JUNE 9, 2003


      The Panel met at 2:30 p.m. in Salons A, B and C of the Hilton DC North, 620 Perry Parkway, Gaithersburg, Maryland, Dr. Mary Jo O?Sullivan, Chair, presiding.




MACHELLE ALLEN, M.D.                      

CAROL L. BROWN, M.D.                           


GARY S. EGLINTON, M.D.                         

EVELYN R. HAYES, Ph.D.                    

JAY D. IAMS, M.D.           



MARY LOU MOONEY, R.A.C.                   






DEBORAH A. WING, M.D.       


JOYCE WHANG, Ph.D.          








NANCY BROGDON                                   Director, Div. of Reproductive, Abdominal and Radiological Devices

COLIN POLLARD                                   Chief, Obstetrics and Gynecology Devices Branch


DANICA MARINAC-DABIC, M.D., M.M. Sc.Office of Surveillance and Biometrics



CATHERINE Y. SPONG, M.D.                   NICHD, NIH

DAVID SWEDLOW, M.D.                             Perinatal Consultant, Nellcor






























      Dr. Colin Pollard........................ 11



      Dr. Danica Marinac-Dabic................. 15

      Dr. Catherine Spong...................... 31

      Dr. David Swedlow........................ 35


                                         2:49 p.m.

            CHAIR O'SULLIVAN:  Okay.  Let's call this open session of the 67th meeting of the FDA OB/GYN Devices Panel to order.

            All attendees will please sign in on the signout sheets outside.  Any and all audience comments must be recognized by the Chair.  Would you please use the mikes either at the podium or at the table for any speaking.  And if you do speak, please give us a full statement of your conflict of interest, including your travel, per diem and relationship with companies.

            For those of you who have not met the panel, we will have the panel introduce themselves as we go around the room.  Start at that end.  Who you are and where you're from.

            MS. MOONEY:  Mary Lou Mooney.  I'm the Vice President of Clinical Regulatory and Quality for SenoRx, and I'm the industry rep to the panel.                            MS. LUCKNER: Kleia Luckner.  I'm a women's health administrator from Toledo, Ohio and I'm the consumer representative.

            DR. LARNTZ:  Kinley Larntz.  I'm Professor Emeritus of statistics at the University of Minnesota. I also work as the independent consultant in statistics.

            DR. WING:  I'm Deborah Wing.  I'm an associate professor of obstetrics and gynecology at the University of Southern California in Los Angeles.

            DR. NEUMAN:  I'm Mike Neuman.  I am professor of biomedical engineering at the Memphis Joint Program of Biomedical Engineering, Memphis, Tennessee.

            DR. SHARTS-HOPKO:  I'm Nancy Sharts-Hopko, a professor in the College of Nursing in the area of women's health nursing at Villanova University in Philadelphia.

            DR. BROWN:  I'm Carol Brown.  I'm a member of the panel.  I'm an assistant professor obstetrics and gynecology at the Weill-Cornell Medical School and a gynecologic oncologist on the staff at Memorial Sloan-Kettering Cancer Center.

            EXECUTIVE SECRETARY WHANG:  I'm Joyce Whang.  I'm a reviewer in the OB/GYN Devices Branch and the Executive Secretary for this panel.

            CHAIR O'SULLIVAN:  I'm Mary Jo O'Sullivan, an obstetrician, gynecologist from the University of Miami.  And currently chair.

            DR. WEEKS:  I'm Jonathan Weeks.  I'm a maternal fetal medicine doctor from Louisville, Kentucky.

            DR. HAYES:  I'm Evelyn Hayes, professor of nursing and family nurse practitioner at the University of Delaware.

            DR. CODDINGTON:  I'm Charles Coddington, professor of OB/GYN from the University of Colorado and Director of OB/GYN at Denver Health Medical Center in Denver, Colorado.

            DR. EGLINTON:  I'm Gary Eglinton, associate professor of OB/GYN maternal fetal medicine, Weill-Cornell and Chairman of OB/GYN at New York Hospital, Queens.

            DR. IAMS:  I'm Jay Iams, I'm professor of OB/GYN at the Ohio State University in Columbus, and I'm a member of the panel.

            DR. RAMIN:  I'm Susan Ramin, I'm at the University of Texas Houston Medical School, and I'm the Director of the Division of Maternal Fetal Medicine.

            DR. ALLEN:  I'm Machelle Allen, assistant professor, OB/GYN at NYU and Associate Medical Director Bellevue Hospital in New York.

            DR. NOLLER:  My name is Ken Noller. I'm a professor and chair at the Department of Obstetrics and Gynecology in Tufts University in Boston. I practice general OB/GYN and a panel member.

            MS. BROGDON:  I'm Nancy Brogdon. I'm the Director of the Division of Reproductive Abdominal and Radiological Devices at FDA.

            CHAIR O'SULLIVAN:  And Dr. Wolfson, would you introduce yourself?

            DR. WOLFSON:  Yes.  I'm Bob Wolfson. I'm a maternal-fetal medicine specialist in Colorado Springs, Colorado.

            CHAIR O'SULLIVAN:  Okay.  From the press perspective, anybody who wants press information will have to contact Colin Pollard, and he has his hand raised.

            And, of course, this is a rather professional group so there should be no outbursts from the audience.  There's no reason to, anyway.


            EXECUTIVE SECRETARY WHANG:  Yes.  We have two remaining dates scheduled for meetings for this panel this calendar year.  They are September 8th and 9th and November 3rd and 4th.

            We have four new voting members on the panel today; Dr. Hayes, Dr. Miller, Dr. Weeks and Dr. Noller who were recently approved as voting members of the panel. And we have two new panel consultants; Dr. Coddington and Dr. Whang.

            I'll now read the conflict of interest statement.

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

            To determine if any conflict existed, the Agency reviewed the submitted agenda and all financial interests reported by the Committee participants. The Conflict of Interest statutes prohibit special Government employees from participating in matters that could affect their or their employers' financial interests. However, the Agency has determined that 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.

            We would like to note for the record that the Agency took into consideration certain matters regarding Drs. Michael Neuman and Hugh Miller.  Dr. Neuman reported an interest in a firm at issue, but in matters that are unrelated to today's agenda.  The Agency has determined, therefore, that he may participate fully in the panel's deliberations today.

            Dr. Miller reported past interests with a firm at issue. Because the involvement is not directly related to the agenda item for the first session of today's meeting and he has no ongoing relationship, the Agency has determined that he may participate fully in those deliberations.  However, due to conflict of interest prohibitions, Dr. Miller has been excluded from participating in the second session today. 

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

            With respect to all 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 on.

            The transcripts for today are being handled by Neal R. Gross & Company. And there's a flyer at the registration desk if you want more information.

