWING: There is a difference in some of the numbers. I based these numbers on the information that was presented in Table A-19 which is included in the appendix. This table is my counting of the numbers.

DR. NORMAND: So if you can just tell me what your table means. Is it the "other only" that are the non-investigational?

DR. EWING: The "other" is the non-investigational.

DR. NORMAND: So 20 out of 108 are not non-investigational in the FDA's opinion.

DR. EWING: Correct. It's definitely non-protocol devices. Then there were also non-investigational devices used after the Cardima catheter was tried and we assume it failed to produce the lesion that the investigator thought. So it's a combination of non-investigational catheters which were used first or after the Cardima catheter failed.

DR. NORMAND: That would be the 20 plus the eight plus the 10?

DR. EWING: Correct.

DR. NORMAND: In other words, anything with "other" in it. I'm being simplistic.

DR. EWING: Correct.

DR. NORMAND: Thank you.

DR. EWING: Should I sit?

DR. GILLIAM: No, I have one question. In your review, was the "other" catheters used to make a line any other place other than only the isthmus line?

DR. EWING: No, only the tricuspid isthmus line.

DR. GILLIAM: Okay.

DR. TRACY: Warren, can I ask her a question? This is a huge difference of effectiveness from 85 to 47 percent. Did you look at including the "other" when there was another catheter used? What would your effectiveness have been if you were calculating?

DR. EWING: That is actually the total number of patients that reached the primary effectiveness point which is 42 out of 88. That included patients who had only the Cardima catheter used and who had the Cardima plus a non-investigational catheter used.

DR. TRACY: Can you explain again why are your numbers so different? I'm lost.

DR. EWING: From Cardima?

DR. TRACY: Yes.

DR. EWING: We included the patients that went on to have another procedure such as AV node ablation if they perceived the linear ablation procedure failed. So the patients that the sponsor called "withdrawn", we included in the effectiveness endpoint.

That's part of the answer but the other answer is that they included patients that had a pacemaker implanted as successes. According to the protocol, patients who had a pacemaker implanted before the six months were to be called failures. We called them failures. There are seven patients included in the increased antiarrhythmic drug group that were different from the sponsor's mainly because they added or increased amiodarone.

But you are right. Our numbers are different than the sponsor's and as I said before that`s why I included the data upon which I made these decisions in the panel pack.

CHAIRMAN LASKEY: But in no instance were these issues prespecified? Your definition? Their definition? No drug manipulation for example.

DR. EWING: There was not a prespecified protocol for antiarrhythmic drug use.

CHAIRMAN LASKEY: Which in term ties into success by your definition.

DR. EWING: Correct.

CHAIRMAN LASKEY: So that was not hashed out in the deliberations leading up to the conduct of the protocol.

DR. EWING: Actually I don't know what was discussed but it was not included in the protocol. Just because it is not in the protocol does not mean it was not discussed with the Agency.

CHAIRMAN LASKEY: It's just helpful to us.

DR. EWING: The honest answer is I don't know that it was discussed.

CHAIRMAN LASKEY: We have some of the history but obviously we don't have all of the history. Any other questions from the panel?

DR. SCHWARTZMAN: Being new to this side of the table, I don't know if this is cricket because beyond what's being asked of the company, but I wonder if some of the clinicians might comment on whatever one year experience they have. Certainly from a clinical point of view, six months is almost no time. One year is very short.

What we really want is multi-year data particularly given the fact that these are relatively young patients. It is obviously in the context of trying to get approval it's not possible. But I imagine given the timing of this study there was substantial one year experience. I wonder if any of the clinicians might comment on whether these patients have maintained a stable program which is hybrid therapy, ablation plus whatever drug they are on, early post- or pre- at one year.

DR. KOCHERIL: Abe Kocheril. Dave, I can give you a couple of different answers to that question. My overall experience with this type of catheter system goes back to the REVELATION_ catheter which preceded the Tx. I had published a study using that off label to deliver linear lesions for paroxysmal atrial fibrillation. This was published in the journal of "Interventional Cardiac Electrophysiology."

That reported on 29 patients who had the procedure done and the follow-up time at the end of data collection was 19.7 months. So it's almost two years. There was significant reduction in AF episodes. Actually this was more of a rigorous look. Seventy-nine percent were free of AF and off medications at the mean follow-up of 19.7 months.

I'm not sure that we have enough data to answer it from the Phase III clinical or even combining Phase IIb and III. There are patients who have completed the one year time point and a few who have completed the two year time point. But I don't think they have enough numbers to say that for sure.

My experience consists of the 29 that were published. I have probably a total experience of 50 using the REVELATION_ catheter. Then I've enrolled 20 some odd patients in Phase III. The overall experience has been positive but that's a very good point. That's a short window of observation.

DR. KAY: There was one thing that I wanted to help clarify. That is this numbers problem that Cindy brought up as far as why the numbers are so different. I think there are two things. One is the way Cardima looked at it which is you started with 81 patients and how many met the bar as far as reduction in symptomatic atrial fibrillation episodes at 50 or 75 percent.

The primary endpoint in the study which is specified in the protocol is reduction in symptomatic atrial fibrillation episodes. A secondary endpoint is clinical efficacy which has to do with the drug effects. So the primary endpoint that we spoke about was 85 percent of patients had a 50 or 75 percent reduction in symptomatic atrial fibrillation episodes not looking at the antiarrhythmic drug motion during the study.

Once you factor that in at least as identified by the investigators that's where our numbers drop down to a 62 percent efficacy. Then as we look at the drugs, it's very hard to figure this motion of drugs up and down. But when we look at the picture of the big circles, we saw a motion down and not a motion up. That's the main point that we wanted to make.

CHAIRMAN LASKEY: Yes, we're actually trying to confine our comments to the Agency but thank you for the elaboration. Again just to remind all of us, we really need to confine our thinking to the material at hand. While the other long term stuff is of interest, we can only judge what's here in the panel pack today. If not, I suggest that we break for lunch and resume in one hour at 1:30 p.m. Thank you. Off the record.

(Whereupon, at 12:39 p.m., the above-entitled matter recessed to reconvene at 1:42 p.m. the same day.)

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

1:42 p.m.

CHAIRMAN LASKEY: On the record. I'd like to call us back to order. In order to stay on schedule or perhaps even be ahead of schedule, let's move on with the open committee discussion. I would like to have Dr. Gilliam present his review please. I'm going to have my timer on here, Rosie.

DR. GILLIAM: Well, then I guess I'll have to move quickly.

CHAIRMAN LASKEY: Correct.

DR. GILLIAM: I have a few questions regarding the investigators first. The time of ablation of 250 minutes seems long. I know I saw one that was 755 minutes. God help us that day. Is there a comment on that? This seems a pretty straight forward procedure. What are the complexities that make this literally a four hour procedure?

DR. CALKINS: It is interesting looking at those numbers how long it ends up taking. I think there's a couple of things. One is it's the old question of how long does ablation really take from when they get on the table to when they leave the table as opposed to in case of the academic institution when the actual attending shows up to be there during the burn. The things that make this procedure less than a 30 minute procedure are (1) the flutter line. Early in the protocol there was a requirement that you had to try this REVELATION_ catheter for ?- You know isthmuses really are not designed to be ablated with a linear catheter so that added a considerable amount of time early on to the procedure. Then (2) it was using the 4mm catheter to get complete isthmus blocks.

So there is a certain amount of time goes for a flutter ablation and then you have two more linear lesions. When you do ablate along the lateral wall, we'll pace to make sure we don't have any phrenic nerve stimulations and so forth. Then at very points in the protocol, you had this requirement to record every electrogram and print things out and measure things so all of that ended up slowing things down.

If you say how long does it take to do just to go in without having to do some of these steps and go straight and use the NavAblator^TM instead of having to use the REVELATION_ first, I think it's a one to two hour ablation time procedure and not a three to four. I suspect the 755 minute case was when some of these cases people were looking out of interest with the ESI system. Was there block and was there not block? As soon as you start throwing in complex mapping systems, you can add an awful lot of time to our procedure that's not the heart of the essence of the procedure as I see it.

DR. GILLIAM: While you're there, another question concerns me. This isn't your chart but it's this chart that's on 79 that shows the episode frequency. Do they mean episode frequency starts out at baseline at 9.2? At three months afterwards, it's 3.5. At six months after, it's 1.2. Obviously it didn't do anything between three months and six months. How do you account for that?

DR. CALKINS: It was interesting looking at that slide because it was strikingly how the number of symptomatic atrial fibrillation episodes seem to decrease over time. Looking at the data the only way I can explain it are two things. I believe the data. 1) You have the whole issue of reverse remodeling that sinus rhythm begets sinus rhythm. 2) I think we're all aware of the fact that you do an atrial fibrillation ablation whether it's a pulmonary vein ablation or any ablation and you're getting transmural lesions, you're creating some information that indicating a healthy phase.

Certainly in these pulmonary vein isolations, we see it in a very dramatic way that lasts for one to three months. It can last up to a fair amount of time, up to three months for this irritability phase. So if you say there's an antiarrhythmic effect of the procedure that happens immediately, you balance that with some healing phase inflammation. That gives you a little more atrial fibrillation early on and then this remodeling to my mind is how I can explain it over time is atrial fibrillation keeps getting better and better. Those are the two things that come to mind to explain this continued improvement in the symptomatic atrial fibrillation episodes.

DR. GILLIAM: My concern again was whether we just got the episodes. Whether people either became frustrated or they had a placebo effect to the event and they just either stopped calling in or gave up. That's the one concern I have because I don't know early on given that the variability of atrial fibrillation from 12 percent to some people are 100 percent as far as knowing whether they are in atrial fibrillation when they send in their transmissions. Did you all consider the potential of doing weekly transmissions before the procedure during that month of time to document what they were doing and then follow up afterwards?

DR. CALKINS: Yes, I think how is the best way to document these atrial fibrillations and quantify it. I think we all struggle with it every day. Other than David's suggestion which was a continuous Holter monitor that you have every single day and trade in, it's basically impossible. I guess the REVEAL is one option. All these things were considered.

This study was designed five years ago. But even today, you say that event monitors seem to be the best we can do. You say that maybe we should do Holter monitors periodically. In this study, patients did transmit weekly pre- and post- to get trained in this routine of Tuesday is your day for transmitting.

All I can say is I think this is the best approach we have. By having it in an outside company and the outside company calling the patient, I think every effort was made to make them comply. This is just a reality. But then when the patient showed up, they felt better and they were in sinus rhythm on their EKG.

DR. GILLIAM: And regarding that, there are five pacemaker patients I think prior to the procedure. Were any of these patients evaluated using their pacemaker counters to determine whether they had a decrease in their atrial fibrillation episodes using the counters before and afterwards?

DR. CALKINS: I'm not aware of that data. I don't know if those were Abe's patients but that's certainly another way to monitor this. But I don't think we have that data. It was not collected as part of the study.

DR. GILLIAM: That would be obviously a pretty easy way to see mode switching episodes or even actual documentation of atrial fibrillation in some way. There is a question raised and I'm not sure that you have the answer. Potentially some of these patients could have had the same episode of atrial fibrillation but lasted a longer period of time. Do you have any take on that?

Could someone have been in atrial fibrillation for a long period of time to actually get their three episodes? Was there a requirement for instance of normal sinus rhythm documentation between episodes of atrial fibrillation in any way?

DR. CALKINS: No, I think your point is well taken. It's amazing now as we look at all this to say that here's another ten things that retrospectively maybe we should have done. You could say that it's clear what happened. The patient one episode of atrial fibrillation that lasted an hour that transmitted 20 times in the hour and that's your 20 episodes. That's why we are seeing such a good benefit.

We went back and looked at the data to address the issue saying if you defined this episode as being separated by at least an hour, the actual time that it was recorded, what does that do to the numbers and does it effect the affect at all. If you say that throughout we only at most will allow one episode per day and separate by days, then you'll only loss two patients.

Our interpretation is no, that's not what we were seeing. These were patients with paroxysmal atrial fibrillation that were in general aware when they were in or out. But I think that was a valid concern. When we looked at it I was relieved to say it wasn't the patients that were cheating the books that they were transmitting 20 times in an hour and one episode of atrial fibrillation. These patients all have multiple atrial fibrillation in multiple days pre- and no atrial fibrillation documented post or little atrial fibrillation.

DR. GILLIAM: The selection of patients. There are 20 centers. Obviously you have a lot more atrial fibrillation patients than 80 you would think that you would have over the period of time. Was there any selection that you spoke with the investigators to see how they selected who would enter the trial other than just doing the monitoring?

DR. CALKINS: I think there's a lot of different centers that were involved in the study. The question is how did a given center find these patients. The reality is you sign up to a clinical study and you start thinking you're going to have a lot of patients and you approach patients with this incredibly cumbersome thing. They have to have this event monitor. They have to transmit for six months. They're going to get called at home if they don't transmit. They say "No thanks. Let's just use the off-label non-thermocouple version of the catheter or the heck with it. Let's skip it." The patients then disappear.

The other thing we saw during this study in some centers when the study started a pulmonary vein atrial fibrillation didn't exist. As the study was going along, that showed up. For centers that wanted to be doing the ultimate novel unknown thing, that was something that people started getting involved with. So some centers initially on put in five or ten patients and then they say "The latest thing is atrial fibrillation of pulmonary veins. I want to be doing that." So the patients would be steered to that so they could build up their experience with that procedure.

What we've learned obviously as we look back, four years ago we were ablating deep into the pulmonary veins creating all this pulmonary vein stenosis thinking we were doing good. At the end of the day, a lot of people were stepping back and saying "We have to be out of the pulmonary veins for that." This study has just been chugging along coming to completion and showing very sound data and very excellent safety.

DR. GILLIAM: Looking at the catheter itself, I'm not sure that you necessarily need to answer this but I think you might be able to. It looks as if it's really difficult to manipulate the tip of the catheter. It needs a delivery system to get it where you put it. Why is this catheter more useful than a regular ablation catheter in doing essentially point by point?

DR. CALKINS: There are a couple of unique features of the catheter. It's incredibly flexible and floppy as you can see. That allows it to conform to the heart, move with the beating heart and so forth. It's delivered with the Naviport_ catheter so it ends up being a deflectable catheter with this floppy curve. Sometimes you can get all eight electrodes in touch with the wall and sometimes only four. Because it's floppy, it's closely adherent.

Then the question about the lesion characteristics when we worked with the dogs, these were nice and linear. When you just try to do a point by point pullback, you get these skips and gaps. Also you're debulking the atrium. You're creating these 5mm marbles every time you burn with a standard RF catheter. With this, they are 2mm to 3mm thick and they are equally deep. So this is more doing what we are trying to do which is not ablate the atrium but put roadblocks up to ablate atrial fibrillation.

DR. GILLIAM: Given that, you would think that for a flutter line this is almost be a preference rather than the point to point for almost the same reasons. The fact that it would conform will allow it to deal with the ridges and the aggravations that we typically run into doing a flutter line.

DR. CALKINS: The issue there is the flutter. I saw Anton Becker (PH) had a presentation at the NASPII (PH) meeting which was just terrific showing the anatomy of the right atrium and the pecta eight muscles that fan out in the flora of the flutter isthmus. There's been many companies that have tried to ablate the isthmus with a linear catheter and all have failed.

There hasn't been a single linear catheter system ever tried in humans. In dogs, the dog isthmus is nice and smooth. It always works. In humans it's these huge pecta eight muscles and you need to drop down into the valleys with a end tip firing catheter that's the critical thing.

In this study early on, people worked with that catheter. We learned you have these pecta eight muscles. It's the wrong catheter for that job. You need an end hole catheter. That's when the NavAblator^TM came in and then the NavAblator^TM is interchangeable with an 4mm tip end hole catheter. It's just whatever you like in terms of the handling characteristics.

