UNITED STATES OF AMERICA

 

           FOOD AND DRUG ADMINISTRATION

 

                     + + + + +

 

    CENTER FOR DEVICES AND RADIOLOGICAL HEALTH

 

                     + + + + +

 

   OBSTETRICS AND GYNECOLOGY ADVISORY COMMITTEE

 

                     + + + + +

 

                SIXTY-FIFTH MEETING

 

                     + + + + +

 

                      MONDAY

 

                  APRIL 22, 2002

 

                     + + + + +

 

            The panel met at 8:00 a.m., in Salons E and F of the Gaithersburg Marriott Washington Center, 9751 Washingtonian Boulevard, Gaithersburg, Maryland, at 9:00 a.m., Dr. Jorge D. Blanco, Chairman, presiding.

PRESENT:

JORGE D. BLANCO, M.D., Chairman

CAROL L. BROWN, M.D., Member

GARY S. EGLINTON, M.D., Temporary Voting Member

JAY D. IAMS, M.D., Temporary Voting Member


PRESENT (continued):

KLEIA R. LUCKNER, J.D., M.S.N, Consumer Representative

MARY LOU MOONEY, R.A.C., Industry Representative

MICHAEL NEUMAN, M.D., Ph.D., Temporary Voting Member

MARY JO O'SULLIVAN, Member

SUSAN M. RAMIN, M.D., Temporary Voting Member

RICHARD E. RINGEL, M.D., Temporary Voting Member

DAVID B. SEIFER, M.D., Member

NANCY C. SHARTS-HOPKO, Ph.D., Member

ALICIA Y. TOLEDANO, Sc.D., Temporary Voting Member

ROBERT N. WOLFSON, M.D., Ph.D., Temporary Voting Member

JOYCE WHANG, Ph.D., Executive Secretary


                  C-O-N-T-E-N-T-S

 

Introductions................................... 4

 

Introductory Remarks........................... 12

Colin Pollard, Chief,

Obstetrics and Gynecological Devices Branch

 

Open Public Hearing

  Raul Artal, M.D.............................. 17

 

Sponsor Presentations

  Karl G. Rosen, M.D., Ph.D.................... 24

  Overview of STAN concept and Swedish RCT

 

  Ingemar Kjellmar, M.D........................ 53

  Perinatal Care in Sweden

 

  Karel Marsal, M.D., Ph.D..................... 58

  Management of Labor and Delivery

 

  Lawrence D. Devoe, M.D....................... 64

  Applicability/Usability of STAN in U.S.

 

FDA Presentations

  Kathryn S. Daws-Kopp......................... 81

  Overview of FDA Review

 

  Julia A. Corrado, M.D........................ 88

  Clinical and Statistical Review

 

Panel Discussion.............................. 120

 

Open Public Hearing........................... 278

 

Panel Deliberations and Vote.................. 282

 


               P-R-O-C-E-E-D-I-N-G-S

                                         8:07 a.m.

            CHAIRMAN BLANCO:  All right, let's go ahead and call the meeting to order.

            I want to remind everybody that there are signup sheets outside in the front.  Please sign in and let us know who you are and who you are affiliated with.

            Just a couple of housekeeping matters before we start:  As usual, let's not have any outbursts from the audience.  If you would like to be recognized, at the appropriate time you will be recognized; you can come to the podium and speak.  Always identify yourself and identify whether you are associated with any particular entity, and whether you have, at least the first time that you come up, you need to state whether you have any possible conflict of interest with the issues at hand today.  That includes any type of employee, employer, consultant, travel, per diem, or any other type of relationship with any of the companies that might have some business before the panel.

            At this point let's go ahead and have some introductions of who the panel members are.  If we could start over here, please state your name and affiliation.

            MS. BROGDON:  Good morning.  I'm Nancy Brogdon.  I have been the Director of the Division of Reproductive, Abdominal and Radiological Devices, FDA.

            CHAIRMAN BLANCO:  And next to her will be Dr. Michael Neuman, who will join us very shortly.

            Go ahead.

            DR. RINGEL:  Yes, I'm Dr. Richard Ringel, Pediatric Cardiology at Johns Hopkins Hospital.

            MS. TOLEDANO:  Alicia Toledano, Center for Statistical Sciences, Brown University.