            And if there are any presenters to the panel who have not already provided FDA with hard copy of their remarks, including overheads, they should do so. They can provide them to Dr. Noel Del Mundo. Noel, please stand up.  And he will collect them from you at the podium.

            CHAIR O'SULLIVAN:  Colin Pollard now will give us some information regarding some updates that are associated with the branch.

            DR. POLLARD:  Thank you, Dr. O'Sullivan.

            First of all, I just wanted to welcome the entire panel today and tomorrow.  We very much appreciate you being here.  We really understand the sacrifices all of you have gone to to be here to help us, and we really need your input. So I just want you to know we appreciate that.

            I'd also like to introduce to you some new people in our group.  Noel Del Mundo.  Noel, would you stand up?

            Noel is a gynecologist who joined us in January from the Indian Health Service in Phoenix.  And we're just incredibly delighted to have him working within the branch and helping us out in all kinds of reviews.  And you individually may be in touch with him on a number of things that he's working on.

            And secondly, I'd like to introduce Paul Ciminera.  Paul is with the Army and he's doing a practicum here in the branch, kind of shadowing the situation here and actually helping us out on a number of things.

            The next thing I would like to do is just update you on a variety of activities or things that we've gotten done since our last panel meeting, which was almost a year ago.

            First off, we have approved 3 original PMAs in that time.  In November we approved the Essure System for Permanent Birth Control. This is a PMA that went before the panel at its last meeting in July. And that approval issued with a couple of conditions associated with post-approval studies; one for a continued follow up of all the patients in both the phase two and pivotal study cohorts up to five years as well as a smaller study to look at bilateral placement rates in the general population.

            We approved the Philips field pulse oximeter in January of this year. Like the PMA for the Corometrics monitor, this device and the PMA is linked to the Nellcor OxiFirst, it essentially embeds the same fetal pulse oximeter technology and essentially uses a lot of the same clinical validation study work.

            And finally, in March we approved the FemCap device. This is a vaginal barrier contraceptive device.  And sort of falls in line with FDA's policy of trying to get more contraceptives out in the market and available to women.

            Earlier this year using a relatively new regulatory tool that FDA has for what we call early collaboration, FDA entered into an agreement with Medispectra for a optical system that would serve as an adjunct to colposcopy for detection and localization of diseased cervical tissue.  And that is for the pivotal study and we would probably see the results of that in the form of a PMA probably sometime early next year, or something like that.  I'm not sure how long that will be.

            In October of last year we approved a 510(k), the first for the new indication of uterine fibroid embolization.  This is for the Embosphere device.  And that now offers a nonsurgical alternative to women with symptomatic uterine fibroids, and the company will continue to follow the women in that study for at least three years.

            We also did a de novo reclassification of the BreastView device, which is a breast lesion documentation system to be used in conjunction with a clinical breast examine.

            And the last thing I wanted to mention is that we oversaw the voluntary withdrawal of Intergel, an adhesion barrier product that's applied immediately after pelvic surgery before surgical closure.

            And finally, with the help of our Office of Surveillance and Biometrics we have planned an update for you on the status of the post-approval studies for the OxiFirst Intrapartum Fetal Pulse Oximeter.  You will remember that when FDA approved this monitor in May of 2000, hard to believe it's been three years already, there were several issues around which questions remained.  We planned to update you on those post-approval studies as well as a Dear Doctor letter on this monitor that issue earlier this year in response to a number of adverse events that were reported to the agency and to the sponsor. 

            Before that begins, I believe there are a couple of other administrative matters for you.

            So, thank you very much.  Good luck today and tomorrow.

            CHAIR O'SULLIVAN:  Well, we're beginning our open public hearing. And as far as I can determine, there are no prearranged speakers.  Are there any other public speakers?  If not, we can move on then to OxiFirst post-approval studies and adverse events to be presented by Danica Marinac-Dabic. 


            DR. MARINAC-DABIC:  No problem.

            Good afternoon, ladies and gentlemen, distinguished members of the panel.  My name is Danica Marinac-Dabic.  I work for the epidemiology branch here at the center of the Office of Surveillance and Biometrics here at CDRH.  And the office that I work at is in charge of the post-market surveillance of approved medical devices.

            I would like to thank you for the opportunity to present today an update on the CDRH post-market surveillance activities related to the OxiFirst Fetal Oxygen Saturation Monitoring system.

            As a part of today's presentation, I would first like to provide a brief regulatory history of the OxiFirst device followed by a descriptive analysis of the adverse event reports that we have received here at CDRH so far. And then I will focus on the labeling changing that have occurred since the approval of this device and a "Dear Doctor Letter" that was issued by the company as a form of public health notification  to inform all users of this device across the country of those adverse events that have occurred and on the labeling changes as well as clinical guidelines notifications associated with the use of this device.

            And finally, I will talk about briefly about the status of the post approval-studies with a note that Dr. Cathy Spong from NIH and Dr. David Swedlow from the company will talk in more detail about those studies later this afternoon.

            The OxiFirst was approved in May of year 2000 as an adjunct to fetal heart rate monitoring in the presence of nonreassuring fetal heart rate pattern.  The device was meant to be used after the fetal membranes ruptured in the singleton fetus with vertex presentation and in the gestational age greater than 36 weeks.

            This approval decision was based on the results of the multi-center U.S. randomized clinical trial that showed a decreased Caesarean section rate for the indication of nonreassuring fetal heart rate in the group where fetal pulse oximetry was used in conjunction with fetal heart rate monitoring alone when compared with the group when used only fetal monitoring.

            As you remember the results of this trial did not show and in fact actually a decreased rate in overall Caesarean section rate, mainly because there was an increase Caesarean section rate for the indication of dystocia.

            As some of you may remember when this premarket application was brought before the panel, there were several issues that were raised by several members of the panel and those members felt that these issues need to be addressed in the post-approval studies.  The first of those issue is actually the indications of OxiFirst placement under general use of this device.

            Several panel members at that time have felt that once this device reaches the routine use across the country, there is the possibility that we will see an increased number of Caesarean section rates.

            Also maternal infection rates were one of the other concerns that needed to be corrected in the post-approval phase of this device.

            Another question that remained unanswered at the time of the approval of the device was what is the critical duration of the saturation that is associated with the higher risk for that or neonatal depression or fetal acidosis. In other words, how long the duration of fetal oxygen situation can remain below 30 percent before risk of fetal injury is seen.

            And two remaining issues that needed to be addressed in post-approval studies were adequacy of labor  and also the information on the neonatal outcomes that needed to be collected including cord blood analysis, Apgar scores, etcetera.

            And what was originally planned to be a one post-approval study evolved into three separate studies.  One being general use studies sponsored by the company.  Another one was the study that was conducted by several investigators who wanted to explore the reasons why the dystocia in the original clinical trial, the Caesarean Section rates for the indication of dystocia were increased. And the third one was the FOX trial that was planned to be conducted by the NIH.