DR. GILLIAM: I have one more question and then I'll turn it over. I may come back and ask some more. This question may not be playing fair but I do need to ask it. When you do your pulmonary vein ablations now, do you routinely do right-sided lines as these lines would dictate?

DR. CALKINS: I just isolate the four pulmonary veins and do a left-sided line, left flutter, left inferior vein.

DR. GILLIAM: Do you do any right-sided ablations in addition?

DR. CALKINS: For redos, we'll bring them back and we'll isolate the superior vena cava and do a flutter line and so forth. But we won't for a standard pulmonary vein case. I think very few people are. Piroget (PH) and the rest of them aren't working on the right side routinely for the pulmonary vein isolation procedures.

DR. GILLIAM: Thanks.

CHAIRMAN LASKEY: Great, Rosie. Thank you. Bill.

DR. MAISEL: Good afternoon. First I would like to congratulate both the sponsor and the FDA on excellent presentations this morning. I think you really highlighted and clarified many of the issues that I had certainly. What I would like to do is focus on several areas that I think warrant a little further discussion and perhaps a little clarification from the sponsor and the FDA on these issues.

DR. WALDO: Could you speak a little louder?

DR. MAISEL: The specific issues that I have include the procedural methodology including the creation of the lesion sets, the procedural endpoints and the use of multiple catheters in individual patients, safety issues with particular attention to the post-procedure pacemaker implantation rate and then the assessment of the pre- and post-procedural atrial fibrillation. We have touched on many of these issues already.

With regard to the standard lesion set, the sponsor states that the protocol did not define a standard lesion set and that the optimal lesion set is not identified by the results of this study nor was it an objective of the protocol. That being said, 83 percent of patients received the posteroseptal, posterolateral and the isthmus lines.

That number increases to close to 90 percent if you include the patients who had a prior isthmus ablation. So in more than 90 percent of the procedures, at least three of the proposed four potential right atrial lines were made. For me, while there is variation lesion set that the patients received, it seems reasonable to group the patients and consider the outcomes of the procedure as a whole.

More challenging to me is the interpretation of that procedural endpoints. As we've discussed already, the procedural success was defined in the protocol as demonstration of at least one of the following: reduction in the amplitude, fragmentation or widening of the local electrogram, appearance as split potentials or an increase in the pacing threshold.

As noted in FDA reviews, these procedural endpoints were not consistently measured or recorded on the data forms. The sponsor indicates that no conclusion can be made regarding the acute procedural success endpoint. I perceive the lack of a clear procedural endpoint as a problem. I think it makes it difficult to instruct physicians performing the procedure on how to assess to the acute success.

So I would be interested in knowing from the sponsor specifically what are your instructions to the physicians performing the procedure as to how to perform the lesions and when is the procedure done.

DR. KAY: Neal Kay. I'll try to give the best I can. I haven't use this catheter but I've used others to try to do linear ablations in the right atrium. It's extraordinary challenging. I think that part of these procedure times that Dr. Gilliam was asking about really are related to trying to figure out did you get block. I think looking for double potentials is great if you know the direction of propagation. If it's going parallel with your line, you may not see any double potentials even though there is block. So it is a challenge.

I think that right now the state of the art at present is to decrease the amplitude of the electrogram and then to try to some kind of mapping method to pace on one side of the line and show block. Just looking back the way this procedure was done, a lot of these techniques we use now with electroanatomic mapping just were not available when this procedure came up.

Let me just also address some of the confusion. I think Dr. Ewing's presentation was outstanding. There is a question of who do you group and how do you include in this based on what the flutter line was ablated with. It's pretty clear that the linear ablation was all done with this REVELATION_ catheter.

DR. MAISEL: If I could just interrupt. I'm trying to clarify for a physician performing the procedure what precisely are the instructions to that procedure about what they should do with the catheter. How long should they be applying RF? Are there temperature guidelines that you recommend? Is there a way for them to know when they create a lesion that they don't need to reapply RF at that site? Can you just spell those explicitly out so that we can understand how the procedure will be performed?

DR. KAY: Let me ask Abe. He actually did the protocol so he can tell you how long they applied RF for?

DR. KOCHERIL: Abe Kochenil. I guess before I go into the specific details. The air has changed so if the study was being redone today these electroanatomic mapping techniques would help in making sure that the lines were delivered.

What was done in the protocol was to deliver lesions. There was a thermocouple so we could measure temperature. We would want at least a 10 degree rise over body temperature to make sure that we were actually doing something to the tissue. The typical application was a minute at a time so each electrode gets a minute worth of burn. During the burn, all of us are trained to look at those electrograms and even with all the caveats we are looking for those electrograms to shrink and there is an internal assessment of what's going on.

This is fairly easily to transmit to an electrophysiologist who wants to pick up the procedure and that would be the current instruction. It's just to use what we already know to be markers of getting an adequate lesion. But the guidelines are to get that temperature and keep it up for a minute.

DR. MAISEL: Do we have any idea on how many of the patients or lesion sets met those criteria? I realize we didn't measure threshold and all those things. I'm trying to understand what those are based on.

DR. KOCHERIL: This is definitely soft and that's what you're reacting to. But the best we have is the investigator assessments and according to that assessment, 110 out of 118 were successful in accomplishing that.

DR. MAISEL: Okay. Thank you. The next issue is the use of multiple catheters in individual patients. As I already stated with the exception of the isthmus lesions, all the right atrial linear lesions were created with the REVELATION_ Tx catheter. A variety of catheters were utilized to create the triscuspid isthmus lesions. The numbers vary in different places in the application but the general message is the same.

The REVELATION_ Tx catheter was used in 31 patients and of those 10 or almost one-third required a second catheter for the isthmus. The NavAblator^TM was used in 59 patients and in 10 of these or one in six an off-protocol catheter was required to achieve bidirectional isthmus block.

So in 70 of 108 patients bidirectional block was achieved with only the protocol catheters although admittedly the NavAblator^TM was not used in some patients because it was not part of the initial protocol. An overall 84 percent of the patients achieved bidirectional block. The success rate of the non-investigational catheter when used alone was 100 percent and was 81 percent when used after another catheter. To me the rate of isthmus block seems a little lower than I'd expected. I would be interested in hearing your thoughts about why that might be.

DR. CALKINS: I appreciate the question. Hugh Calkins. I think it really comes down to how hard you try with a given catheter before you switch to a new tool. If you used a tool for three years and you are used to using whatever manufacturer's catheter that's what you're really comfortable and that's what you want when things aren't going real quick and the day is getting long. That's when you want to go to what you're most familiar with. That's what we faced.

I firmly believe and there's been animal studies to show that and the company has data showing that a lesion generated with the NavAblator^TM is exactly the same as the lesion generated with 4mm catheter made by XYZ Company and there's absolutely no difference. I think you and I would both believe that as an electrophysiologist understanding the principles of RF energy.

It comes down to handling of the catheter and what you're comfortable with. What we saw was that investigators would ?- The protocol said you had to use the NavAblator^TM. It's not the catheter you really like. You humor this protocol for so long, given it 10 or 15 minutes. If you don't get block, you're going to switch to the catheter you have been using longer and just feel more comfortable with. I think that's what we saw.

If you were to do a study where you put me in a room with the two catheter and say get isthmus block. I think no one gets 100 percent with a 4mm catheter. Some patients have very thick isthmuses. You need irrigated catheters or 8mm catheters or something else. So the 100 percent I don't believe either for the non. But all of us will get isthmus block in about 90 percent with any 4mm catheter and 50 watts of RF energy and that's the reality of what we are seeing. It's how long do you try before you switch.

DR. MAISEL: Thank you. I think the use of multiple catheters including some off-protocol catheters makes it a little challenging to interpret the data but you do provide a breakdown of the individual catheters and success rates. It's in the clinical summary which I think helps allow for interpretation of isthmus block.

I would like to move on to some of the safety issues with particular attention to the post-procedure pacemaker implantation rate. Five of 116 patients per the FDA had major complications within a week of the procedure. An additional three patients required pacemakers within two weeks and overall 20 or approximately 23 percent of patients had pacemakers implanted during the course of this study, nine of these following AV node ablation.

I think the issue of post-procedure pacemaker is an important one for several reasons. For one, there is certainly data that for patients with sick sinus syndrome pacing the atrium may reduce the incidence of atrial fibrillation. We all acknowledge that these patients are at higher risk for receiving pacemakers. They are on multiple drugs. They have sick sinus syndrome.

The questions for us are is the procedure associated with a significant risk of post-procedure pacemaker and if it is, is that increased risk justified by the important clinical benefit. Some patients and physicians may be willing to accept a higher risk of pacemaker if the procedure is successful. Do we have any data maybe historical data regarding atrial fibrillation patients who are refractory to medications on multiple drugs and what the expected pacemaker implantation rate might be in those patients? It just strikes me as a high number, 23 percent of patients receiving the pacemakers.

DR. CALKINS: Again Hugh Calkins. I appreciate the question. A couple of things to bring out. One is that the 23 patients we're looking at data from six months. So from six months follow-up, 13 patients have pacemakers. So the 20 is adding in pacemakers that had occurred in the second six months which isn't the data that we are focusing on today. So 13 patients had pacemakers.

I'm not aware of any comparable data taking patients with paroxysmal atrial fibrillation refractory to two or more drugs or to amiodarone and seen over six months of follow-up how many will end with pacemakers. I suspect it's a fair number.

Certainly while this study was going on, this is the same period where some of the studies were coming out, the St. Jude, DAO pacing algorithm for atrial fibrillation. There was a lot of hype during this period about devices for atrial fibrillation. I think now once the dust has settled, most electrophysiologist believe there's very little role of pacing for treatment of atrial fibrillation.

Certainly the magnitude of improvement we saw in this study has not been seen in any pacemaker study that I've been aware of for atrial fibrillation. When I look at the population, look at the current understanding of the data and then look at the procedure, except for that initial anterior line that was dropped from the second phase of the protocol, we're ablating the septum which is far away from the sinus node, the flutter line which is way far away from the sinus node. Even though the posterolateral line is not an anterior line along the crista where the sinus node would be but it's posterolateral. So there is no lines in the region of the sinus node.

I would have a hard time understanding from a pathophysiologic basis. When we go to ablate the sinus node and some of the inappropriate sinus tachycardia, it's virtually impossible even when we deliberately try to do that. For many reasons, I don't think the pacemakers are a complication of the procedure. In most of the patients who had it, there's good documentation of sick sinus syndrome prior to the procedure.

DR. MAISEL: Okay. I saved what I consider the most important issue for my last question which has to do with the assessment of the pre- and post-procedure atrial fibrillation. I just wanted to review what we stated as the primary effectiveness endpoint which was the reduction in symptomatic episodes of atrial fibrillation assessed six months post-procedure compared with the patient's baseline frequency.

For subjects with greater than or equal to five episodes, they were required to have a 50 percent reduction. For subject with three to four episodes, they were required to have a 75 percent reduction. These reductions were to occur in patients taking the same or a reduced dose of medication.

There are several important potential sources of bias in the data collection that I would like to highlight. One has already been mentioned that patients were aware that a certain number of episodes were required to be admitted to the study. I'm not that troubled by that. Patients didn't know exactly how many episodes they needed to get in the study. If anything it would make them more likely to make transmissions. We may get a truer incidence of the pre-procedure atrial fibrillation.

To me much more significant are the post-procedure monitoring issues. During the six month post-procedure, patients were supposed to transmit recording when they were symptomatic and transmit weekly recording whether or not they were symptomatic. That would be a minimum of four recordings for all patients and close to two-thirds of patients did not follow the protocol. They did not make the minimum required number of transmissions. Just over one-third made the required number of transmissions.

I understand there were other ways of trying to follow them up with EKGs and visits, etc. The lack of required follow-up has very important implications for the interpretation of the data. A patient who doesn't follow up in my mind should not be classified as a success but should be classified as missing. My question for the FDA is was any data or analysis performed classifying those patients that didn't meet the minimum of four transmissions classifying them as missing rather than as a success?

DR. EWING: This is Lesley Ewing. That's a very good point. No, we did not do that. If the patient at the six month post-procedure said they didn't have any episodes but they could have had them in the fifth month or the seventh month and they did not have transmissions, we did give that to the sponsor. I think that's a generous assessment.

DR. MAISEL: I would agreed with you. Certainly there's recall by us and patients who show up may not remember two weeks ago that they had a brief spell or what have you. I'm very troubled by the large number of patients who did not have adequate follow-up data. My final comment would be to note that patients that have withdrawn from the study prior to the six month for things like AV node ablation, I agree should be classified as failures and just not evaluated for the effectiveness endpoint. These patients clearly likely had frequent atrial fibrillation and should be failures.

So even if you accept the transmission data as valid, then 42 of 88 patients or 48 percent reached the primary effectiveness endpoint and of those 24 of the 88 or 27 percent have what we initially in the protocol defined as clinical success. So these are the main issues that I had. Perhaps we can discuss them further later.

CHAIRMAN LASKEY: Great. Thank you, Bill. Why don't we go to Al Waldo and have him render his comments. Dr. Waldo.

DR. WALDO: Thank you.

CHAIRMAN LASKEY: Yes, sir. You have about 15 minutes. Okay?

DR. WALDO: I don't think I need that much. I do have some important comments to follow-up on the remarks you've just heard. One of the things that concerns me is what appears to be the lack of rigor in obtaining some of these data. I think for me the most worrisome thing is the lack of the transmission of the transtelephonic monitoring.

The crux of the study is the decrease of the burden of atrial fibrillation. As my old mentor used to say, "I would love to believe you but you have to provide me with some evidence." The trouble is as we've just heard my colleague say before there's precious evidence missing. You just can't assume that because they didn't send anything in that nothing was going on.

I'm also concerned especially with since so many patients had gotten pacemakers symptoms may have disappeared from that. There are a bunch of people that got AV node ablations. There are a bunch of people who got more amiodarone. If you start adding all these little things together, they built the stone statue but on clay feet. That's the problem. That's my biggest concern.

I'm not sure if they would like to comment further on that I would like to hear what they have to say. But I'm very concerned that the critical data that we really need to believe them are not there. That is not sufficient data rigorously obtained to show that indeed with this technique. You heard both Dr. Kay and Dr. Calkins were skeptical at the beginning of this study that this would work. Those of us who may still be skeptical about it need to be shown.

I'm quite willing to be a believer but they have to provide me with enough evidence so I can believe them. I think the missing evidence for the most part for me is the transtelephonic monitoring transmissions. I saved the page. Data submitted on only 83 patients 22 patients with no transtelephonic monitors at all, 41 patients only had from one to three. So therefore almost 64 percent were poor compliance with the transtelephonic monitors over six months. This is to me is the Achilles heel of what's been shown to us today.

DR. CALKINS: It's Hugh Calkins. Just a few comments. The goal of treatment of atrial fibrillation in the patients was improvement of symptoms as reported by patients. So I would agree that when there's no claims that this procedure cures all atrial fibrillation and you can stop Cumidin, this is to make patients feel better.

Again it's very easy now five years later to look back at this as to how the study was designed but the essence of the study was to say six months after the procedure we wanted you to transmit every time you have symptoms. If they don't have any atrial fibrillation, they don't have any symptoms. They don't do any transmissions.

Then all of a sudden, you say there's no transmission so you can't count the data. But that is the data. So the question really comes down to when the patient's feeling well, when they are doing just fine, going about their life and going to work and whatever else, do they take the time to put this thing on to transmit into the company and so forth?

Now I think the strictest look at the data you'd say if there aren't four transmissions or some months there might be five transmissions, then we should say that these patients are non-compliant and ignore all the data even though the patient showed up and sitting there in clinic looking you in the eye and say "I'm feeling great and I've had no atrial fibrillation and I feel tremendously improved. Thank you very much for doing this procedure." And you ignore that the patient that you had at the end of the day.

That's what we are trying to accomplish. We're trying to look at the patient when they come back to follow-up and say "Are you better? Are you satisfied with what I did for you?" They say they are. And you say "Because you didn't take the time out of every week five weeks in a month to transmit the episode data, you might as well not have been in the study and let's exclude any look at the data."

I think one could argue that we want to at least prove that they have these monitors and they know how to work them still and they are still alive. We know they are alive. So let's look at a more reasonable endpoint and say "Did the patient transmit at all that month when they are feeling fine" and call that a compliant patient. To my mind if you look realistically at patients, they're feeling fine at six months and we're learning about this. This will be important for future designs of studies about having things that patients don't have to trigger themselves.

There are some new technologies coming out now that can autorecord and that may be what we need in the future but this technology was not available. To my mind that is a more reasonable way to look at compliance meaning did they transmit once.

If you look at the way you slice the numbers and the way the numbers would get cut either look at the FDA's way of looking at it or our way of looking at it, there are two places where we really lose patients. One is this flutter line. The FDA statistician brought up the issue about the picture and you have part of the picture so how do you know what the picture shows. I think in this case you know what the picture shows because it's a flutter line. So you have a duplicate of the missing corner of that picture. It's a flutter line which is interchangeable regarding what catheter you do it with. The flutter issue at least from a clinical perspective to my mind is inappropriate and it excludes those patients unless you look at the true meaning of what we're trying to do here.

Then with the event monitors, you may say "Only include patients that transmitted once that are at least somewhat compliant." At least, they know how to use it and then they'll transmit symptomatic episodes. So if you add those two things in, then you quickly go from a 44 percent success rate which I think is still very good for a safety procedure in a refractory. Then you add in the flutter line patients and you end up with a success rate somewhere between 40 and 85 percent or 40 and 62 percent or whatever it is, somewhere around a 50 percent number.

I think your points are well taken and we're all learning about study design from this study which I thought was very well designed at the time and very well carried out. To say that it just has clay feet I think ignores the fact that these patients showed up in clinic and said how they felt and they felt better and said they had no atrial fibrillation episodes or a marked reduction which at the end of the day is what we are trying to achieve.

DR. WALDO: I have to tell you. I respect you very much and I respect what you are saying but in the end we have to try and be very objective. I think objectively the data just aren't there. I hope you're right. I wish you're right but I don't know how we know.

To me, it's not just the TTMs. When you add up the number of pacemakers, I think pacemakers alone make the symptoms sometimes. Who know what the symptoms were even if the patients had bardycardia before and that's why we were symptomatic? There are so many reasons. Then a bunch of patients got AV node ablation beside. A bunch more got increased amiodarone which I believe and I think we have to agree may play a role in controlling rate.

So when you add all the other little things to the fact that there's poor compliance and objective evidence with TTMs and what's going on. I was probably the very first paper back in the days when we did the TTMs and patients called in saying they had atrial fibrillation and they didn't and patients called in saying they had sinus rhythm and they didn't. We know about all those things and those data are so well documented now.

It's not just the TTMs. The spin you put on it I understand and I'd love to believe you but I think in the end it's soft. It's not hard data and I'm worried that this is so critical to everything you're trying to say. On top of that, a bunch of patients had increased amio or got amio for the first time. I know some patients went off it. That's good but a bunch of patients got more amio. A bunch of patients got pacemakers and a bunch of patients got AV node ablations. All those are going to impact symptoms.

Controversial or not, just the pacemaker alone some people still claim even a single site patient if the rate's fast enough will impact the incidence of symptoms and atrial fibrillation. So without trying to be contentious in the least way, I'm just trying to be as objective as I can as my old boss said "I'd love to believe you but I need the evidence." That's where I am at the moment. It's very troublesome to me that the critical chunk here that I think is needed to hold your statue up is too soft.

DR. KAY: Al, let me just address a couple of these things. The pacemaker issue was something that I asked the company to show me as well. If the patient has AV node ablation, I think we would all agree that those are failures of this procedure. I don't think we would argue with that. That's a failure. They would be willing to concede that too.

For those patients who got a pacemaker after the six months when the efficacy was obtained, I don't think that should count them as far as the efficacy. Those data came in before they got a pacemaker. If you look at the patients who actually got a pacemaker carefully looking at their baseline sinus rates and I went through everyone one of them, I think there are two patients that I would agree are a complication or that I would think ought to be classified as a complication. The other ones really weren't. Their heart rates weren't the same.

Your point about amiodarone I think, Al, that's not quite right. There were actually 18 patients on amio before and there were 16 afterwards. Your point that a bunch of people were put on it isn't really true. There were more people that came off it than were put on it. I don't think that really relates.

Then the last issue which I think is a troublesome issue is the TTM compliance. Just asking ourselves, how would you do it differently? I think it's the art of possible. I don't know back when this study was designed that there was any better way to do and I'm not sure there really is any better way to do it today to find out. You have an independent person calling them, you have independent people reviewing this and I think transtelephonic monitoring is about as good as we have.

It's really the art of the possible. The company did what they were asked to do and they did it probably the best they knew how to do it. I'm not sure any of us could do any better. Even if you say "We'll cut down and look at these numbers" one thing that's hard evidence is that data you can look at anyway you want. That procedure is actually safe.

Efficacy whether it's 60 percent or it's 40 percent for a right-sided procedure with very low risk is actually probably clinically significant and it's something that I think a lot of physicians would actually use because it's simple to do. Those are the ways I would respond to your very legitimate comments.

DR. WALDO: Again I remind unconvinced. I think that there are a little things. Maybe we can add up the numbers. Was it seven or eight patients being put on amio and how many patients got pacemakers before the six months and so on? They just add up and it just makes data soft. Let me ask you one last question then. On the basis of this study, my three EP colleagues who presented at this meeting for Cardima, are you planning to do these kinds of lesions on a regular basis on your patients who are symptomatic and who you can't cure otherwise?

DR. KOCHERIL: Al, this is Abe Kocheril. I will answer that question. I just wanted to inject that in our trying to figure out how to address the TTM issue one of the last things we did was to look at a multi-variate analysis of what predicted transmissions. This is also going to be on the soft side but what was found was that it was people who were of relatively younger age and people with fewer symptoms who were less likely to transmit. That again won't satisfy the legitimate questions you're asking.

The other question is would we use this catheter and I probably have more experience with this catheter than most people around. What I do in my practice is this is the first ablation procedure that I offer to patients who have drug refractory atrial fibrillation. If this low risk procedure fails, we still have the option of going to the left side and doing a pulmonary vein isolation.

Despite the lack of rigorous data to satisfy your mind, patients do well and it's very little of the time that I actually have to go to pulmonary vein isolation. When I do, I get about a 50 percent success rate there. Yes, we will use the catheter. One of the issues that we are struggling with is how best to use it. This study as Dr. Calkins mentioned was developed awhile ago and things would probably change if we were trying to study it now.

One of the things that I'm looking at at our center is trying to map atrial fibrillation from the standard point of Chaos theory and trying to analyze it from an oscillator model and trying to determine where the ideal lesion location should be rather than just saying we're going to put in three lesions and compartmentalize the right atrium. I don't have any data to present on that. I just highlight that I think it's safe and effective the way it is now but as more clinicians have a chance to use it, we'll get a better handle on how best to use it to benefit the population of patients with atrial fibrillation.

DR. CALKINS: Al, I just have a couple of comments about where it would fit into the armamentaria of the electrophysiologist. I think it would have a very important role. Right now, in the absence of this catheter being available, people say that their one option is a pulmonary vein isolation. Anyone who is doing this procedure realizes there is a learning curve and the learning curve is very rocky as you go up on it and the complications are like no other procedure that's ever been done in an EP lab.

So right now we have the electrophysiology community in a tough spot. They're hearing about pulmonary vein isolation. These articles are showing in circulation every week. They are getting pressure to do something for atrial fibrillation. All of a sudden they start doing it and the complications occur.

We don't know where this procedure is going. It's evolving rapidly from periostial ablation to perianatomic ablation to different mapping techniques to different energy sources. That's very much of a moving target.

What's nice about this procedure is that it's not a moving target. Here's three lesions which can be understood or reproducible. It's clear how to deliver them. It's a first step. So where it fits in for a center like mine where we can do whatever we please, we'll tell the patients that they have an option. If they want to go for the home run, the cure, and they're willing to accept a procedure with higher risks that's in evolution and so forth, we're happy to go ahead and do a pulmonary vein isolation.

If, on the other hand, that scares them, the idea of doing a procedure where pulmonary stenosis (PH) or stroke ?- are clear risks but they say "I want to do something. I'm refractory to drugs". We'll say "Let's go head and do a right-sided procedure, get that done and see if you improve. The data suggests that there's a very good chance that will make you better. Then if you don't improve we can always three, four, five, six months or a year from now do a pulmonary vein isolation and hopefully by then we'll know what tools are the best to do it with and we'll have a little more information so we can do it more safely."

For the other hospitals that don't have the ability to do pulmonary vein isolations safety or effectively, this is a procedure that instead of just doing a flutter line which I think does very little in this setting they can go in, do an atrial fibrillation ablation procedure, give the patient some benefit, make them feel better and then in the future, you'll have to a pulmonary vein isolation as an adjunctive procedure. That's a better way to be approaching atrial fibrillation rather than just doing a high risk procedure right from the outset.

DR. WALDO: I really have no other questions. I have one last comment though. I was surprised in the material that we got from the FDA that the company provided. I was surprised that in 2003 that the company still stated atrial fibrillation was due to multiple ?- I think that was the premise for a lot of this which I still find as a principle that the technology is way ahead of the science. That's always a problem. I think that's where we are now and we have made advances in this field. That's for sure. Thank you very much.

DR. EWING: I'd like to present the panel with some additional information if I may. This is Lesley Ewing. Just to give some information that the Agency has about transtelephonic monitoring, there is another submission that's under review. The data eventually will be available but they're in a follow-up of another ablation study. They are able to get near 90 percent compliance with transtelephonic monitoring in asymptomatic patients. So the Agency knows that it is achievable. Thank you.

CHAIRMAN LASKEY: Okay. David.

DR. SCHWARTZMAN: I think I have more comments than questions. First of all, I would like to echo the others on the panel in stating appreciation to both the sponsor and the FDA for excellent presentation and clarification. At the end of the day, the numbers can only take you so far and then there's interpretation. I respect the bias to either side.

I'm going to bring to bare the reason I'm here which is based on experience, primarily personal experience not with this particular technology but with the electrophysiologic, anatomic and clinical target and the vagaries and difficulties therein and also considering the data as presented. I first want to make comments against this proposal and then I'll make comments in its favor.

First of all, against, the starting point to me is that this claim of continuous linear lesions is certainly unproven. Again not having primary experience with this and without data presented otherwise based on my own experience, I think it is likely that the vast minority of these lesions were continuous in the sense of transmural and contiguous, therefore acting as arcs of conduction block as conceptualized.

Secondly, I think the electrophysiologic efficacy that is EKG burden reduction is really not clear to me but my guess is it's probably under 50 percent at six months. Again bringing experience to bare, longer follow-up is going to probably likely be worse. I know what the reducing burden data had shown from three to six months but again primarily from my own experience this is not a disease that remodels favorably over the long haul.

In that regard, I'm concerned that we're approving a protocol or a technology technique that is going to end up being a waste of patient/physician time and money. My own personal experience with right atrial linear ablation although with similar numbers early at two years only 10 percent of these patients remained without another procedure whether that be pharmacologic or non-pharmacologic. So at the end of the day long term benefit is hard to find.

Finally I have concern regarding the safety of this technology in its inevitable use in the left atrium as soon as it gets into the hands of those of us who have our own visions about this technology. Some of the charring issues, the entity of thermometry which I think is a terrible method of guiding power titration and generally under represents interfacial temperature which is what relates to charring and presumably cardioembolic material. We get away with a tremendous amount in the heart particularly in the right side. I'm concerned about its extended use whether or not it's labeled.

On the favor side, it is clear that some ablation burden is accomplished here. It is unclear clinically that this concept of continuous linear lesions are either needed or important. So in that regard the technology because it achieves some degree of ablation burden is a viable ablation technology.

A substantial minority of patients appear to experience a significant burden reduction. That is electrocardiographic burden. A majority of the patients appear to feel better. So clinical burden reduction is apparent and notwithstanding the limitations that have been laid out pretty clearly. But again practically it is very difficult these kinds of studies without a control arm.

As already stated in the absence of longer term data notwithstanding our arrogance about pulmonary veins and left atrial ablation and left atrial being the seed of all evil, it is probably reasonable to start with the right atrium. This is particularly true and a very important point for the masses of electrophysiologists who are new to this area who are getting tremendous pressure from their community to do something as these guys said. It is a time extending situation that allows that experience to be gained and refinement of left atrial ablation technologies that will have also served a viable purpose. Those are my comments.

CHAIRMAN LASKEY: Wonderful. Thank you. Sharon.

DR. NORMAND: I also echo the comments and enjoyed the presentations by both the sponsor and the FDA. I actually have several comments and not surprisingly they have to do with analysis. I have some questions about really which patients should be included in the analyses. I do have questions about sample size.

I have some very important questions I feel about determination of the primary endpoint. I have some comments about the secondary endpoint as well. I know not much has been said about that but I do some comments about it. I do have some questions regarding some of the assumptions and verifications of the assumptions that have been made both by the sponsor and as well by the FDA.

Part of my question at the very beginning about clarification of the numbers related to really what are we studying here. What's the investigational device? When I look at the numbers that I ask for clarification for, it seems to me that some patients shouldn't be included.

This is not a clinical trial. This is an observational study. There is no sense of intention to treat in this type of analysis. So I'm a little confused I guess as to why some patients would be included in the analysis. I'm stating that. I don't think anybody can argue me over to the other side on that. I just want to lay that out.

That is some patients and I think it's at least 20 if I understand the numbers correctly that I think I would not have included certainly in the primary endpoint. They didn't receive the procedure, the device for which we are supposed to be assessing.

I do have questions that hopefully someone can answer regarding sample size. This has a little bit to do with maybe the bouncing of numbers that I've seen both from the FDA as well as from the sponsors in terms of the number of available observations at baseline and at six months.

I just want to get a sense of why some of those numbers were going away. I understand that some people are lost to follow-up. I'm just trying to figure out what the right denominator here for the primary endpoint. Is it 81? Is it 87? Is there any meeting or convergence of the minds in terms of that number?

The reason why I'm raising this issue is for some reason there was longitudinal data collected yet no use of it certainly in terms of analysis of the primary endpoint, some use of it for other answers. I'm a little concerned about that. I just want to get a sense of why we're missing some data for the primary endpoint. I know maybe the sponsor can tell me. Is it 87? Is it 83? Is it 81 and whether or not the 20 or 22 depending on how you count people are included in that denominator?

Let me be specific. You have a number for your primary endpoint. Either it's 87, 83 or 81. It's in the 80s. Are the 22 people who didn't report included in that number that is who report about their symptoms included in that denominator? It should be simple. It's yes or no.

DR. KOCHERIL: Abe Kocheril. A very good question. Of the 22, we had redundant measures as we've seen. Based on symptom reduction by other measures, the office visit, case report forms and by the AFSS, 10 of the 22 did have the requisite symptomatic reduction. They were in the analysis as successes. The other 12 did not.

DR. NORMAND: But are they in the denominator. I don't care about the numerator. They are in the denominator.

DR. KOCHERIL: Yes.