            With this I will wrap up the regulatory background and we'll move on to the description and analysis of the adverse events received by the FDA so far.

            As you can see, during  the period from July 2000 to June 2003 we have received a total of 14 adverse events reports. The first event that we received occurred in July 2001, and you can see on this slide the distribution of those events based on the month when the event occurred. The last report that we have received was a report of an event that occurred in November of last year.  Since then we have not received anymore adverse events associated with the use of OxiFirst.

            Three of those reports were deaths. Nine of those were injuries. And then two of those were reports of malfunction without an injury to the fetus.

            All those adverse reports were discussed in depth with the company.  And based on the information that was in the reports and as well as on the information that the company additionally provided to us these categories were made. They actually flagged the clinical outcomes of the cases on which the reports were made.  As you can see, in three of those cases the stillbirth or immediate neonatal death occurred. In three cases there was a significant clinical depression and acidosis present.  In two cases there was a transient acidosis with normal clinical status.  And in five cases there was a transient depression or acidosis post difficult delivery. And one case was an incidental observation unrelated to use of fetal pulse oximetry.

            I'd like to point out here that in 5 of those cases there were clear deviations from the clinical guidelines.

            In order to estimate a true incidence of adverse events associated with the use of this device, I'd like to talk briefly about both the denominator that we were able to use and the numerator. And first I will talk to the denominator.

            Best estimate that the company could provide to us point toward a number of 15,000 monitor patients during the same period where those 14 events occurred.  Highlighted in yellow on this slide you can see two events that occurred during the same time that in our view could possibly effect of the number of the monitored patients, and those are the ACOG Committee Opinion that was published in September of year 2001 when the Committee stated that it cannot endorse the use of this device because it was concerned that it will further increase the cost of the obstetrical care without necessarily improving clinical outcomes when this device is used. 

            And approximately a year after that in October of 2002 the company stopped active marketing of this device and currently they're serving only their existing customers. 

            Again, this is the best denominator that we can come up with when estimating the incidents of the adverse events.

            With the numerator we'll have a little bit more problem, because we have to keep in mind from where these 14 reports came from.  And on this slide I tried to illustrate the adverse events reporting systems that FDA uses.  And to make sure that you understand that these are passive surveillance systems, which means that there is limited information that we get from those reports and we cannot require, in all cases, additional information to be provided to us. 

            This is a surveillance system that collects tens and thousand reports on any device that is approved. And just to illustrate the number of reports, you can see here that in 2002 we have received over 120,000 reports associated with all devices that CDRH regulates.

            You're probably the most familiar with the MDR system under which the manufacturers are required to report all deaths or serious injuries that occurred.  They have to report that to us. And, in fact, 93 percent of the reports that we receive through the system comes from manufacturers.  Approximately 3 percent of reports come from user facilities throughout the  United States.

            On the other hand, MedWatch system is entirely voluntary system and it's responsible for approximately 3 percent of overall reports that we receive.

            The two remaining systems are MedSun and International Vigilance.  MedSun is currently under development. We have 80 hospitals throughout the  United States that participate in this surveillance network. Soon we'll have 200.

            And International Vigilance is responsible for approximately 100 reports per year, and we have 15 countries including Australia that contribute their reports and share their adverse events reports with us.

            With all that said, I'd just like to underscore the major limitation of these surveillance systems is under reporting.  We know that those 14 reports are not what's out there. And I would like to caution in using the simple mathematics and dividing 14 by 15,000.  This is not what the true incidence of the adverse events associated with OxiFirst is.  We almost never use this type of calculations in estimating the real incidence of those reports.

            On other hand, what we do is we use the information that's given to us through this system, and we use that as a signal and enter previous information, try to get more information through the dialogue with users and with manufacturers. And try to come up with a strategy through another regulatory mechanism that we have in our hands.

            And in this case, that strategy was the labeling changes.  We did held several meetings with company, we're in constant communication with them in written and verbal communications.  And as an outcome of this, the sponsor submitted PMA Supplement 10, this was approved in November 2002.  And the main purpose of this was really to provide the clinical modification, guidelines modifications in form of labeling changes.  And this next 4 or 5 slides I will illustrate what those changes are.

            As some of you may remember, the definitions that were included in the original clinical management under the original randomized clinical trial that served as the base of the approval of this device, there were several definitions of nonreassuring and reassuring fetal heart rate. In addition to that, we had the definition of ominous fetal heart rate pattern.  And that original definition included those heart rate patterns that had prolonged deceleration to below 70 beats per minute for longer than 7 minutes.  However, the experience over past two years with this device suggested that this definition needed to be expanded to include some additional patterns that were seen in those cases that we have received with adverse outcomes.  And in addition to that original definition, these two were included in the labeling change; the markedly decreased to absent variability with persistent late decelerations, or markedly decreased to absent variability with severe variable decelerations.  And also in the footnote here I'm going to read it from the slide, that "It is not necessary to wait more than 7 minutes of prolonged deceleration before initiating intervention."  And intervention meaning anything from evaluation of the cause or nonsurgical intervention to preparation for delivery, even when we have the reassuring fetal situation value.

            What will happen in those cases when the signal is not available?  This is the second thing that appeared in the modified clinical guidelines.

            If the signal cannot be obtained, despite the sensor adjustment, this device does not have any adjunctive value.  And in those cases, the fetus should be managed by fetal heart rate and clinical signs alone.  And what's very important, and it's emphasized here, that no inferences regarding fetal status should be made on the basis of earlier FSpO2 value.

            Again, the labeling change, we emphasized the adjunctive role of fetal oxygen situation as an instantaneous measure of fetal oxygenation. It is not a measure of fetal arterial blood pH.  And prior hypoxia may not be detected by subsequent FSpO2 monitoring.

            As a part of the labeling change and as a part of clinical guidelines modification, there were tables of updated fetal heart rate tracings and the classification and also the clinical management guidelines submitted. And here I would like to point out that these guidelines are recommendations only and they are not intended to substitute for clinical judgment. Even though they were tested clinically, they may not apply to every clinical situation.

            In February of this year a "Dear Doctor Letter" was sent by the company to all the clinical users of this device.  And the purpose for this letter was to inform those clinicians on the changes that we were just discussing.  Again, I'm going to go over them one more time.  Four major issues in the "Dear Doctor Letter."  Expansion of the definition of ominous fetal heart rate pattern.  If the signal cannot be obtained, fetus should be managed by fetal heart rate and clinical signs alone. 

            Again, this letter was a reminder that OxiFirst is to be used only as an adjunct to fetal heart rate monitoring.  And as a part of this letter those two tables with the new classifications of the fetal heart rate balance as well as clinical guidelines were sent to the users with a note that these tables are for recommendation only.