DR. NORMAND: Terrific, okay. So I'm going to ask some more questions about the primary endpoint. I'm just curious as to why quasi experimental design wasn't used. In other words, you have an observational arm. You're looking at prospective data collection. You could have had some sort of other group. Certainly the other that I'm throwing out here could have been used. Why did you just go for a interruption time series we would call it in terms of not having a control group?

DR. KOCHERIL: I don't understand the question. Please phrase it again.

DR. NORMAND: Why is there no prospectively a control group?

DR. KOCHERIL: There's a history as to how this protocol was developed. It was a collaborative effort between Cardima and the folks at the Mass. General and the FDA. For people who are drug refractory, this was the recommended setting design. If they weren't drug refractory, then a randomized control arm would have been used.

DR. NORMAND: Okay. I noticed that you have a multi-center study. Did you adjust for the fact that you have multiple sites in the analysis? Just so you know, all my questions are just going to add more variability to your answers. I just need to know whether or not that was adjusted.

DR. KOCHERIL: It was not weighted if that's what you mean. Basically all the patients were thrown in --

DR. NORMAND: No, it's not a weighted question. It's whether or not there was an extra component of variation due to the fact that you had multiple sites. So probably not.

DR. KOCHERIL: I'm not aware if we did.

DR. CALKINS: Hugh Calkins. I'm not a statistician and it is true that many sites participated in the study. I think actually that's a strength of this study.

DR. NORMAND: I'm not arguing that it's not a strength.

DR. CALKINS: Single center studies where Center X can 100 percent success during atrial fibrillation with a pulmonary vein approach. In that person in that center with whatever they used for follow-up, they can get those results. In this study, you had 20 different centers and each center may have more than one investigator doing the procedure and the data was pooled. There was not a subanalysis but I think that speaks to the value of widespread of hands that these are the results that you will get with this technology.

DR. NORMAND: I understand. I do think it's a strength but statistically you also need to adjust for the fact that you have multiple centers. So didn't adjust?

DR. CALKINS: No.

DR. NORMAND: I guess the next thing is whether or not you did a stratified analysis. I noticed in your sample size calculations it looked like you just did a single analysis but it looked like it really should have been a stratified analysis depending on the baseline entry. In other words if I understood correctly and I may not have understood correctly depending on the number of episodes at baseline, it was a different expectation in terms of the size of the benefit that would have counted as a success. Is that true?

DR. CALKINS: Yes, if they had more frequent episodes of atrial fibrillation, they only had a 50 percent reduction.

DR. NORMAND: Got you.

DR. CALKINS: But what we saw in the data and if you remember the slide showing the 81 patients, the most symptomatic patients didn't have a 50 percent reduction. Those were the patients with the 90 percent reduction in atrial fibrillation episodes. So although that was the cut-off in fact it was a non-issue because the most symptomatic patients had a 75 percent reduction just like the less symptomatic.

DR. NORMAND: But it wasn't a stratified analysis.

DR. CALKINS: No.

DR. NORMAND: I just want to get a little better understanding of the definition of the primary endpoint. Let me just be specific so you understand exactly what I'm asking. I don't want to waste time. My understanding is that the number of symptoms, let's say in the baseline period and then the sixth month period, the patient was supposed to report weekly. Is the total number of symptoms that's represented in the baseline period over the four week reporting period or is it an average per week?

DR. CALKINS: No, it's a total number. They were instructed to transmit when they had symptoms of atrial fibrillation at baseline and then during the sixth month point. So the routine weekly transmissions when they were asymptomatic do not affect that number. The number is how many symptomatic episodes of atrial fibrillation. What we found is baseline when the patients thought they were in atrial fibrillation only 50 percent of those transmissions actually showed atrial fibrillation --

DR. NORMAND: I'm sorry to interrupt but I think you misunderstood my question. I probably wasn't clear in the question. My question is the reporting is supposed to be on a weekly basis. Sometimes some people reported four weeks in baseline period and maybe some people just transmitted for two weeks. Is that correct?

DR. CALKINS: No, the patients were instructed to transmit for a month. Every time they had symptomatic atrial fibrillation, they would transmit anytime throughout that month whether it's once a day, whenever they had symptomatic atrial fibrillation they were supposed to transmit. Even if they were asymptomatic, they were supposed to in addition transmit once a week regardless of symptoms.

DR. NORMAND: I got you.

DR. CALKINS: The number we're looking at is how many times they had symptomatic atrial fibrillation that was transmitted and confirmed to be atrial fibrillation baseline versus at six months.

DR. NORMAND: See my concern is coming in if there's a different exposure period for the reporting ?-

DR. CALKINS: The same 30 days.

DR. NORMAND: But if some people even though they were supposed to report all the time don't then you are measuring on a different period of exposure period if that's what you want to think of it.

DR. CALKINS: Yes, I think in the absence of an auto-triggering device you implant in the patient that takes them out of the picture all you can do is hope that they are as compliant as they can be. There's no way to continuously monitor them with today's technology.

DR. NORMAND: Then that led me to the question of whether or not it was adjusted in the analysis in terms of let's say at month six when they are reporting when they are supposed to be transmitting weekly. If you had that I transmitted three times in month six and you transmitted once in month six and let's say for me there were two episodes, did you account for the fact that I transmitted three times and you transmitted once the information?

DR. CALKINS: No, because the patients were supposed to instruct when they had symptomatic atrial fibrillation so they may just transmit 20 symptomatic atrial fibrillation episodes that occurred in a week but not transmit the asymptomatic times. So there was no real way to look at that or correct for that or divide that. Although half the patients had no atrial fibrillation episodes at all. So depending on how you correct for zero, it's still zero.

DR. NORMAND: I'm going to act on the concern that other people raised about the 20 or 22 people who didn't report. One can't assume that it's zero. I guess I'll argue the other way even if I went to see my doctor, I might be afraid to say that I did have a problem but forgot to transmit. I think there are some issue with actually assuming it's zero.

I understand that there was some verbal communication which I take exception to in terms of the reliability of that. I understand you did the best you could. But that also means that you have differential ascertainment bias in mixing the information from those that actually transmitted and those that reported verbally. I'm just wondering about the bias that it might introduce. You probably don't have an answer to that but it does just add another level of concern about the quality of the data for the primary endpoint.

I think a prudent approach should say it's missing. Then the missing data one needs to think about other assumptions about the missing data. That is whether or not you assume things were missing completely or random. Missing at random I realize these are technical terms but I'm sure somebody does your statistics. There is some things that you worry about in terms of those types of assumptions.

I'll go on to the secondary endpoint. I'll be brief about that one. This is just my being naive. I'm not sure what a clinically important change in that particular SF-36 scale is. Is 10 point clinically meaningful? I have no idea.

DR. CALKINS: I think that is felt to be that clinically meaningful.

DR. NORMAND: Okay. I think in the reading of materials I understood that you did something with that other scale which I'm going to forget, that AFSS scale. I didn't understand what was written. It sounded like to changed the scoring of that second scale. Is that correct? In other words, how you actually score a patient on the second scale.

DR. CALKINS: If you look at the two scales, the atrial fibrillation severity scale is particularly designed for patients with atrial fibrillation. So if you look at the results, that's where you see consistent improvement and greater than an average of 10 point improvement all across the board. For the SF-36, there were just two or three of the seven markers that showed that 10 point improvement. Those were the ones that were most significantly depressed but that's a very non-specific marker that doesn't specifically look at atrial fibrillation. Let me have our statistician speak to it.

MR. STUBBS: Hello, I'm Harrison Stubbs. I'm a paid consultant for Cardima. I have no equity in the company. I'd like to comment on a couple of things, just to step back a few steps if possible.

First you asked about a stratified analysis of the success episode reduction according to baseline categories. We did that in response to an FDA deficiency and we turned that into them. Frankly we saw no trend or real difference in the response rate according to baseline number of episodes.

Secondly regarding the AFSS instrument, that is not scored on 100 point scale. We wanted to convert that to a scale that would be congruent to the SF-36 domains. Consequently we scaled it to an 100 point scale with low numbers being bad and high numbers being good.

We were concerned about compliance too. In response to FDA's concerns, we took a good look at compliance. We categorized compliance a couple of different ways. We looked at those that didn't transmit versus those that did. We looked at those that transmitted three or more episodes in the month six versus those that transmitted fewer than three. We chose that break point because it divided the sample in about half so we figured we optimized the discal power to address any differences.

We looked at AFSS instrument in particular because the three scales of that instrument are episode frequency, severity and duration. Those three go in to make up the total AFSS score. In looking at that and also the SF-36 according to these two ways of categorizing compliance at six months, we did not see any significant differences in any of those parameters.

It is important in particular that the AFSS scales showed no differences because the concern one always has with non-respondents and non-compliers is that they are doing characterizing worse. At least according to the AFSS score which is an AF specific quality of life instrument there is no evidence of that in the results that we looked at.

DR. NORMAND: Thank you. I think one of the concerns I would have with the analysis of the secondary endpoint is power for you not to find differences. I think you have some small sample sizes to make comparisons. Moreover you have several subscales that you are comparing. If anyone cares about multiple comparisons, I don't think you adjusted for the multiple comparisons you were doing.

MR. STUBBS: No, but nevertheless there were none that were statistically significant.

DR. NORMAND: I'm just saying that it's power. A lack of power would say you might not find differences.

MR. STUBBS: That's true but you have to balance that against the multiplicity of the number of comparisons.

DR. NORMAND: I actually think that's all I have for now. Thank you very much.

CHAIRMAN LASKEY: Chris.

DR. WHITE: Thank you, Warren. I actually don't have much really to add to what's been said by my colleagues. I would like to ask the PI though maybe in the sponsor that as many of us are involved as being investigators in trials, I don't quite understand the freedom of your investigators to deviate from the protocol. For example, the 20 times that study catheter was never used.

Hugh was mentioning up here that if they were familiar with the catheter, they might abandon a little easier and gone with another catheter. I don't understand never trying the experimental catheter. What was going on in their minds? These would be protocol deviations that would come up as red flags right off the bat. I don't know why the waters were allowed to get so muddy down the road.

DR. KOCHERIL: Abe Kocheril. I'll try to tackle that one. In Phase IIb before the NavAblator^TM was available, the investigators were told to use the REVELATION_ Tx and it was tried. Hugh pointed out the variables that are specific to the operator and basically several of the investigators right away had bad experiences trying to get an adequate lesion and the investigators started deviating in Phase Iib.

In Phase III, the NavAblator^TM was offered as an option but by that point there was already this mindset that a flutter line is a flutter line and that what you used for it didn't make a difference. This deviation was tolerated because on the part of the investigators we felt that way too that no matter what you use to get a flutter line, it doesn't matter. You just need to get a flutter line as part of this procedure.

DR. WHITE: Was there insight at the time that you would have that price to pay today?

DR. KOCHERIL: No, I think we can.

DR. WHITE: Was there insight that you were contaminating your data that you weren't following your own protocol that you set?

DR. KOCHERIL: Looking back on it, that's certainly an issue. At that point, this was the subject of investigator meetings and we would get together as a group from all the 20 sites. In all the discussions of the flutter line, the prevalence of opinion was that a flutter line is a flutter line. There wasn't a distinction made.

I also add that the getting a flutter line with Tx is also a matter of persistence and dedication to it. Having had the earlier experience with the earlier version of the REVELATION_, I did persist longer. What it takes to get an adequate flutter lesion if you are going to use that catheter to do it is basically making multiple loops and getting into the nooks and crannies of the flutter isthmus to be able to do it. That's time consuming. You have to be persistent to do it. It can be done. For most investigators since the flutter line did not seem to be what would make or break this protocol, it was easy to slip over to a catheter that they were comfortable with.

DR. WHITE: Were any of your sites terminated for non-compliance with the protocol?

DR. KOCHERIL: No, I don't think they were.

DR. CALKINS: Just one other comment, it is true that with the flutter line we had these 28 patients that were ablated only with some alternative catheter and the NavAblator^TM or linear catheter was not used for the flutter isthmus right from the get go. This really reflected the bias of the investigators to do what was best for the patient. They felt the other catheter got the job done better.

DR. WHITE: That may be so but you're asking us to approve this data upon protocol you wrote and then you deviate from that protocol. I don't know how to make up for that. I understand what you just said. When I enroll patients in protocols, I do things according to protocol. I don't get a choice to say what I think is best for the patient.

DR. CALKINS: Two points. The linear lesions which is a really unique feature of this protocol were all done with the REVELATION_ Tx catheter. I think that's what we're mainly focused on today. The issue about the NavAblator^TM and the NavAblator^TM ended up being only used in 60 patients and whether that should be officially part of the system where we now have an improved flutter catheter on the market that should be available to create a flutter line with whatever catheters you want. That makes sense.

In the beginning of the protocol the way it was initially designed was it instructed the investigators to first try the flutter line with this linear catheter but it did not say you cannot use any other catheter to get the job done. It said first try this. Then use what you otherwise would do as far as clinical practice.

So the investigators that started going to use another catheter, it wasn't that they were deviating from the protocol. The protocol said first try theirs but to get the flutter line, you can use whatever you want that's available to you to get the flutter line as part of the protocol. So it was not viewed from the company's perspective as pull the plug in the center because that's built into the protocol that the flutter line can be done with some other catheter as part of the protocol.

What we learned as this study started five years ago, there were 10 different companies out there with linear catheters trying to ablate the isthmus. Guess what? It doesn't work with any of them. So we were trying to do the impossible. Then it just became this preference of catheter feel issue. But the protocol was not designed if the flutter line was done with another catheter that's a failure. It was first try with our catheter, get a flutter block with your standard clinical care.

DR. WHITE: But you just said "Try with our catheter and then if you don't..." You have a whole bunch of patients in here who weren't tried with your catheter.

DR. CALKINS: I know. By the end of the protocol, some of the investigators really preferred X catheter and these were all these 4mm catheters.

DR. WHITE: But I don't understand that. I'm in protocols right now where it's clumsy to do the thing the study way. It takes a long time. I think if you sign up for a protocol, you play by the rules. If you don't play by the rules, you get terminated. You get kicked out of the protocol if you don't follow the rules. Or you pay the price you're paying today for the confusion this generated. I think that's the lesson.

DR. CALKINS: Yes, at least as investigator for me there is no confusion. It's very clear that a flutter line is a flutter line from a physiologic standpoint, from a therapeutic standpoint. That's not what we're here today is to approve a flutter catheter if you will or to look at the safety or efficacy of a flutter catheter. We're really focusing on a linear ablation system.

DR. WHITE: Actually you don't get to focus. I'm focusing.

DR. CALKINS: You're focusing.

DR. WHITE: Thanks. What my problem is if you're sloppy enough to do the protocol this way then I think that runs through the whole study. That contaminates the study. It's not that a flutter line isn't a flutter line. It's that you wrote the protocol and then you didn't stick to it. I don't like that. That bothers me in terms of the reproducibility.

Al talked a little bit about the rigor. I think this is part of the rigor of following the protocol. I think it's the sponsor's job and the PI's job to make sure people are sticking to the protocol. If they are not, then there is disciplinary measures and you exclude them or you terminate them or you say "You're not playing by the rules so you can't do this anymore."

I just think you didn't do that. You didn't take that very seriously. I'm telling you from my perspective that it's a very important part of doing these kind of investigations. You have to have the discipline to follow through with what you and the FDA set about as being your protocol. I just see those deviations as problems. That's all I have.

CHAIRMAN LASKEY: I'm confused now. Are you backing away from the NavAblator^TM portion of this PMA?

DR. CALKINS: If the question is what is the tool that I and the EP community would like to have access to, it's a safe linear catheter ablation system. It's not the NavAblator^TM. It's the linear ablation, the REVELATION_ Tx. Quite frankly, we have plenty of catheters we can make flutter lines with either on-label or off-label now that there are some officially approved. I feel strongly there are all completely interchangeable.

And this catheter was never, the NavAblator^TM design, to be the best feeling, best handling, best driving flutter catheter on the market. That was not the purpose. That was made to address the FDA's concern about this issue Dr. White brought up about having a tool so that you can complete it as best you can.

If there is a tool that we need, it's quite frankly not the NavAblator^TM. Yes, we would like to have the whole system approved but the critical thing is the REVELATION_ Tx which is the unique technology that I think is most critical to making these other thin linear lines that are essence of the procedure.

CHAIRMAN LASKEY: All right. Is the only reason we're spinning our wheels here about this because your sample here is enriched or contaminated by the AF? How often do you encounter this in general practice? Is this an unusual group? You had a quarter of your people with flutter. Is that usually an issue with your garden variety AFs?

DR. CALKINS: The issue here is part of the protocol. The three lines that were prescribed were the flutter line, the SVC line and the septal line. They are the three lines of this right atrial linear approach that we're discussing. Much of the discussion of both Dr. Ewing and our discussion today is on what catheters pulled out to make that flutter line. Now as far as patients in the study, they got in because of atrial fibrillation.

CHAIRMAN LASKEY: Right, but they have a large presence of atrial flutter here.

DR. CALKINS: Yes, and in any population you'll see both atrial fibrillation and atrial flutter. If you show an EKG and they show rapid flutter one time, atrial fibrillation the other time, I think the two disorders go hand in hand to a large degree.

CHAIRMAN LASKEY: For that matter, what goes on in the RA may or may not go along with what goes on in the LA. Can the left atrium be fluttering and the right atrium be fibrillating or visa versa? How does that effect this approach?

DR. CALKINS: Dr. Waldo may correct me but either the heart is fibrillating or it's fluttering. Yes, there have been some cases where you do some ablation procedures and electro-isolate one part of the atrium which you can refibrillate and the rest can be fluttering. In general at a given moment, you're in atrial flutter which is a regular rhythm or you're in atrial fibrillation which is more rapid irregular rhythm.

CHAIRMAN LASKEY: We just need to be clear about what it is we're treating here or what you're treating. I just had one comment statistical in passing. When I was bored on a plane ride, I went over your sample size calculation and I went through the upper and lower confidence limits that you had quoted on page 14 for a 50 percent success rate.

Actually the calculations that I come out with using the standard roughly 1.96 times your standard error giving you the upper and lower limits, you don't make it at the .05 level. Did you go through the exercise of your beating your lower 95 percent confidence limit for your clinical success rate? If so, can you point out where I erred?

MR. STUBBS: Harrison Stubbs again. When we put the confidence interval figures in the original protocol, that was done with the exact method as opposed to an approximate method which you are using. But I'm unclear on your question regarding whether or not the lower confidence limit was met.

CHAIRMAN LASKEY: Generally you want to beat that since you don't have a control arm here. You just have a single arm study. So basically you want to beat some estimate of your estimated success rate. Generally you pick your lower confidence level and try and beat that.

MR. STUBBS: Okay.

CHAIRMAN LASKEY: Using the numbers that you actually observed and computing the lower, it doesn't work.

MR. STUBBS: Yes, I follow you. In the protocol, we chose the sample size of 80 as providing clinically acceptable precision for a wide range of possible success rates going from 30 to 70 I believe. It turns out that a success rate of 50 percent gives you the worse possible case.

We did not set in any OPC or any particular minimum confidence level that we needed to achieve because in discussions with the FDA statisticians it was decided that it's not clear what level of success for this procedure would be clinically useful. We were going to leave that to the panel to decide. So our responsibility in designing the study with 80 patients was solely to achieve an acceptable level of precision. The success rate fell where it may. It's up to the EP community and the clinical community to decide whether that's acceptable. Does that explain your question?

CHAIRMAN LASKEY: Yes, thank you. It's just quite a gamble. I'm sorry, Cindy. It's your turn.

DR. TRACY: I'm a bit more confused than I was earlier this morning I'm afraid. One of the questions that I think is pretty relevant for thinking about this whole thing is what is an acceptable level of effectiveness. We have a range between 47 percent and 81 percent. For a condition that in terms of treatment, palliation is probably an acceptable outcome. Is 47 percent an acceptable level effectiveness? I would argue that it probably is.

Often the thing that you do one day is not going to be the last thing that you're going to do with that patient for the rest of their lives. So even though you may end up a year later needing to do a pulmonary vein ablation, that probably is okay to give them a year with some improvement in their symptoms. So even if you accept the lowest level of effectiveness, it's not terrible.

I guess though I'm troubled by the mixture of catheters that were used. I agree with you that a flutter line is a flutter line regardless what catheter it's made by. I was playing with this across my knuckles and there's no way this thing could ever make a lesion that would be effective for flutter.

I'm a little troubled that a flutter line was made at all. I guess it comes to the question of what is atrial fibrillation in the first place. How does it start and why does it perpetuate and what good is a flutter line in fibrillation? That's a very confusing piece of this whole puzzle.

The other thing that I'm not so troubled by is the number of pacemakers having taken care of a number of post-Maze patients and having put in Dr. Cox's patient's pacemaker in about probably 10 to 17 percent of them after they've had successful Maze. They always presume to be on the basis of preceding undetected or poorly recognized sinus node dysfunction. That's not such a terrible thought to me that a fairly high percent would end up with pacemakers although certainly those that had subsequent AV node ablations and pacemaker implantation are failure. I think that's pretty clear.

Another piece of how the study was run is it's clear that people were trying to achieve bidirectional block with their flutter ablations and they knew that they weren't successful somehow. Either they had ESIs in or they had Cardos in or they were doing pacing techniques to determine bidirectional block. It's not so clear to me that the other linear lesions were that systematically looked at to know whether there was bidirectional block around the linear lesions. So maybe somebody when I stop talking can comment on that.

DR. CALKINS: Hugh Calkins. Cindy, the only line that was confirmed was the flutter line in a conventional fashion with a halo catheter. The other lines the protocol said nothing about confirming block or demonstrating block. It was you put the line in these positions. You demonstrate lesion creation by seeing some change in the electrical properties of the myocardium. You usually see a decrease in electrogram amplitude in a very striking fashion.

But that was it. There were requirement for testing verification and finding out if there's lines as David pointed out. I would agree that it's very unlikely that every one of these lines is continuous, transmural and anchored at both ends. The chance of that is very unlikely.

The goal was lay these lesions down. Not all of them are continuous, transmural and anchored at both ends. But does that strategy decrease atrial fibrillation burden? This is really the question that we're asking. David's point is correct. We really don't know. Does it have to be completely transmural? What if it's transmural in part but not all? Does that have an effect?

We were really looking at the strategy not with concomitant mapping systems which makes the procedure go from one that's straight forward to one that's very complex. It has to be performed in labs and have all this fancy technology and adds tremendously to the overall personal time, lab and equipment expense of doing that.

DR. TRACY: Just to get a sense again for the procedure, there are eight areas where energy is delivered a minute each so that's eight minute per line. But some of the procedure times were much longer than the 32 minutes. So how many lines per patient was average?

DR. CALKINS: The way the catheter is delivered is it goes through the Naviport^TM catheter to be delivered. Then you lay the catheter against the wall. In some patients and in some lines, you can have the whole catheter lines against the wall and be able to burn sequentially eight electrodes and that would go fairly quickly. If it's a lateral wall, you would have to do pacing to make sure you don't have phrenic nerve simulations so that would add some time.

But more typically you wouldn't have perfect contact in all eight electrodes. Maybe only the first six or the first four would have good contact. You do that burn and then you just pull the catheter down to where the bottom ones started and then put another four to six to eight burns in and then pull down until you're down in the IVC.

So let's say if there's three lines, each line took at least two catheter positions to create it. Then you have placement of standard catheters, creation of the flutter line, validation of the flutter line and now we know that flutter lines are best created with other types of catheters, irrigated tip or 8mm tip or high energies because it's thick. That is one way that makes the procedure shorter today when we are doing flutter ablation just because quickly go to higher power technologies for that part of the procedure.

DR. TRACY: Just a couple other things. A safety issue is that I noticed that there were a relatively high number of skin burns. My guess is that those aren't as patiallated (PH). Those are related to radiation burns. I noticed some where were the skin patches were, the R2 patches were but those act as antenna basically for the radiation. The fluorotime was at least in one of these cases 247 minutes which I would think that anybody would agree that's excessively long. How hard is this procedure? Are people going to inducing radiation damage in patients by doing this procedure?

DR. CALKINS: I think your question is a good one in terms of radiation exposure. At least my interpretation data there was no ulcer formation. There was no clear demonstration of radiation skin injury. I'm unaware of any data saying that the defibrillator tab patches attract radiation skin injury.

It's clear that it's greatest near the tube right where the tube is located. If you have biplane or what system, that's where you're going to get the highest skin exposure with some evidence of skin injury usually occurring a week or two later.

Overt evidence of radiation injury is clearly when you get these nonhealing ulcers that are a complete disaster. There were none of those in this protocol. It is true. You should have a radiation limit. In the old days in some of the earlier studies, 40 minutes was an average fluorotime. In the study, we looked at the cardiorhythm study where we looked at radiation exposure times. Some got up to as high as 300 or 400 minutes. I think 450 minutes was the longest for a standard SVT ablation.

It's clear that there's a cutoff that we proposed in that paper. I think, realistically when you do that procedure without using other technologies to try to validate lines when you go straight for the flutter line to a flutter catheter that's designed for that either an 8mm irrigated tip or whatever, the procedure takes one to two hours with 40 minutes of fluorotime.

Any electrophysiologist should know that there's a point where you should stop any ablation procedure whether it's four or six hours. That's a message that goes with saying for any of these procedures. I think your point's well taken. Let me turn it over to Abe.

DR. KOCHERIL: Abe Kocheril. On the burns that was analyzed further and those were actually related to cardioversions and probably not related to the radiation exposure. On the actual time of the procedure, I mentioned earlier in my presentation that there is a learning curve to it. When an investigator is starting out with this procedure, they want to convince themselves that they are getting an adequate linear lesion. They will take longer.

In my center, we've gotten it down to under three hours for total procedure time and 30 minutes or less for fluorotime. So if you take away measurements, it could possibly be even quicker than that to deliver these three lesions.

DR. TRACY: Okay and then just a couple of final points on the transmissions. There was a period of time at the beginning when patients were coming to enroll in the protocol. AF was between a month or if they didn't make it in a month they could be rescreened and they might go as long as three months making symptomatic transmissions. As we all who take of atrial fibrillation patients realize that not all of our patients are 100 percent accurate at determining it. But the ability of the patients to determine whether they were in atrial fibrillation or not seemed to decrease after the ablation procedure. Is that some nerve damage that was done that reduced their ability to detect atrial fibrillation or did they just really want to feel better after having gone through this procedure so that they couldn't detect the difference?

The other part of that is you're comparing one to three months of pre-data to six months post-data. My guess would be that as time went by the people especially the young and busy stopped making their transmissions.

Can you take a three month piece or a one month piece of post-procedural data and make a comparison recognizing that one month free of atrial fibrillation means pretty little? But is that possible to look at that? Would the compliance be good enough at one month or three months to make a look at that data be worthwhile? I don't know who I'm asking these questions to.

DR. CALKINS: This whole event monitoring stuff is very difficult. We have been going around it all day. Your point about the patients seem to be less aware of the atrial fibrillation in follow-up. I think when you looked at the number of transmissions baseline 50 percent of the ones were atrial fibrillation. At the end there was about 30 percent that transmitted atrial fibrillation. That denominator for that is the total number of transmissions whether it's the weekly ones that not everyone did. It's the just of the total transmissions we had how many showed atrial fibrillation.

DR. TRACY: These aren't symptomatic transmissions.

DR. CALKINS: These were the lump transmissions is my understanding. We just looked at all transmissions and how many of those were atrial fibrillation. I think right now the question is if you can put a REVEAL monitor in would that be the better answer. It seems you need something where the patient doesn't have to transmit himself or herself.

Dr. Ewing said in that another study they had better compliance. I'm not sure if the other study was atrial fibrillation or what the rhythm disorder was but I think a lot of it depends on who the population is that we're dealing it.

I'm not aware of better compliance in an atrial fibrillation study in a paroxysmal population like this at six months. Maybe there are better ways to do it. We can learn from the other studies that have been done however there may be better techniques to prod patients.

DR. EWING: This is Lesley Ewing. I'm going to provide to Dr. Tracy the raw data on transmissions that was submitted from Cardima to the FDA.

DR. TRACY: Okay. I think those were my major points. I'll take a look at this but pass the question over. Are you ready?

DR. ZUCKERMAN: Okay, before we continue now that the clinicians have gone around for the first round. I do think it's important to clarify the remarks made by Dr. White which were quite important because I've heard from the clinicians, many on this side of the table and that side of the table, that in atrial flutter line is an atrial flutter line perhaps from a clinician's perspective. The people on this side of the table wear a clinical, regulatory and statistical hat today.

The bottomline is that the only atrial flutter catheter that is approved right now by the FDA is an 8mm tip catheter. A lot of what you've been hearing or all the 4mm tip catheters would be considered off-label use which just underlines the point that Dr. White made about the necessity to follow protocols once they've been developed or the point that's made on slide nine of the FDA presentation again talking about what the protocol was during part three of this investigation.

CHAIRMAN LASKEY: Would you care to respond?

DR. CALKINS: At least my interpretation of Bram's question was more of a statement rather than a question. It is true that the only approved flutter catheter is the 8mm catheter. It is also true that the EP community uses whatever they like in the common irrigating catheter and so forth to create that line. His point is well taken but I took as a statement. Bram, did you have a question specifically that hasn't been addressed that I can answer?

DR. ZUCKERMAN: No, it's just protocol dictates that you have the opportunity to respond.

DR. CALKINS: The protocol as initially specified said initial attempts not that you create a block with the Cardima catheter but you initially try it and then you can use another clinically available catheter to create this part of the investigational procedure. If you look at it in that sense, there was never anything in this protocol or the FDA had agreed, that yes, try this catheter but you can use whatever's clinically available to do the flutter line. That is in fact what was done.

For the sake of patient benefit and the speed of the procedure and so forth, above all we're taking care of patients. A step was skipped and they did not use that first catheter. Dr. White is correct in his interpretation. But that's what was done. At the end of the day, it was taking care of patients.

DR. TRACY: Actually that's not at all my understanding of what the FDA presented. I thought clearly that the option at the point where the decision was made to permit a catheter other than linear catheter. An opportunity was given at that time to enter a contract with another catheter manufacturer or to provide your own ablation catheter for the study.

So maybe there was just an enormous misunderstanding in the investigators that they thought they could try with this and if this didn't work, then they could go whatever happened to be their favorite catheter. Maybe somebody needs to clarify that. That definitely was not my understanding.

MS. BALDWIN: Marianne Baldwin, Vice President, Regulatory, Quality and Clinical at Cardima. The original protocol in Phase IIb did in fact specify that the investigator could use standard institutional procedure to create only the isthmus line if the linear wasn't able to do that. Now that was the approved protocol that we went into the study with.

Actually we had also used it in IIa. In the course of IIb and IIa, it was made clear that we needed to have a ?- In fact let me back up just another step here and say that in the initial development in this protocol which is to treat AF, it was discussed and debated at some length about whether or not a flutter line was necessary.