            Because the FOX trial is still ongoing and there were some concerns raised when the "Dear Doctor Letter" went out, that possibly the content of the letter as well as the labeling changes might influence the continuation of this clinical trial, the FDA issued the letter and sent the letter to the NIH, NICHD, again emphasizing the four issues from the "Dear Doctor Letter."  And also stating that despite the adverse events that we have received and those labeling changes that occurred, OxiFirst labeling changes do not undermine the results of the pivotal trial and that FDA believes that when used according to newly revised and approved labeling, OxiFirst is safe and effective in reducing Caesarean Section rates for nonreassuring fetal heart rate patterns when compared to intrapartum fetal heart rate monitoring alone.

            In conclusion, I would just like to give a brief status of those three studies; the general use study, dystocia and the FOX trial.

            General use study was designed to be a prospective general use study with end points Caesarean Section rates with OxiFirst compared with the historical control from those three participating hospitals with the estimated sample size of 1750. However, this study is currently on hold, primarily for two reasons. The one, the usage rate of this device is approximately 1 percent of all births in those hospitals as opposed to 30 percent that was planned when the study was designed.

            Another problem was that the Caesarean Section rates were very unstable in those hospitals.  They changed from 4 to 6 percent, which was the trend of increasing the Caesarean Section rates nationally was also mentioned.  Even if the user rate was higher, it would be a problem to compare with the historical control data.

            The dystocia study was prospective nonrandomized study that was designed to fit hypothesis that significantly nonreassuring fetal heart patterns identified patients that are increased risk for dystocia in a setting where fetal well being is defined by normal fetal pulse oximetry.  274 patients in 5 clinical sites participated in the study.  Final report is in preparation and you will hear more from the company on the results.

            The FOX trial is conducted under supervision of NIH Maternal Fetal Medicine Group.  The end points are Caesarean Section rates for all indications.  The safety issues and critical threshold and duration analysis are the things that I discussed previously were among the end points.

            The study is multi-center, randomized trial, estimated sample size of 10,000 deliveries in 15 study sites. And current enrollment is a little bit over 1,000.  The study is expected to have a duration of 2 to 3 years.

            And with this, I would like to turn it over to Dr. Cathy Spong who will continue on the progress on the FOX trial.

            Thank you very much.

            DR. SPONG:  Thank you, distinguished members of the panel, members of the FDA and guests. I'm delighted to update you on the NICHD Maternal-Fetal Medicine unit's network study on fetal Oximetry.

            As many of you know, the Maternal Fetal Medicine Unit Network is comprised of 14 academic sites across the country. The network was formed in 1986 with 7 sites, and there's a competitive recompetition every 5 years. And currently we have 14 centers. The dotes on the map show the locations of those current sites.

            The network currently includes over 100 deliveries per year, and these are the different sites with their deliveries per year. 

            And, of course, whenever you do a Power Point presentation and you move from one computer to the next, the spacing always changes.

            The Fetal Pulse Oximetry trial began as an idea following the SMFM presentation by Dr. Greavey's group. The network was interested in conducting a randomized trial with a large enough sample size to address the concerns regarding both safety and efficacy.

            The protocol development began in May of 2000 and recruitment for the trial began in May of 2002. 

            The research questions that this study will address include does fetal oximetry reduce the risk of Caesarean delivery both overall Caesarean delivery as well as Caesarean delivery for fetal distress and Caesarean delivery for dystocia?  Does the knowledge of the fetal oxygen saturation effect the risk of fetal compromise?  And what is the natural history of oxygen saturation values during labor?

            The overall aim of the study is to determine if fetal pulse oximetry effects overall Caesarean delivery rate.  It is a randomized controlled trial. The eligibility criteria includes singleton gestations, nulliparous women at greater than or equal to 36 weeks of gestation and labor.

            It's an interventional trial with two arms.  In one arm oximeter data is available.  It's called the open arm. And in the second arm the oximeter is also placed and monitored to get adequate values.  However, all of the data is masked to the caregiver.

            The primary outcome is Caesarean delivery. And our sample size is 10,000 women.

            To date we've had over 6,000 patients screened.  Thirty-eight percent were ineligible. And the consent rate is around 30 percent.

            The median percent registration time of the oximeter is nearly 80 percent, it's 78.4 percent.  And we've had 28 protocol violations.

            Baseline data, mean age is 23.2 years for the women.  We have around 50 percent caucasian patients and nearly 30 percent Hispanic or Latino.  The mean years of education is a little over 12 years. And around 6 percent are current smokers.

            The recruitment as of April 2003, so we're up a little higher than this now, was 1250. And, as you can see, we have over 8,000 to go.

            And I'd be happy to answer any questions.

            CHAIR O'SULLIVAN:  I'm sitting here doing math.

            Okay. Our next speaker is Dr. David Swedlow and he's from Nellcor.  Dr. Swedlow?

            DR. SWEDLOW:  Good morning.  Just a minute while I pull up the correct version.

            Good morning. My name is David Swedlow. I'm retired, however I'm currently a paid consultant, clinical and technical consultant for Nellcor Division of TYCO Healthcare. And I've been with the fetal pulse oximetry project since its beginning as a dream in 1987, the actual work starting in 1990 and then continuing through to today.

            Thank you for inviting us to speak or inviting me to speak. And I appreciate the opportunity to bring you up to date on our view of where we are.

            I'd like to cover four subjects.  And then if there's time permitting or interest from the panel, I'm prepared to go into a detailed case history.

            The items are I'd like to bring you up to date on what's currently out there.  The MDR experience, which has already been covered, I think, in more detail.  I'm going to very briefly talk about the labeling changes and focus most of my attention on the status of the post-approval studies.

            This is a slide that shows the current status of -- one month actually. The current status of the installed base of monitors shown in the top curve with closed squares, and in the middle curve in blue diamonds.  The cumulative number of sensors shipped since the beginning back in July of 2000.  I call your attention to a seminal event that occurred, as Danica mentioned back in September of 2001, which is the publication of the ACOG bulletin.  After that, the rate of new introduction or the new installation rate tapered off fairly dramatically. And then in October of 2002 the company decided to cease active selling efforts and active marketing simply because literally it wasn't making enough money to sustain the expense of the full time sale force.  The company continues to take orders. There's a waiting list. And new devices will be shipped out as soon as the new training materials are completely finished and approved.

            In the meantime, however, the sensors that were going at an increasing rate, which is what you would expect from an increasing placement, have stayed very constant. It's quite remarkable that those people who had the instrument in place continued to use it at about the same rate as before the ACOG announcement. And the average number of sensors shipped out on a monthly basis is 1,052 over the past year and a half or so.  And that hasn't really changed.

            There are couple of blips here in April and another one out here in the next April, which represents an order of the NIH FOX trial asking for a large number of cases.