It was concluded clinically that it wouldn't really be fair to a patient to go in there and do a catheter ablation and not do the flutter line. This is irrespective of an indication for flutter. This is to assure further effectiveness of the atrial fibrillation treatment. In fact I think one of the investigators, one of the people involved in the development of this protocol, said "It just wouldn't be ethical to not include the flutter line in this treatment scenario."

At that time, Cardima did not have a 4mm catheter. In fact as you've all heard, there was no flutter catheter approved. So the standard institutional procedure was an off-label catheter. There was no way to implement this protocol without that during Phase IIb and in fact during Phase IIa. So Cardima developed its own 4mm catheter for Phase III.

A large number of those patients that you've seen in the study whose flutter line was treated with a non-study catheter came from Phase IIB before the NavAblator^TM was available. So yes, there was a lot of confusion and by this time, a number of the investigators had been involved in the study in Phase IIb and in Phase III, the protocol itself still said "If you fail you still have the option of using the standard institutional procedure" because how you can not complete the lines. How can you not complete these lesions? That's not ethical. That comprises the characteristic of the confusion about that particular part in the protocol.

DR. WALDO: Al Waldo. Can I just make a comment?

MS. TILLMAN: Dr. Waldo, this is Donna Tillman from the FDA. I just want to make one clarification. If you look at the FDA presentation slides seven and eight, I just want to reiterate that nothing that we do ever is intended to interfere with the physician's ability to treat a patient where necessary.

However we did repeatedly instruct the company that use of non-investigational catheters in the trial would be considered a failure. If you look at our slides, historically we've told the company that on numerous occasions. So I think that we need to separate what was appropriate to do in terms of treating the patient which nobody is going to argue with the physician's best judgment versus what FDA told the company in terms of how the data in the study would be ultimately available.

DR. WALDO: Can I make a comment?

DR. TRACY: Al, the 42 percent that you reported as your success rate that's counting those as failures.

DR. EWING: The 42 out of 88 includes the success of the procedure. Only 24 out of the 88 were considered a success using only the Cardima catheter. So it's a difference in terms of evaluating whether the procedure is a success, taking care of the patient using whatever methods you have versus the FDA's evaluation whether a device works for that procedure.

DR. WALDO: Can I make a comment here?

CHAIRMAN LASKEY: Al, hang on a second. We're just trying to handle these one at a time because the slippery slope just got extremely slippery. We have a dialogue here but the company gets to respond as well. I don't limit this between Cindy and the FDA. I'll have your response to Cindy's question. Do you have a follow-up? I think you were very helpful, Marianne, but do you want to further clarify this?

DR. KAY: This is Neal Kay. Cindy, I think that this is getting at important issues. What that percentage of success is depends on a lot of these definitions. If you do that number, it comes out 30 something.

DR. TRACY: Twenty something.

DR. KAY: The question you guys are going to have to weigh obviously is in view of everything the safety I don't think we can argue very much about that but whether that efficacy is good enough in this patient population. I think it's going to be in that 40 percent range. That's my only comment.

DR. TRACY: So it would be 20 something percent successful strictly by protocol. Forty something percent if you permit the non-protocol catheter. Eighty percent if you go by the company's definition.

CHAIRMAN LASKEY: Dr. Waldo.

DR. WALDO: Can you hear me?

DR. ZUCKERMAN: Yes.

DR. WALDO: I think there's something important to add to this. Hugh earlier when talking about the flutter lesion I think I heard him correctly. He suggested that there was some question about whether that was even needed.

I think it's important to keep in mind where those other lesions are made. There's a big lesion between the vena cavi that's made. That's a setup for making the atrial flutter circuit. In many ways, the atrial flutter lesion is critical to prevent the development of flutter.

In other words, a lesion of atrial flutter isthmus is almost a requisite regardless of whether you think it has any value in treatment symptoms per se. It has value preventing flutter that may be caused by one of the lesions that makes a line of block between vena cavi. That's just a mechanistic comment that wherever this goes that probably is an important aspect.

CHAIRMAN LASKEY: Again you have the prerogative of responding.

DR. CALKINS: Al, I agree with your comment completely and that's why when the study was designed in terms of the flutter line, it would unethical to go in there and do an ablation procedure and not put a flutter line in. That speaks to Marianne's comments and the way the study was designed that the flutter line was part of this protocol even though the REVELATION_ Tx is designed to create linear lesions in other regions of the atrium.

DR. WALDO: Right.

CHAIRMAN LASKEY: All right. I think we've been around this. I suggest we take a 10 minute break. No more and then we will regroup around for the panel process. Off the record.

(Whereupon, the foregoing matter went off the record at 3:32 p.m. and went back on the record at 3:45 p.m.)

CHAIRMAN LASKEY: Back on the record. I'm on the finish line here so everybody can get to their plane. I'm going to have Geretta read the FDA questions at this point.

MS. WOOD: I will read each question and then the panel members will have a chance to respond. The sponsor conducted a single arm, multi-center study. The study protocol identified three linear ablation lesions to be performed with an optional fourth lesion. The indications for use statement is based on the procedure dictated in the study protocol.

Question No. 1. The sponsor presented acute procedural data on 116 patients all of whom had the posterolateral and posteroseptal linear lesions performed with the REVELATION_ Tx catheter. The tricuspid isthmus lesion was made with a variety of catheters throughout the two phases of the study. Please refer to the table following retrieved from table A-19 of the submission. A. Please discuss how the multiple catheter combinations affect the conclusions that may be drawn from this study.'

CHAIRMAN LASKEY: In order to help engender the consensus process, let me just start off and say that it seems to me the entire afternoon was spent discussing the limitations of multiple catheter combinations as articulated by Dr. White and almost everybody else. It went from protocol violations on down from there. Do we have any other things to add to that consensus statement?

DR. TRACY: No, I think that's right. I think the success rates go if you just look at only using the protocol catheters 20 percent and if you look at procedural success by FDA calculations 47 percent and if you look at procedural success using the sponsor's calculation 80 percent. So again it's an issue of variation in the amount of success.

CHAIRMAN LASKEY: So at best they confound the interpretation of the conclusions and at worse they preclude making conclusions. All right, B.

MS. WOOD: Please discuss the ability to analyze the device outcomes versus treatment outcomes in this study. In particular, can you comment on whether the safety and effectiveness results for this study may be attributable to one specific catheter? Do the treatment strategies employed in the study support the proposed indications for use statement?

CHAIRMAN LASKEY: Well again reflecting on the comments, I think part B stems from part A that the different treatment strategies necessarily employed the different catheters. There are two sides of the same coin. I think the answer then will be the same.

DR. TRACY: Except you might be a little bit more generous saying that if you look at treatment outcome and you take the most pessimistic analysis, it's 40 something percent treatment success.

DR. GILLIAM: But I'm not sure you can even say that because we're making the assumption that there is some success in the six month follow-up. Given the possibility of actual data, the fact that if you are making the assumption that no news is success and it may be that no news is continued atrial fibrillation. So my suggestion that we may be being very generous to suggest that it's 20 percent or even 40 percent. We may even have no difference between a control group and an interaction.

CHAIRMAN LASKEY: Dr. Waldo.

DR. WALDO: Yes.

CHAIRMAN LASKEY: Are you still with us?

DR. WALDO: Is that Rosie talking? I agree with him.

CHAIRMAN LASKEY: Thank you.

MS. WOOD: The study identified the acute procedural effectiveness endpoints as demonstration of either the reduction in the amplitude, fragmentation or wideness of local electrograms or appearance of split potentials. These endpoints were not consistently measured and/or recorded in the data collection forms.

The sponsor indicated in the PMA submission that no conclusion can be made about the acute procedural success endpoint. It is unknown what measure or electrogram characteristic was used by the individual investigator to determine when the linear lesion was completed. Therefore each investigator may have performed a slightly different ablation procedure. Please discuss how the lack of measurable procedural endpoint affects data analysis for this clinical trial.

CHAIRMAN LASKEY: I think this was nicely summarized by Dr. Maisel who wondered when to tell the operator to stop. What were the endpoints and the responses were not crystal clear. Is that fair?

DR. MAISEL: I agree that it seems unclear exactly what the procedural endpoint is. You can argue that the procedure as performed and simply look at the effectiveness data to decide if the procedure is effective with these vague and general statements about the procedure. I think Part IIb which we'll get to in a minute is the more important aspect of the procedural endpoint.

MS. WOOD: Please discuss whether the study provides sufficient information to instruct the user of the catheter system as to procedural goals or endpoints when treating an individual patient.

DR. MAISEL: In my mind, this is the more important aspect of the procedural endpoints with regard to both safety and effectiveness and the physician knowing how much to do and when to stop.

CHAIRMAN LASKEY: Therefore does the study provide sufficient information to instruct the user as to the goals or endpoints when treating the individual patient?

DR. MAISEL: I would say no.

CHAIRMAN LASKEY: I think that reflects the consensus of the group.

MS. WOOD: The primary effectiveness endpoint was based on comparing number of atrial fibrillation episodes captured and transmitted by event recorders during a baseline period and the six month post-procedure in patients on the same medication or a reduced dosage. It is unknown if each transmission represented a discrete atrial fibrillation episode. Given that the patients knew that a certain amount of episodes were required to be admitted into the study, please discuss the potential problems with accuracy in the counting of episodes at baseline and at follow-up.

CHAIRMAN LASKEY: I thought the sponsor actually answered this in a credible fashion but I'm not sure that's the essence of getting your arms around an event which has tremendous variability. The variability and frequency of atrial fibrillation or just editorializing PVCs for that matter requires a different approach. To answer your question here, there are problems with accuracy in counting an event which is admittedly highly variable at baseline and more so at follow-up.

DR. WALDO: Most studies with transtelephonic monitoring asks the patient to call back again when they think they've been in atrial fibrillation or when they were normal rhythm. So they establish that they return to normal rhythm so that a new episode is generally a new episode. That's part of the rigor of what we would like to have seen here. It asks a lot but a lot is needed. Unfortunately I think a lot of it is missing.

DR. GILLIAM: I would further suggest that a Holter monitor done before and after could have added some information. Perhaps informational people were having paroxysms so that they do have not just one episode of a persistent atrial fibrillation.

CHAIRMAN LASKEY: All right. Is it helpful then to say that particularly for the follow-up period better ascertainment needs to be ensured?

DR. TRACY: I would echo that. I think you need it for both periods because your primary endpoint is a difference between the baseline and the follow-up.

MS. WOOD: During the sixth month post-procedure, the protocol required patients to transmit weekly recording and in addition whenever they had symptoms of atrial fibrillation. Fifty-three out of 83 patients or 63.9 percent transmitted fewer than four times in the sixth month. Thus 36.1 percent of patients completely complied with the protocol requirement. Please discuss how incomplete compliance with transmissions of rhythm strips impacts measurement of the primary effectiveness endpoints.

CHAIRMAN LASKEY: I think the group has said repeatedly that it precludes interpretation of the endpoint.

MS. WOOD: The safety endpoint was listed of adverse events over 24 months and major complications that occurred in the first seven days post-procedure. A total of five out of 116, 4.7 percent, with the upper limit of 95 percent confidence interval being 9.4 percent, patients that had a major complication within a week of the right atrial ablation procedure. If all patients who required pacing within two weeks are included, the major complication rate would be 6.9 percent, eight out of 116 with the upper limit of a 95 percent confidence level being 13.0 percent. Please comment on whether the results of the clinical study provide reasonable assurance of safety for the intended use.

CHAIRMAN LASKEY: I don't think we were as concerned about the safety of the procedure as we were clearly efficacy. We dwelled on that but are there safety concerns? The rates of perforation and so forth are standard.

DR. TRACY: I agree. The safety concerns are not great with this. I think it's more the efficacy that's questionable.

DR. WALDO: Just one of the concerns by abandoning the anterior lesion, there was some concern about getting sinus nodes. I don't think safety concerns are a major problem here.

MS. WOOD: A total of 20 out of 88, 22.7 percent, patients had a permanent pacemaker implanted during the long term follow-up period. Nine of the 20 also had a AV node ablation procedure. In this patient population, please comment on whether this rate of permanent pacemaker implantation represents a significant safety concern for the device. Given the lack of a control group, please comment on how one would determine an acceptable rate of permanent pacemaker implantation.

DR. TRACY: You have a comparison. You can compare to any type of surgical intervention and see what the rates are on that. This does not seem to be out of line. Having AV node ablation represents not so much a complication as a procedure failure and therefore 100 percent of those who have an AV node ablation are going to end up with pacemakers. I think those are thought of differently so you are left with 17 or 20 percent. I'm not sure what the right number is at this point. You're left with some percent of the 20 that are having it because of sinus node dysfunction. That's probably not out of line with what you would see following a Maze procedure.

DR. WALDO: In my opinion this represents some of the problems with rigor in the study because we do know that there were some patients who came into the study already having a pacemaker. That I think was a whole lot better in some respects because it eliminated one of the imponderables at least for me. I think that it would have been much better to deal with sinus node dysfunction at the beginning so that it doesn't provide a confounding factor in this whole thing. I think that was a problem.

Whether it's excessive I'm not sure because I think the sponsors were right that there is a fairly high incidence of sinus node dysfunction. Some people estimate as many as two-thirds of patients with atrial fibrillation have sinus node dysfunction. So it wasn't a surprise that these people have sinus node dysfunction.

Then many of the drugs that we used to treat them be the AB nodal blocking drugs or antiarrhythmics further add to that sinus node dysfunction. I don't think it was excessive. It might seem to at first but probably it really isn't.

CHAIRMAN LASKEY: Are you satisfied with the determination of an acceptable rate of permanent pacer implementation? Is that response satisfactory? Just concurrent controls or historic controls or do you need more?

DR. ZUCKERMAN: Unless someone has a better way of making that comparison. Do any of the other electrophysiologists have suggestions?

DR. WALDO: I don't know how we can answer that definitively but I do think if you think about it, it really probably isn't a surprise. That would be my surmise.

DR. MAISEL: Maybe one other possibility would be to do formal sinus node testing just prior to the ablation while the catheters are in. Granted that's not the most sensitive method for evaluating the sinus node but would be an option.

MS. WOOD: The primary effectiveness hypothesis was: subjects with greater than or equal to five episodes during the 30 day baseline period. A reduction of 50 percent or more is required to constitute a clinical success. For those subjects with three to four episodes during the baseline period, a reduction of 75 percent is required to constitute a clinical success. These reductions were to occur in patients either on the same medication regime or decreased dosage. The results of the clinical study have been described in the presentation. Do the clinical data provide a reasonable assurance of the effectiveness of the system?

CHAIRMAN LASKEY: Well, very bluntly, I think everyone up here has said no. Any qualifications? Do you have consensus? We have pained looks but do we have consensus? Okay.

MS. WOOD: Please discuss the clinical utility of the primary endpoint of a percent decrease in atrial fibrillation episodes as opposed to a cure for paroxysmal atrial fibrillation. Does a significant decrease in PAF episodes constitute adequate evidence for effectiveness?

DR. GILLIAM: I think the problem again comes under defining whether there is a decrease. We don't have a handle on whether the numbers actually support whether there has been any change in the frequency, duration and the burden of atrial fibrillation post-procedure as compared with pre-procedure.

DR. MAISEL: But I would say that the endpoint as defined is adequate. It's just that there wasn't evidence to support it.

DR. GILLIAM: Yes.

DR. WHITE: The consensus is that the percentage would be an adequate one if you could measure it and met your prestated goals.

CHAIRMAN LASKEY: But we shouldn't be looking for obliteration, a zero frequency at follow-up. We're looking for a given percent decrease. Is that right? Okay?

MS. WOOD: A secondary effectiveness endpoint of the study was improvement in quality of life. Given the potential bias introduced with a nonrandomized, unblinded study, please comment on the device system's demonstration of improvement in quality of life.

DR. TRACY: It's all subjective. You have to have some assessment of quality of life because you're dealing with a condition that we have pretty good data that suggests that it's not a life-threatening arrhythmia nor is it terribly likely to affect longevity, etc. So you need an assessment of quality of life. No assessment of quality of life is perfect. It needs to be in there. I tend to believe that these people did feel significantly better. It's just that unfortunately we don't have a better way of determining why they felt better. The quality of life data is fine.

CHAIRMAN LASKEY: Yes, most of these tools were developed in the randomized trial literature. They were not applied to single arm, unblinded trials. It makes it extremely difficult as the statisticians said to discount placebo effect or to adjust for placebo. It's very hard.

MS. WOOD: If you believe that additional data are necessary to demonstrate assurance of safety and effectiveness of the Cardima ablation system, please address the following questions. Please clarify if additional analyses on the current data set may be performed to provide adequate information to support safety and effectiveness.

DR. WALDO: I have one thought on that. It's possible that if they go over that data, there are some patients who Cardima indicates had no episodes at all of atrial fibrillation. It would be really nice to see how many of these patients were reliable in performing the transtelephonic monitors. Maybe they'll get lucky and find that they all did for instance. That would be useful. I didn't see any of those data sorted out that way and to me that might be some degree of important support to their contentions.

DR. MAISEL: In addition there are still a number of patients who apparently were enrolled in the study and have not reached the sixth month clinical endpoint. I don't know what the status of those patients is. Certainly an effort could be make to get full data on all of those patients.

MR. MORTON: Dr. Laskey.

CHAIRMAN LASKEY: Yes.

MR. MORTON: Could I follow on a question that you had and it had to do with looking at this PMA as it was presented to the panel as a system. We have spent the afternoon talking about those data especially with the catheter in there. If we look at only the REVELATION_ Tx microcatheter, do we have adequate data for only that device?

CHAIRMAN LASKEY: The hooker with that as I've come to understand is that it's felt to be now unethical. That is to just do the linear ablation without treating the isthmus. That one needs to do the whole thing. Is that right? Are there people who feel strongly about that?

DR. MAISEL: The only other comment I would make is the lack of effectiveness data which is true for no matter which subset of catheters we look at.

DR. TRACY: Yes, I think you can't just put a couple of linear lesions in there and not do an isthmus burn. There's pretty good data that would support the need for putting in an isthmus burn. But I think your question is not just the linear catheter but the system. Am I interpreting it correctly? Are there enough data on the patients who had the system used?

But I think it's only 24 percent of the patients that have the system used or some percent. I doubt you'd have statistical significance with such a small percentage of the patients. It's unlikely that the compliance was better in those that had the total system used versus those that had a mix and match used. I don't think there's probably anything salvageable but it could certainly be looked at.

MS. WOOD: Please comment if the collection of additional data using the current patient selection criteria and outcome measures would be adequate to support safety and effectiveness.

CHAIRMAN LASKEY: I think Bill's comment about trying to get better sixth month follow-up or more complete sixth month follow-up will only help. I don't know to what extent. I can't estimate. That's the only thing we've come up with.

DR. SCHWARTZMAN: What about a more detailed post hoc assessment of EKG AF burden if each of these patients were subjected to a rigorous month of ambulatory monitoring and the company lived and died with the result of that relative to the baseline notwithstanding the limitations of baselines? It sounds to me like the month of baseline was probably not the limitation. It's the follow-up that's the problem. If we took a snapshot month of any patient beyond six months and lived and died with that versus the month pre-, would that be something that people would be interested in?

DR. MAISEL: I think it gets very complicated when the patients are their own control. If you're not collecting data in the same way in both sides of the procedure, then I think it gets very difficult to interpret.

DR. SCHWARTZMAN: We would be talking about daily transtelephonic monitoring for a month comparing that to patient's baseline. The problem I see in that is that they may have evolved so far away from the procedure that the drug changes would really muddy the waters. Not knowing that seems to me that it's not inconceivable that it would be a risk worth taking if I were looking at from the vantage point of the company.

DR. NORMAND: If I could comment, I think that statistically it's somewhat of nightmare in terms of how you're actually going to handle. You've already collected the data at baseline using a certain mechanism. Now you're proposing a different mechanism later on. It's only going to be on a smaller subset of patients I think we're talking about now if we are talking about the system versus not the system.

It may be interesting but I think statistically given it is an observational study, I actually think having the patients serve as their own control is a good idea. I wouldn't take that away but because they are using the patients as their own control that means the mechanism by which we obtain the data and look at that difference has to be done in the same way. So I don't think there's an easy fix that way to use the baseline data which we have to use because the patients have to serve as their own control.

DR. SCHWARTZMAN: So your problem with the baseline data relative to that proposal of 30 days of daily irregardless of symptoms, ambulatory monitoring would be what? That the baseline data was not as ?-

DR. NORMAND: The frequency with which you're collecting the data at baseline is different if I'm understanding you correctly than what you're going to do in the post-period. Moreover the mechanism by which you're going to collect the data in the post sounds like it's going to be different. It's already fixed. It's been done in the pre-period. How that information was collected was done in a certain way. So it's a different mechanism. I think you adequately characterized it as ad hoc.

DR. SCHWARTZMAN: I respect the statistical mischief here but frankly these are patients who had a lot of atrial arrhythmia burden as a practicing atrial fibrillationologist. If a large majority of them had 30 days of minimal burden, I would respect that as a clinical endpoint.

DR. NORMAND: But I think you want to be able to justify it scientifically. We'll have this fight. So I disagree I guess. I agree with your point and I just think it's going to be very challenging to justify it.

DR. MAISEL: In potentially satisfying both of those things if we are going to go back and monitor patients again would be to monitor them the way it was designed for them to be monitored and just pick a different time. So instead of at six months do it at a year out or eighteen months or whatever we decide and then monitor them according to the protocol.

DR. SCHWARTZMAN: Right, but whether it's a fixed time or anytime after six months, I don't think that it's any more relevant to get them all to a year. I think it's clinically relevant but in terms of answering the issue of effectiveness, I don't think it's any more relevant to wait until they all get to a certain point as opposed to saying again respecting the statistics any time after six months just stop them in their tracks, do a month of ambulatory monitoring daily and live or die by that data.

This is not a subtle population in that regard. Unless if they have a very impressive low burden score irrespective of how they feel assuming that the changes in drug therapy haven't muddied the water, that to me would be convincing.

DR. WHITE: David, the problem though is that even if you improve the assessment mechanism, the fundamental methodology flaw of the deviation from the protocol will make it very difficult to then take your information and be happy with what's been done with this catheter. So even data assessment improvement doesn't fix the protocol deviations.

DR. SCHWARTZMAN: This comes back to this isthmus lesion. I have to say I side with the group that looks at an isthmus lesion as an isthmus lesion.

DR. WHITE: That ought to have been clear when the protocol was designed. I think it's hard to do it post hoc. You may be correct but why didn't you know that when you wrote the protocol?

DR. SCHWARTZMAN: Yes.

DR. TRACY: You have two fundamental problems. The isthmus problem which we should just deal with separately but in terms of if there is something salvageable you have 36 percent of the patients who were compliant and made the transmissions as stated by the protocol.

As I see it, there are two solutions to what you can do to try to meet the prespecified endpoints. You can either accept as Bill is suggesting taking month 11 on the patients who were not compliant and make that be their compliant month. That's one option. The second option would be to go and recruit an additional group of patients to bring it up to statistically significant group of patients. So there are two fundamental ways of going about this.

The question for the group I guess is would a month of appropriate monitoring or an additional two months or three months of appropriate monitoring in the cohort that's in this study satisfy this group in terms of an effectiveness endpoint? Or do we think that we begin to run into the ambiguities of the disease progression at some point?

DR. WALDO: Can I make a comment? Dr. Schwartzman I think earlier commented that he was concerned that six months wasn't enough long term follow-up.

DR. SCHWARTZMAN: Yes, I think if anything it puts the company at a disadvantage because in my own experience and what I anticipate from ablation in general is that it has a time-based failure that is progressive. I think that's in no small part because of the issues regarding aging of the atrium plus the co-morbidities effect on the atrium such as hypertension.

But on the other hand to set it at a year for example if 30 or 40 or 50 percent of their patients are months away from a year, I don't think is necessarily fair which is why I'm saying notwithstanding the statistics and the protocol violations related to the isthmus lesion, I don't think it would be undermined clinically to create a snapshot wherever those patients are as long as they are at six months.

DR. GILLIAM: We don't really have a clear understanding of exactly what the burden was pre-procedure. We just have essentially three transmissions done over a 30 day period except for five people who were recycled, the very fact that they could get nine over three months whereas the other month they couldn't get three.

It's just that to me this getting three transmissions with atrial fibrillation as an entry criteria. Then if in six months, 11 months, 12 months or two months period after procedure show that we don't get any transmissions doesn't really convince me. Let's say that a person has a relatively persistent atrial fibrillation episode that lasts two to three days. They send in three transmissions in the study. They either get cardioverted or they spontaneously go out of it and for the next six months they have nothing.

That could be explained that it may be has nothing to do with that procedure or that their medication goes up. You know as well as I do that you could stop amiodarone today and amiodarone is not going to get out of your system in the foreseeable future. If you add another medicine, you just in effect even though you've technically given them less of an accelerated dose of medicine by our standards if he's not taking amiodarone anymore but it's still there at least for a six month period.

My problems with it fundamentally starts with the initial data collection and understanding just how much of an atrial fibrillation burden existed and whether we just saw a snapshot one month later, would that be sufficient to convince me? I'm not sure that is now.

DR. SCHWARTZMAN: I agree with your comments on fixing AF burden whether you define it as number of episodes or total duration pre-operatively. That being said having been involved with patients that are like this, these are very symptomatic and I would call high burden patients to be able to get the number of episodes they got within a month assuming that they were instructed to identify individual episodes and I have no reason to disbelieve this. That's the type of patient they were. I would say that particularly if you put stringent criteria on what we would define effectiveness in a follow-up 30 days of daily EKG monitoring that if they met that effectiveness number then that would be convincing to me as being effective burden reduction.

CHAIRMAN LASKEY: Right. That goes back to the pre/post. I think it would be hard to sell. While clinically we all understand that, still there has to be a pre/post comparison and it's not apples and oranges. It's apples on the front end and oranges on the back end. That's the problem. The horns of the dilemma.

MR. MORTON: Just a quick comment to remind the panel that this sponsor actually conducted the study in compliance with the earlier panel recommendations. Out of fairness, I know you are trying to work within so that they can salvage this.

MS. WOOD: Please comment if a new perspective trial is needed to provide adequate information to support safety and effectiveness.

CHAIRMAN LASKEY: Well, you know that's often the easiest thing for the panel to do. I'm not sure we suggested that up here today. We're still grappling with either how to salvage or how to dig deeper to get some useful information here. I'm not sure anyone is prepared to recommend. What do you think, Cindy?

DR. TRACY: I would like to know back to sponsor's slide 71 baseline symptomatic AF episodes per month mean is 10.1 plus or minus 8.9. There are at least some salvageable data here in terms of people. We have a good sense of what the atrial fibrillation burden was. I think we should be very clear that there is a group that we can make a comparative.

I would be willing to accept an arbitrary period of time really rein these people in and the investigators in to get the data even if it's month nine to 13 or something like that on the cohort that's present. We can see how much can be obtained that way and/or get an additional group of people.

I agree that this was the directive these people were given at the start of the protocol. Here's how you will do it. Here's how you'll assess this. So for us to recommend a change in how we assess things at this point is not fair. We have to work with the guidelines and rules that were set up ahead of time. I'm not talking about the flutter burn. I'm talking about just the guidelines in terms of symptomatic assessment. I think we have to stick with what that was at that time that the protocol was created and try to salvage as much as we can out of that.

CHAIRMAN LASKEY: That being said, it's very hard to account for regression to the mean, for placebo effect, for everything else that the statisticians pointed out are limitations perhaps unavoidable in this kind of study. So yes, we're saddled with this but I don't think we have to like it.

DR. MAISEL: I would also say that demonstration of effectiveness if new data were achieved, in my mind there are still a number of outstanding issues. I don't think we should be sending a message that if effectiveness is demonstrated then we're comfortable with this. At least I'm not comfortable with some of the other issues.

DR. WHITE: No, I think that's why we have to suggest that a new perspective trial is necessary. I think that needs to be said.

DR. ZUCKERMAN: Dr. Tracy, you've asked the Agency and the sponsor to live within the rules and to understand the history of this application. I'm looking at FDA slide no. 9 which was the pivotal Phase III trial design which was 80 patients using a defined system to look at results. That's what the Agency suggested and this is a hypothetical based on what you've learned today, the Agency might still suggest that type of approach given the other approach suggested which was to go back and to try to recall information from perhaps what may be problematic data sets may be problematic. Do you have a problem with this approach as stated on slide 9?

DR. TRACY: Slide which?

DR. ZUCKERMAN: Slide 9 page 3 of the FDA presentation. It's the May 2000 protocol description.

DR. TRACY: If I'm looking at the same slide you're looking at, we're slipping into the isthmus question and using other catheters. Sample size N equals 80, Phase III. Am I looking at the right slide?

DR. ZUCKERMAN: Yes, I think there was an intent with 80 patients to get a user-defined system and look at what the results are. This was considered the Phase III pivotal protocol. That's what was requested.

You've helped us perhaps today in 1) you've given us better performance estimates for what safety should be: 2) you've stated that these present safety estimates look okay; and 3) you've also indicated that a certain effectiveness estimate for reduction atrial fibrillation burden of around 40 percent is acceptable from a clinical perspective for you. If the sponsor and the Agency chose to look at a new data set, would that be appropriate if one were to summarize your comments today?

DR. TRACY: Yes, I think that would be appropriate. I don't know if you want to grapple with the isthmus question right now or not but if you do, my thoughts on that would be that you can't limit people to using a catheter that they don't like. If the standard ablation catheter that was part of this protocol is one that I've never felt or touched, it may not be a very good catheter. What you want is a procedural success.

I would argue that you should open the door to permitting other catheters. If you want to specify the single approved ablation catheter for flutter as being the alternative catheter, that's fine. But I don't think you should tie people's hands to using something they don't like.

DR. ZUCKERMAN: Right, I think we can write a protocol that's acceptable from a clinical perspective. The other question I have is would effectiveness in a new protocol be acceptable if it's just a percent reduction in symptomatic atrial fibrillation. Are we looking also for a certain number of true cures?

CHAIRMAN LASKEY: Percent reduction compared to what?

DR. ZUCKERMAN: Baseline. The primary effectiveness endpoint in this trial was for looking at a percent reduction and thought to be clinically useful.

CHAIRMAN LASKEY: You don't have that information in this data set on the front end.

DR. ZUCKERMAN: Right. But we're referring to question 8C if in a hypothetical context we're designing a new trial. Can you help the Agency and the sponsor answer this question? Is a certain percent reduction as suggested by Dr. Tracy of perhaps around 40 percent clinically useful or should there be a certain percent reduction and a certain number of true cures?

DR. TRACY: We don't have a standard to measure the cures with. The best we have is the Maze procedure or now the pulmonary vein procedure which as we've heard there is very little randomized control information on that. So a percent reduction from my perspective would be adequate.

DR. MAISEL: A percent reduction in symptomatic atrial fibrillation is clinically relevant. If there's a reduction in symptomatic atrial fibrillation I might also expect to see a reduction in the burden of atrial fibrillation. In my mind, the total burden of atrial fibrillation is less subject to bias if you were doing Holter monitoring for example. While I do think the symptomatic endpoint is the most clinically relevant, it would be nice to see another endpoint with that, a surrogate endpoint, to see that it also is reduced so that the bias that might be inherent in post-procedure monitoring might be answered by that question.