            Those are facts.  The green line or the green shaded area is our best guess.  And that's the problem that Danica was talking about in terms of what is the denominator.  That is how many patients have actually been monitored with the device.  The truth is we don't know.  This green area represents kind of a minimum number of patients and maximum number of patients based on a lag time between the time the device is shipped and actually used.  At a steady state they're equal because there hasn't been any change at all in the shipment rate.  So we know that about a 1,000 sensors a month, ergo a 1,000 patients a month or 2 patients per box is another way of looking at it, are being monitored.  But the actual number of patients, let's say at the end of April, was somewhere between 16,000 at a minimum or 15,000 a month earlier, and 25,000 at a reasonable maximum.  Last month there were another 1,000 sensors shipped. So these numbers are now probably closer to 17,000 at a minimum and 26,000 at a maximum.

            In the bottom row of symbols I've drawn in a sort of an orange closed circle the reported occurrence of the 14 MDRs.  And what you see is that during the early use, if you will, where the number of instruments and clinicians using it was fairly small and the number of patients being monitored was fairly small -- I should point out that the vertical axis on the left is for the number of monitors. The vertical access on the right is for the number of patients.

            There were a fair number of MDRs early on. As the number of patients being monitored, and by inference the experience level of those users expanded, the number of MDRs began to trail off and there haven't been any reports to us or rumors, or any other interesting drum beats out there since about October -- no. I guess it was November of 2002.

            About the same time we started two different processes and decided to do both of them at the same time. One was to introduce the new software, the software upgrade, which was designed to substantially improve the ergonomics of using the device and also the ease with which one could get good sensor placement with audible and visual feedback.  And it was heartening this morning to see Dr. Spong saying that the registrant rate is 78.4 percent.  You may or may not remember, it's kind of a minor detail, in the randomized controlled trial that we presented back in 2000 it was only 64 percent in the registration time.  That is the percentage of the case that numbers are available. 

            All of the NIH sites, by the way, are using the new software.

            And at the same time, we had approached the FDA back in August of 2002 with these MDRs and our concerns about clinicians not really following the guidelines the way we would like them to. And we thought that we needed to reenforce them, update them.  Reenforce, I guess, is the best word. So we started an effort with all of our clinical specialists back in October to not only go out and install the new software, which made it easier to use, but assuming that we would eventually reach agreement with the FDA on the exact wording of all of the guidelines and letters, which we did without any changes, we started training people on the new guidelines back then.  And that continues. And, hopefully, will be complete with all of the customers updated by the end of this calendar year and retrained.

            So, it turns out to be rather difficult to get all of the physicians who use the technology to be in one place at one time. It's really quite difficult.

            As Danica mentioned earlier, there are 14 cases altogether.  This is a slightly different categorization.  It's the same data.  Three patients with stillbirth or perinatal death.  One came out with severe clinical depression and acidosis.  And then the other cases, as she described. There were 5 cases as she pointed out with a very significant deviation from the guidelines. And then there was this incidental observation that's interesting, but had nothing to do with it.

            I won't go through a discussion of the numerator divided by the denominator because of the material that Danica discussed.

            So, when we went back to the FDA in August, we came and said we're concerned.  There are common factors, there are patterns here that we'd like to do something about.  There were a lot of situations in which clinicians ignored periods of no FSpO2 reading. They ignored what we call pulse lost or pulse search.  These are situations in which in some cases the sensor is lifted up off the fetal face and therefore cannot give an accurate measure, but in other cases the pulses are so small and attenuated that the instrument cannot detect an adequate pulse to make a measure.

            Now, in the anesthesia world, losing a pulse in a pulse oximeter is, I won't say a catastrophe, but it's a big deal and we get very excited about it. It's a high alert, high status problem. Quite frankly, I expected that to be the case in obstetrics as well.  I was wrong. I apologize.  I was incorrect on that. It turned out that lots of obstetrics appear to be ignoring a condition of pulse search or pulse lost, no detectable pulses and assuming that if a value was present 2 hours ago but it no longer present, then it's still the same as it was 2 hours ago. That, quite frankly, was just a surprise to me and I'm sorry for it.

            In some cases we saw the sensor placement and usage criteria ignored.  As I'll explain later if you want the details, we had a patient in whom the sensor was put in after there was a true catastrophe.  It was just ignored. Never mind.

            And there was a common fetal heart rate pattern that Danica mentioned, and that was at least in a couple of cases with adverse clinical outcome, a pattern of absent variability and repeated severe deceleration. Some were late, some were variable, but they were all horrible.

            So we went back in August and made the case that we needed to change the guidelines.  We went back and forth several times.  We started in October to upgrade the customers to the new software and coincidentally started training them on the proposed guidelines.  And we finally got that nailed down and Danica mentioned the mailings were finished by mid-February, either mid or third week in February, to everybody who had ever seen an Nellcor oximeter or who had ever ordered a sensor, whether they were a Corometrics user or a Philips HP user, or a -- no matter how they got it. Anyone who had gotten the sensor, we sent the letter to.

            Post-approval studies.  There were many concerns expressed by both the FDA and the panel.  Why were the sensors being replaced, what happens to the C-section rates, what happens to the mom in terms of infection, what's the critical duration, how do we prove that labor is adequate?  This is in reference to the Dystocia issue. And what happens to the kid.

            We, quite frankly, could not design a single study that would answer all of these questions. Some of them were mutually exclusive. So together with the FDA we proposed and eventually settled on splitting these questions into three distinct studies. One to address the question of what happens in the real world when it's used by real people. Another one specifically targeted to the dystocia question of adequacy of labor.  And then the NIH study that Cathy described.

            I'm not going to repeat Cathy's material. They're up to about 1150 patients, 1200 I guess, and they're going quite well.

            The dystocia study.  The critical analysis of the randomized controlled data suggested, at least the possibility, that the fetal heart rate patterns that we had chosen for entry criteria, which we thought at the time would be reasonable criteria for fetal distress, might actually be a marker for a dystocia.  And many people thought that that was possible, more people thought that was impossible, improbable and silly.  But we designed along with our investigators a multi-center study that was specifically designed to test that hypothesis; that is that the more significantly nonreassuring of fetal heart rate pattern you had, the more likely you were to have Caesarean for dystocia given a prospectively defined labor management method and a prospectively defined definition of dystocia.  Two things that we did not have the randomized controlled trial.

            So we categorized -- "we" meaning myself, the company and most importantly the investigators from the 5 sites -- we categorized the fetal heart rate patterns that we had used form the randomized controlled trial and which are still in the guidelines, into two classes.  Class Is were the not so very bad and Class II were the more so very bad.  The idea being that the test essentially being were there more Caesareans done for definitional dystocia in Class II than in Class I given that labor management protocol that included fetal oxygen saturation monitoring.  We were looking for a trend. Worsening of that rate in Class II.