DR. WALDO: You know I'm even a little troubled by the concept of 40 percent. Let me tell you why. One of the things I wondered about in reviewing all this was the fact that if I had 10 episodes of atrial fibrillation a month and I reduce them to five I guess I would be happy that I had only five. But I wouldn't be happy that I had the five because I would be so symptomatic.

I think there's a question here that we really haven't addressed. You can get a number and a decreased burden but how that translates into something meaningful is much more important. So the fact that some of these patients seem to have a cure was a big surprise to the investigators and certainly to me. That would be nice if some patients had a cure.

That's where the quality of life issue has really become important. If a 50 percent difference really makes a difference in a patient's life, I'd like to know that. I'm not sure I learned that from this study that that kind of difference if I go from 10 to five is that really a whole lot better. With five terrible episodes a month, I would think I'm still suffering. We really didn't address that too much and that has some impact to our considerations.

DR. TRACY: It's a subjective disease though. It's ultimately the patient who has to choose how far they want to go in terms of achieving palliation versus cure. So if the patient is happy with a 50 percent or a 75 percent reduction, I'm not going to push them any further than that.

The quality of life issues are very difficult because there isn't a placebo group. There isn't a better way of doing this. It would be relevant to have if you have cure you want to know that you have a 100 percent reduction in the episodes. I wouldn't throw that data away but I'm not sure that I would expect that.

Even when you are palliating patients with medication you work hard to indicate to them that you may break through at some point and we'll deal with it at that time. However we're necessarily going to change the antiarrhythmics just because you have a breakthrough at six months that we have to cardiovert you for. So it's a very complex disease. The endpoints that we use here are different from ABNRT or WW or something else.

DR. GILLIAM: I agree with you fully. It's a symptomatic illness. If you can make someone feel well, that's a big part of what we would be very happy with. But if we're going to approve an invasive procedure to go in, we need to document that what we're doing is actually doing something very positive and what we're seeing is not only a placebo effect.

It would be great if we actually had some measure that showed that there was some decrease in atrial fibrillation burden and that is actually what's causing them to feel better. That we would be doing something positive. That's why I would suggest that not a cure perhaps but certainly some objective outside measure beyond the patient that we are doing something that's good for them to decrease their symptomatic episodes.

CHAIRMAN LASKEY: Such as repeat hospitalization. Al's point is well taken that even a 70 percent reduction leaving someone with three episodes per month requiring hospitalization once a month for cardioversion is not going to improve things. So maybe we ought to tack on a hard surrogate endpoint if you are going to rethink this.

DR. GILLIAM: But I would imagine that a lot of these people don't get hospitalized.

CHAIRMAN LASKEY: No, but they come in to get cardioverted. They may not get hospitalized. They come in and get buzzed and go home.

PARTICIPANT: That's the minority.

DR. GILLIAM: That would be easy. If it were that way, it would be fairly easy to look at this group even in retrospect to say that these people who called in three or four times didn't show up to the hospital three or four times. They called in.

I guess the hard thing is that there is clearly no easy way to evaluate AF burden. You can evaluate how many times someone calls you and complains. You know that a person feels better. I can tell you that a person feels better this month as opposed to last month but I don't know if I did anything in that intervening time to make them feel better or if they just were just lucky this month and next month they're going to be calling me again.

That's the hard part of this. It's not an easy issue that we are asking them to show. Show us that you did something with catheter or this system if you will to make the patient better and we have to measure in some way. Certainly there are less complaints if we were to say that. The data clearly showed that because they didn't call in. That doesn't mean that the patients were actually feeling better.

DR. WALDO: Good points have been made and it shows how difficult this area is.

MS. WOOD: Labeling for a new device should indicate which patients are appropriate for treatment, should identify potential device related adverse events and should explain how the device should be used to optimize its risk/benefit profile.

If you recommend device approval, please address the following: Does the indications for use as stated adequately define the patient population and procedural use for which the device will be marketed?

The Cardima Incorporated REVELATION_ Tx microcatheter ablation system is indicated for treatment of atrial fibrillation in patients with drug refractory, paroxysmal atrial fibrillation by mapping, pacing and ablating with a set of continuous linear lesions in the right atrium. Based on the study results, please discuss whether the proposed warnings, precautions and indications are acceptable?

CHAIRMAN LASKEY: Of course we're not quite there yet but the patient population seemed to be well defined. We didn't quibble about the patients in the study. They were all hard core with a high burden of atrial fibrillation. From that standpoint, the labeling is accurate.

DR. MAISEL: I might take exception with the use of the word "continuous" which hasn't really be demonstrated by the data presented today.

CHAIRMAN LASKEY: Then I'm not sure we should have another half hour discussion about the procedural use for which the device will be marketed. We've been talking about the procedural use which implicit and in which is the use of alternative catheter or off-label catheters. We're going to have to modify that language. Any other input?

DR. GILLIAM: I'll just echo again the instructions again to the physicians doing the procedure. There's no clear endpoints of the linear ablations when you are to terminate the procedure. At some point, we would have to come up with some set of instructions for those people doing it.

DR. WALDO: One of the honest points of this is that it's an empiric approach that assumes that a dog is a four-legged animal so all four-legged animals are a dog in a sense. You're going to do the same thing for everyone. It's an inherent limitation that we have to accept but that's one of the major problems in this regard too. We have to live with it at the moment but it's a real problem.

MS. WOOD: Please discuss whether the instructions for use adequately describe how the device should be used.

CHAIRMAN LASKEY: EP folks?

DR. SCHWARTZMAN: You know again as a newcomer to this, I don't know quite how to respond. If someone gives me an approved ablation catheter for SVT, they tell me how to hook it up and tell me I have a generator and a catheter and they don't tell me whether I need a sheath or how to angulate to get whatever focus I'm getting. As you know and as I've commented on, it's doubtful to me that this catheter creates contiguous linear lesions.

That being said, I don't know how far one can expect a company to go in terms of instructing how to be creative with a catheter to get the job done. That's my quandary. I think the catheter ablates. It certainly doesn't create in my mind contiguous linear lesions but if you take that away it ablates. So how much then do you have to stipulate in that regard?

DR. TRACY: I think the instructions as they are stated here are pretty specific. They take you from point A to point B. Tell you how to hook it up. Tell you how to flush it out. Tell you how to deliver energy. What degree. What number of seconds, etc. The only thing that perhaps is missing is how do you know when you are done. At some point, we might have a better way of adding that information. But just as it's stated here, I would be able to follow this.

DR. ZUCKERMAN: You're referring to roughly pages four through seven of the labeling, correct?

DR. TRACY: Yes, correct.

CHAIRMAN LASKEY: Does the FDA have any additional comments, questions before the vote?

DR. ZUCKERMAN: No, we don't.

CHAIRMAN LASKEY: Or the sponsor.

MS. BALDWIN: Marianne Baldwin again from Cardima. I would first of all like to thank you for taking your time today to listen to this very complex and difficult discussion. It hasn't been easy and the job for you was almost as difficult as it was for us. I realize that.

We did make every attempt to follow guidelines and work collaboratively with the FDA. In point of fact, times have changed and everyone has a different point of view today than they did five years ago. But I do appreciate your taking the time to listen to this. Thank you.

CHAIRMAN LASKEY: Thank you. Mr. Morton.

MR. MORTON: I'd like to thank the sponsor and the investigators for a very good presentation. Dr. Ewing, thank you for a good review. The rest of the FDA reviewers, very good.

I would like to bring up the point that earlier there was a comment made about a compliance rate from another study that the FDA presented. I know that we hold sponsors to only discussing information that is in their PMA. I would urge the panel also to limit that information.

CHAIRMAN LASKEY: That's a point well taken. Thank you. Dr. Hughes,

DR. HUGHES: Thank you. Thank you for the opportunity to provide a few comments. I too want to offer my appreciation to the FDA staff and the sponsors as well as voting members for delving into this very challenging issue.

The items have been brought up very well for debate. Again as I said this is very challenging kind of decision to be made. I feel very confident that the appropriate decisions will be made by the panel and the FDA in their effort to review this for the benefit of the consumer. Thank you.

CHAIRMAN LASKEY: Thank you. Before we move to the vote, I'd like to have another go at the open public hearing. Is there anyone who wishes to come forth and address the panel? If not, close the public hearing and have Ms. Wood read the voting options.

MS. WOOD: The Medical Device Amendments to the Federal Food, Drug and Cosmetic Act (Act), as amended by the Safe Medical Devices Act of 1990, allows the Food and Drug Administration to obtain a recommendation from an expert advisory panel on designated medical device premarket approval applications (PMAs) that are filed with the Agency. The PMA must stand on its own merits and your recommendation must be supported by safety and effectiveness data in the application or by applicable publicly available information. Safety is defined in the Act as reasonable assurance, based on valid scientific evidence that the probable benefits to health (under conditions on intended use) outweigh any probable risks. Effectiveness is defined as reasonable assurance that, in a significant portion of the population, the use of the device for its intended uses and conditions of use (when labeled) will provide clinically significant results.

Your recommendation options for the vote are as follows:

1. APPROVAL - If there are no conditions attached.

2. APPROVABLE with conditions - The panel may recommend that the PMA be found approvable subject to specified conditions, such as physician or patient education, labeling changes, or a further analysis of existing data. Prior to voting, all of the conditions should be discussed by the panel.

3. NOT APPROVABLE - The panel may recommend that the PMA is not approvable if the data DO NOT provide a reasonable assurance that the device is safe; or if a reasonable assurance HAS NOT been given that the device is effective; under the conditions of use prescribed, recommended or suggested in the proposed labeling.

Following the voting, the Chair will ask each panel member to present a brief statement outlining the reasons for the vote.

CHAIRMAN LASKEY: I'd like a motion on the PMA. Anyone? Dr. Gilliam.

DR. GILLIAM: At this time, I would move the PMA application be voted not approvable at this time as at this point a reasonable assurance has not been given that the device as used and directed is effective.

CHAIRMAN LASKEY: Is there a second?

DR. MAISEL: Second.

CHAIRMAN LASKEY: It has been moved and seconded that the PMA as Geretta just stated be denied approval. Is there some further discussion?

(No response.)

CHAIRMAN LASKEY: Then I think we should vote on the motion on the table starting with Dr. Tracy.

DR. TRACY: I agree. Do you want my reasons now or are you just taking the vote?

CHAIRMAN LASKEY: Do we do this by hands, Geretta?

MS. WOOD: You can go around the table and take the vote.

CHAIRMAN LASKEY: So we'll do a voice vote.

DR. TRACY: I agree that it is not approvable.

DR. MAISEL: William Maisel. I agree that it is not approvable.

DR. WHITE: I agree not approvable.

DR. GILLIAM: I agree not approvable.

DR. NORMAND: I also agree it's not approvable.

DR. SCHWARTZMAN: Agree not approvable.

DR. ZUCKERMAN: Dr. Waldo, are you there?

DR. WALDO: I am. Not approvable.

CHAIRMAN LASKEY: So that's seven for the motion. Let's just go around the table quickly and state your name please and the reasons for voting as you did. Cindy.

DR. TRACY: Cindy Tracy. In a way it pains me to vote this way because I think this thing will have a role in the armamentarium for treatment of patients with atrial fibrillation but I think that there's too much ambiguity about the effectiveness endpoint. I'm hopeful that perhaps some of the data can be salvaged and not make it too burdensome to go forward.

DR. MAISEL: William Maisel. I also agree. I would like to see a linear ablation catheter on the market and hopefully this application can be salvaged but right now, I think there are too many questions regarding the effectiveness of the device.

DR. WHITE: My name is Chris White. I think that the methodology is the reason why I was uncomfortable with the effectiveness. I think the assessment of endpoints was inconsistent.

DR. GILLIAM: Roosevelt Gilliam. Likewise, I was concerned about the effectiveness endpoints. This is a PMA we're asked to for a new indication for a specific very common problematic clinical entity in which we were evaluating a new catheter, the procedure itself as well as its use in RF generators that will be varied. I think we needed to be ensured that there was adequate efficacy demonstrated in this study. The study did not do that.

DR. NORMAND: Sharon-Lise Normand. My main concerns had to do with the definition of the primary endpoint, the ascertainment of the data for the primary endpoint and the data analysis of the data for the primary endpoint.

DR. SCHWARTZMAN: David Schwartzman. The two issues I had were 1) the lack of evidence that a histological endpoint consistent with the concept had been met and 2) lack of clarity as to the electrocardiographic burden reduction.

DR. ZUCKERMAN: Dr. Waldo.

DR. WALDO: Yes, I think this is very important because it was a new direction in catheter treatment of atrial fibrillation. Sadly the rigor was missing that we would have liked to have seen because this is potentially very important. We just didn't get enough data to permit us to vote favorably. There is no question in my mind that this is of interest and potentially important but we need to be shown better data before we can vote approval.

DR. SCHWARTZMAN: Am I allowed to ask a question regarding the future?

CHAIRMAN LASKEY: Yes, let me just ask Mike and Dr. Hughes for one final comment and then we'll get back to you, Dave. Any final words, Mike?

MR. MORTON: No.

CHAIRMAN LASKEY: Dave.

DR. SCHWARTZMAN: It's always been a point of confusion to me regarding this clinical issue and I think this situation is adequate evidence of that. Given the fact that as Dr. Waldo mentioned that we are in an era of empiricism, it may put an overly onerous burden on the manufacturer technology to achieve a clinical endpoint. I wonder that an electrophysiologic endpoint when one could be had would not be adequate as a approvable endpoint.

With respect to this, the initial example was the isthmus lesion which is characterizable. It's much more difficult than we are led to believe but it is characterizable. In the future whether it's in the right or left atrium, these things will be a characterizable siting in the left pulmonary vein isolation.

I wonder because we're in near of empiricism and because what will come to fruition is an increasingly understanding as to substrate so by nature if we do studies now, subsets of patients will benefit and other subsets won't but we can't separate right now. Would an electrophysiologic endpoint alone ever be adequate to achieve device approval?

CHAIRMAN LASKEY: You mean as a surrogate for a clinical or just in and of itself?

DR. SCHWARTZMAN: As a surrogate for a clinical endpoint. Safety obviously needs to be met but safety combined with an electrophysiologic success. Would that effort be adequate?

DR. TRACY: The real problem though is defining even what that electrophysiologic parameter would be. We don't know if you defined that there was block across the linear lesion if that's why it would work clinically. Maybe it's because you are cooking the nerves. You could certainly find something that you could look at that you could concretely measure the same way you can concretely measure a flutter burn.

I'm not sure that would have any relevance clinically. I don't think I would be willing to accept that as a surrogate for effectiveness. Some clear definition of effectiveness burden reduction something has to be a part of what the endpoint is.

DR. ZUCKERMAN: Right, in general it usually works the other way around. Our definition of effectiveness does include the requirement for clinical utility. However, after the actual carrying out of a number of studies where we get reliable good data where we can pick an appropriate surrogate that reliably predicts clinical utility, then we can consider racheting back requirements. I think we need to understand the definition of device effectiveness.

CHAIRMAN LASKEY: At this table or in this room, we're looking at the boundary between empiricism and metrics and that's what we grapple with every time. That's where this is. It's a moving target but still to not abandon the metrics aspect of what we do here. Thank you, sponsor and colleagues on the panel. This concludes the report. Recommendations of the panel on PMA 020039 from Cardima Incorporated for the REVELATION_ Tx microcatheter with NavAblator^TM RF system for treatment of patients with drug refractory, paroxysmal atrial fibrillation. Again that you all for a hard day's work. Off the record.

(Whereupon, the above-entitled matter was concluded at 4:51 p.m.)