            Dystocia was formerly defined at arrest of active labor for 3  hours or more. There was a lot of argument whether it should be 2 hours, 3 hours or 4 hours.  We finally settled on 3.  The definitions are there.

            Labor management was prospectively defined.  Everybody got EFM and saturation. These were all nullips.  Vaginal examines were supposed to be done every two hours.  Actually, remarkably, they were.  Oxytocin was used according to protocol.  And an IUPC was placed when there was abnormal progress as defined in a previous slide.

            The key results are in table 4 and table 5 of the manuscript, which I'll get to in a minute. And what they showed was that the Caesarean delivery rate, any Caesarean, whether it was for dystocia, for any cause at all was higher in Class II than in Class I, even though these were all within the definition of nonreassuring fetal heart rate. And the frequency or the rate of operative vaginal delivery assisted was also greater in Class II than in Class I.

            The key slide, however, is this one.  And that is we're now looking at only those Caesareans done for the indication of dystocia given the strict definition of dystocia after using the strict management in that strictly defined patients, so it's rather heavily filtered.

            And what the investigators found was that the rate of dystocia performed for formal definition -- wait a minute.  That didn't make sense.

            The rate of Caesarean performed with the formal definition of dystocia was 22 percent in Class II and 8 percent in Class I, and that supports actually the hypothesis and also explains, I believe, and supports the explanation of why we saw an increased number of dystocia in the RCT.

            Just to finish up, there were no significant differences in the usual risk factors and outcomes between the two groups, which is what you'd expect. And the conclusion is that, in fact, they found that significantly worse nonreassuring fetal heart rate does indeed predict for -- is associated with is probably a better more statistically accurate way of saying it -- Caesarean performed for dystocia given all those definitions and management among nullip patients with normally oxygenated fetus in a setting of a standardized labor management. And the investigators and myself believe that that does explain the results that we saw in the U.S. randomized controlled trial.

            Now, that takes us to the general use study, and this is a problem.  The issue of what happens or what would happen when this technology hits the real world with general uncontrolled use was that there was a real concern expressed, among all parties actually, FDA, staff, the panel, the company, everybody, investigators, etcetera was that with widespread uncontrolled routine use of FSpO2  monitoring there was a concern that the C-section rate would actually increase. And this was based on an assumption that it would actually be uncontrolled, widespread and routine.  And the numbers banded about by the company and by the panel, and by everybody else, was roughly 30 or 40 percent of all births.  That was the real issue, the real concern.  So the FDA and the panel asked us to do a post-approval study to look at the question of does the C-section rate go up or does it change at all in general use?

            The study design was prospective, multi-site, general use comparing the C-section rate in an epoch of saturation monitoring to a historical control epoch.  And we choose not to do another randomized controlled trial because by definition the randomized controlled trial cannot be a general use study.  It's impossible.  That's one of the conflicts that I mentioned earlier.

            We estimated, along with the FDA staff, that we were going to target hospitals with a priori historical C-section rate of 25 percent. We agreed that a clinically meaningful change would be about 3 percent either way, so it would go from 25 to 28 or 25 to 22.  With those kinds of changes we estimated we needed about 1800 patients at a 30 percent usage rate, which was on the low end of whatever one expected, and a 75 percent consent rate, which was the experience of the U.S. randomized controlled trial. We thought that somewhere between 3 and 4 studies would take about a year to complete.

            The site criteria included finding a site that was willing and able to provide accurate or variable Caesarean delivery rate for the 12 months prior.  So first we had to find sites with data. Then we had to go in and look at their records and find out whether it was real or not.  And lots of sites fell out, some sites didn't.  But that rate had to be between 20 and 30 percent, preferably higher rather than lower.

            That the site had to promise that their rates were stable, that they didn't fluctuate more than 3 percent per quarter because we wouldn't to avoid the noise situation of wildly fluctuating rates. Not all hospitals had rates by quarter. Some only had by year.  So we didn't want to abandon those general use places that didn't have monthly data.

            That they did at least a 1,000 births a year and that they guaranteed that 75 percent or more of the on-site clinicians would accept training and be willing to use the technology.

            The FDA also imposed additional conditions, and that is that they not be current customers and have no prior experience. Because they didn't want to contaminate it with prior experience.  Sorry. Contaminate the results with prior experience.

            That we had to train them exactly the same way we trained a new customer, so we did.  And that they had a sort of training period or a break-in period, if you will, to gain experience with the technology prior to actually starting the data collection, which it turns out was critically important because none of the three sites came even close to the estimated usage values.

            The first one was Doctor's Medical Center in Modesto, California. That site was started May of 2002. Their historical C-section rate on a basis of 4,000 births was 28 percent.  We were able to record the number of C-sections by month when we started. By the time we started in April of 2002 it was 32 percent, varying between, say, a low of 26 and a high of 33-ish, 32.9. 

            This column here, the second one from the right, is the number of sensors that they used during that month. And the last column on the left is the sensors per birth.  This is essentially the usage rate. And what you can see here is that their overall usage rate was about 1.3 percent, not close to the 30 percent.

            Sarasota Memorial Hospital in Sarasota, Florida was started in August of '02.  We were able to get monthly data for about 4 or 5 months prior to the actual initiation of that site on August 15th. Their prior historical rate varied between 24 and 28 percent. You can imagine it was difficult to find somebody who had data that showed it was not varying. So this is the best we could do there.

            By the time we started getting data, it had already gone up to 32 percent. And their overall utilization rate was about half percent.

            And by the way, this wasn't exactly uncontrolled general use. They didn't have to pay for the sensors. They were provided free. So there was zero economic barrier at all.

            And finally, Abbington Memorial Hospital from outside of Philadelphia, my old stomping grounds, really showed the same thing.  Their initial rate was about 20.5 percent. Before the technology was introduced it it had gone up to about 24 percent. And their overall utilization rate was one percent.

            So we really have a big problem here. The assumptions upon which the general use study was based just are not true.  First of all, the baseline C-section delivery was not stable year over year, which makes use of historical controls problematic, to say the least.  The rates had increased dramatically prior to the introduction of the technology.

            The usage rate of 30 percent was not even close. The actual usage rate is about one percent in these hospitals.  In our well established customers, those people who continue to order the two sensors per box per month making up the 1,000 or so per month that are shipped, it's actually about 2.8 percent call it 3 percent in round numbers of actual usage.  What that means is that out of every 100 births, 3 of them are monitored. 

            And interestingly, to implicate the introduction of fetal saturation monitoring for that rise in Caesarean Section in those three hospitals, you'd have to have nearly four supplemental -- four extra Caesarean Sections for every monitored patient, which would be a neat trick if you could do it.

            It turns out that these three hospitals are not unique, not even close. This is data from National Center for Health Statistics. I tried very hard getting the 2002 data. It just has not yet been published and I can't find it. But this shows the total Caesarean section, that's primary, secondary, VBACs, everybody, for the country from 1989 to 2001.  And what you see was a drop from '89 to nadir in '96 and then as the social customs changed in the country and VBACs fell out of favor and it became fashionable to schedule one's delivery with no negative reimbursement phenomenon involved, the C-section was rapidly rising during this period and shows no signs of abatement.

            And so if we ignored that last slide and simply ignored the question of the background rise in C-section rates of the same magnitude that we were looking for in the study, with a one percent usage rate, we'd need about 30 years or so to complete the study which did not seem reasonable.

            As I said, the primary condition driving what I believe was the panel concern for general use of widespread use, simply does not exist.  The actual use is somewhere between 1 and 3 percent. And, we therefore along with the FDA, came to the conclusion that the general use study as currently proposed didn't seem practical and was incapable of reaching a valid conclusion.

            Well, the rest is the same thing.

            That's what I have on the presentation.  I'd be delighted to answer questions or take comments.

            CHAIR O'SULLIVAN:  Does the panel have any questions?  Any comments?

            I do have one comment, Dr. Swedlow.

            DR. SWEDLOW:  Yes.

            CHAIR O'SULLIVAN:  I think obstetricians don't ignore the fact that the pulse oximeter wasn't necessarily of any value, i.e., they ignored it altogether.  If they were paying attention to the tracing, they were using the tracing and if it was perfectly okay, then that's what they probably went on especially when you have to realize that at the time we were talking about this, it was a known fact that somewhere around 60 to 70 percent at the time you would not get any evidence of anything that was happening. So I kind of take issue with the fact that they ignored it altogether, unless you had them answer that question for you, assumed that everything was really okay.  I think that would only the case of in the presence of a normal fetal heart rate tracing.

            DR. SWEDLOW:  May I address that question?

            CHAIR O'SULLIVAN:  Sure.

            DR. SWEDLOW:  I believe what you just said was that they would ignore the fact that there was no FSpO2 being displayed during periods of normal fetal heart rate tracing?

            CHAIR O'SULLIVAN:  I'm assuming that they were paying attention to the fetal heart rate tracing, and that is the reason why they didn't pay any attention to the fact that it wasn't picking up.

            DR. SWEDLOW:  Right.  The facts are, however, that the periods of ignoring the fetal heart rate tracing were the periods of very, very abnormal -- the time of concern is a time of very abnormal fetal heart rate tracing --

            CHAIR O'SULLIVAN:  Yes.

            DR. SWEDLOW:  And no pulses.  So it's not a --

            CHAIR O'SULLIVAN:  Yes.

            DR. SWEDLOW:  Perhaps I misunderstood you.  I think what you asked was -- your comment --

            CHAIR O'SULLIVAN:  I didn't ask. I made a comment.

            DR. SWEDLOW:  I understand. I think your comment was that it's not surprising that they would ignore the absence of pulses if they were paying attention to the fetal heart rate tracing.

            CHAIR O'SULLIVAN:  And it was normal.

            DR. SWEDLOW:  And it was normal.  Right.  Those periods of concern that we brought to the FDA were periods of very abnormal fetal heart rate tracing and no pulses.  Repeated decelerations absent variability, no pulses sometimes for hours.  That was what was being ignored.

            CHAIR O'SULLIVAN:  I see.

            DR. SWEDLOW:  The second comment you made was that you were not surprised that they would ignore periods of no pulses since 67 percent of the time there was, in fact, no pulses.  You got it backwards.  Sixty-seven percent of the time there--

            CHAIR O'SULLIVAN:  Yes. I'm sorry. I meant it the other way around.  Thirty-three percent of the time the machine, the pulse oximeter did not pick up.

            DR. SWEDLOW:  Correct.

            CHAIR O'SULLIVAN:  Sorry. Thirty-three percent of the time.

            DR. SWEDLOW:  I forgot the third comment.  I'm sorry.  What was the third comment?

            CHAIR O'SULLIVAN:  I don't remember.  I don't know if there was a third one.

            DR. SWEDLOW:  Okay.

            DR. WOLFSON:  I'm just curious about utilization. Did you interview practitioners in these three hospitals where you started this trial find why they chose not to utilize the technology?  Was this after the ACOG letter so people were discounting it?

            DR. SWEDLOW:  No. Yes, we did and no it wasn't.  And we also went out and talked to a lot of our regular customers, the ones that were continuing to use it all along.  And the answer is very simple.  The obstetricians are not terribly concerned about the condition of the patient most of the time. They only use it in those cases in which they are very worried. Our own customers tend to use it in about 2.8 percent on the average of all patients. And they typically only use it or put it in when the fetal heart rate is quite worrisome, very confusing, not clear. They just don't use it as a routine let's drop one in, you know.  It's a special thing.

            DR. WOLFSON:  Right, but --

            DR. SWEDLOW:  There was no difference in that conversation pre or post-ACOG.  It seems to have no impact.

            DR. WOLFSON:  Okay.

            DR. SWEDLOW:  I remember there were 40 comments that we received from the field from customers after ACOG, that was after "Dear Doctor."

            DR. WOLFSON:  Because wasn't the original trial that it was placed because there was a nonreassuring fetal heart rate tracing?

            DR. SWEDLOW:  Yes.

            DR. WOLFSON:  So if you had 30 percent utilization during the trial and you're now -- you just said, I thought I heard you just say that your overall usage is about 3 percent.

            DR. SWEDLOW:  Right.  You may remember that --

            DR. WOLFSON:  What happened?

            DR. SWEDLOW:  Well, you may not remember.  During the January panel meeting when I was being asked about this, I also showed a slide that demonstrated that for those patients who never progressed past Class I or what is today Class I on this trial, there was no impact of using the technology at all.  It was only those patients whose fetal heart rate pattern had progressed up to that Class II version.

            DR. WOLFSON:  So the proportion of candidates declined based on experience?

            DR. SWEDLOW:  Correct.  Remember that during the original RCT nobody knew how to use this and neither did we, for that matter.  And what were the sort of rock solid criteria after 5 years of experience, people know how to use it now.

            Did I understand the question?

            DR. IAMS:  During the RCT the enrollment was capitated?

            DR. SWEDLOW:  Capitated?

            DR. IAMS:  Centers enrolling patients received a payment when they enrolled a patient?

            DR. SWEDLOW:  No.

            DR. IAMS:  No?

            DR. SWEDLOW:  No. During the RCT the hospital received financial support for a in-house full time research nurse. Neither the physicians, the hospitals nor the patients received there -- there was no bounty, if that's what you're asking.

            DR. IAMS:  No, I don't mean bounty to the patient, but investigators respond to the reward of participating in a study and typically you want to please the study master, which is means enrolling patients.

            DR. SWEDLOW:  Right.

            DR. IAMS:  In answer to Bob's answer, I don't think the indications for the use of the device disappear.  I think what you saw was his definition of what could be a nonreassuring tracing during the RCT.  And so devices were inserted because those criteria were met.

            DR. SWEDLOW:  Correct.

            DR. IAMS:  When you asked practicing clinicians who were not participating in a trial if they would be interested in using the device under those circumstances, they frequently said that tracing might fit your definition of what's not reassuring, but it doesn't worry me enough to bother wheeling this extra device in there and doing all the stuff I have to do to apply a second piece of technology.

            DR. SWEDLOW:  Your statement is absolutely correct, consistent with the facts with the sole exception that there was no per-patient fee paid to anyone during the -- well, ever, as far as I know.

            CHAIR O'SULLIVAN:  Any other comments?

            Thank you, Dr. Swedlow.

            DR. SWEDLOW:  Thank you.

            CHAIR O'SULLIVAN:  Does the panel have anything else that they want to discuss while we're on this subject?  Okay.  With that --

            DR. IAMS:  Mary Jo, I have a question.  Our original approval was conditional upon receiving results of the studies that we've just heard described.  Is it now our task to revisit those conditions and make some judgment or are we simply to receive the report and pass it on to FDA, and what do we do now?

            MS. BROGDON:  If you have comments to make, we'd be happy to hear them.  FDA will have to decide whether to cancel some of the conditions of approval.  So if you have views on any of this, we'd be happy to hear them.

            DR. IAMS:  I would just say I don't really know what to do right now, but I distinctly remember casting an affirmative vote for approval of this device based on the condition that studies be forthcoming and based on the correct assumption in retrospect that it would take the network a long time to get another trial up and running. So we thought it would be a reasonably safe device to approve, we just weren't sure it would have the beneficial effect that it might have, but we thought we would see that by now.  But, obviously, you know it's been difficult for the company to comply.  I'm not being critical of them. But I don't know what we do now.

            My conditions haven't been met and my affirmative vote is still hanging out there waiting for what? 

            CHAIR O'SULLIVAN:  Well, I don't think that unless the company were to decide to expand its study to multi other institutions to try to cut down on 33 years, that the company can really complete what we asked them to do.  It would cost them a great deal of money to do that, too, if that were the case.

            At this point in time I don't think that that particular requirement should exist.  I think we're really wait on the trial, on the NICHD trial to see whether we feel -- I mean, I think that the company can certainly sell its equipment, but we don't know whether it's really valuable or not until we get that FOX trial.

            DR. LARNTZ:  That's an amazing trial. It's going on, it accruing.  I think you're going to have an answer with respect to the monitoring that it's worthwhile.

            DR. IAMS:  Yes, it's just taking a while to recruit patients.  It's been a difficult trial to recruit to.

            DR. LARNTZ:  What else is new in this world? I mean, it's always difficult.

            CHAIR O'SULLIVAN:  I think if we really want to know how it's going to work out there, that's going to be the one that's going to tell us.

            DR. POLLARD:  Yes. I would just add to this.  We're -- I don't know what the right word is -- troubled.  It looks like there's a problem with this study.  To be honest with you, we're still looking at this and trying to see what does this mean and what does it mean with respect to the condition and what are we going to do about that.  Obviously, if you all have some ideas you want to put on the table, we'd be interested in that. We certainly will take those into account as well.

            CHAIR O'SULLIVAN:  Yes, Dr. Wolfson?

            DR. WOLFSON:  Colin, I guess the question aren't we somewhat reassured, though, relative to the dystocia trial that's going on?  Because I thought wasn't that one of the other key issues is that the value of the technology in reducing the Caesarean Section rate was, in a sense, marred by the dystocia question. But now we've got what seems like, I haven't gone through the details, a relatively well designed trial that has indeed demonstrated that this is a marker for dystocia.  So that means that we can take dystocia data out and it further reassures us that there is an improvement in the Caesarean Section  rate with the utilization of this technology when it is appropriately applied. Are am I incorrect?

            DR. POLLARD:  Yes, we haven't reached that conclusion.  We just got a draft copy of the manuscript, you know, like Thursday or Friday.

            DR. WOLFSON:  Okay.

            DR. POLLARD:  So we're going to look at that, too.

            DR. WOLFSON:  Yes.

            CHAIR O'SULLIVAN:  So, Jay, if you're asking is there anything we can do about bringing back our approval, the answer is no.

            DR. IAMS:  Well, I don't think that's true.  We were told at that time that we could rescind our approval. I distinctly remember that.  I mean, I don't know that we want to rescind our approval, but we were told at the time that we could rescind it at a future date if the conditions weren't met.  But what's happened is that the conditions haven't been met, but for reasons that are difficult for anyone to either foresee or manage at this point, I don't clearly quite know what to do. But we are in the position of having approved an device under conditions that haven't been met -- conditionally and not having those conditions met. And now we all kind of hope they will be met by the Network trial.

            Is the company still providing, I assume, additional post-market surveillance without actually conducting -- the general use study is terminated, is that right?

            DR. SWEDLOW:  Yes.

            DR. IAMS:  So we can sit here and do nothing and just let the FOX trial be the answer, which I think is apparently the consensus of the panel. I have no desire to tell the company to stop marketing their device based on what we've learned today. I'm just feeling somewhat, like I said, responsible for bringing some closure to that because it was an expected condition of our votes.

            CHAIR O'SULLIVAN:  Yes. I think you're right. I think the fact is that since they cannot complete the trial and they are doing the post-marketing surveillance, then we cannot hold them back from continuing to market their product. And we are going to have to wait on the FOX trial.

            DR. POLLARD:  Dr. O'Sullivan, could I just make one comment. Because I think we did address this, I think it was 2 years ago when we brought a PMA supplement back to the panel with the revised post-approval study plan, you know, once we appreciated what Dr. Spong's group was going to do and essentially broke up the original post-approval study protocol into three protocols. 

            You know, there may have been misunderstanding at the time of the vote in January of 2000, but there was never the notion that the PMA approval would be rescinded. The approval stands on its own.  The requirement attached to the approval order is that these studies be conducted. How they're evaluated, the impact they have on the device on the post-market setting could be probably as diverse as the results could be. They could turn into other requirements FDA might want to apply and it could turn into change labeling, change in training, registry.  I mean, it did not speak to rescinding the PMA approval order per se.

            CHAIR O'SULLIVAN:  Are there any other comments?  If not, this meeting is adjourned until tomorrow morning.

            (Whereupon, at 4:08 a.m. the meeting was adjourned, to reconvene June 10, 2003.)