            DR. WOLFSON:  Good morning.  I'm Robert Wolfson.  I'm an independent maternal fetal medicine specialist and obstetrician/gynecologist and engineer.

            DR. SHARTS-HOPKO:  I'm Nancy Sharts-Hopko, Professor of Nursing, Villanova University.

            DR. IAMS:  Jay Iams, Professor of Obstetrics and Gynecology at Ohio State University.

            DR. WHANG:  I'm Joyce Whang.  I'm the Executive Secretary of this panel.

            CHAIRMAN BLANCO:  I am Jorge "George" Blanco.  I am a perinatologist in private practice in Odessa, Texas.

            DR. RAMIN:  Good morning.  I'm Susan Ramin.  I'm a maternal fetal medicine specialist at the University of Texas, Houston, Medical School.

            DR. EGLINTON:  Gary Eglinton, maternal fetal medicine, Cornell and New York Hospital, Queens.

            DR. O'SULLIVAN:  Mary Jo O'Sullivan, maternal fetal medicine, University of Miami.

            DR. BROWN:  Carol Brown, gynecologic oncologist, Memorial Sloan-Kettering Cancer Center.

            DR. SEIFER:  David Seifer, reproductive endocrinologist, UMDNJ, Robert Wood Johnson Medical School.

            MS. MOONEY:  I'm Mary Lou Mooney, 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 clinical administrator at Toledo Hospital in Women's Health, and I'm the consumer rep.

            CHAIRMAN BLANCO:  Thank you very much.

            Michael, I introduced you, but please do the honors for yourself, if you would like.

            DR. NEUMAN:  I think you did such a good job, Jorge, I think it's fine.  I'm Mike Neuman from the Memphis Joint Program in Biomedical Engineering of the University of Tennessee Health Science Center in the University of Memphis.

            CHAIRMAN BLANCO:  No, you did much better than I did.  Thank you.

            (Laughter.)

            DR. NEUMAN:  Can't do it in one breath though.

            (Laughter.)

            CHAIRMAN BLANCO:  We'll have to train you.  We'll have to invite you back more, so we can train you.

            DR. NEUMAN:  That's quite all right.

            CHAIRMAN BLANCO:  All right, at this point I would like to introduce the FDA press contact, if she would please stand.  The FDA contact for this particular meeting is Nancy Brogdon, Director, Division of Reproductive, Abdominal and Radiologic Devices.  Thank you.

            All right, at this point I will turn it over.

            DR. WHANG:  Good morning.  We have two additional OB/GYN Devices Panel meetings scheduled for this year.  They are July 22nd and 23rd and then October 21st and 22nd.

            Today's panel includes seven temporary voting members, and I will now read their appointment to temporary voting status.

            "Pursuant to the authority granted under the Medical Devices Advisory Committee Charter, dated October 27, 1990, and amended August 18, 1999, I appoint the following individuals as voting members of the Obstetrics and Gynecology Devices Panel for this meeting on April 22nd, 2002:

            "Gary S. Eglinton, M.D.; J. D. Iams, M.D.; Michael Neuman, M.D., Ph.D.; Susan M. Ramin, M.D.; Richard E. Ringel, M.D.; Alicia Y. Toledano, Sc.D.; Robert N. Wolfson, M.D., Ph.D.

            "For the record, these people are special government employees and are consultants to this panel.  They have undergone the customary conflict-of-interest review, and they have reviewed the material to be considered at this meeting."

            This is signed by David W. Feigal, Jr., M.D., M.P.H.  He is the Director for the Center for Devices and Radiological Health, and it is dated April 18th, 2002.

            Now I will read the conflict-of-interest in for the record.  "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 for this meeting and all financial interests reported by the Committee participants.  The conflict-of-interest statutes prohibit special government employees from participating in matters that could affect their or their employer's financial interests.  However, the agency has determined 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, a waiver under 18 USC 208(b)(3) has been granted to Dr. Richard Ringel for his stockholding, between $5,001 to $25,000, in the parent of a competing technology firm.  The waiver allows this individual to participate fully in today's deliberations.  Copies of this waiver may be obtained by submitting a written request to 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 certain matters regarding other panelists.  Drs. Gary Eglinton, Michael Neuman, and Robert Wolfson reported current or previous interest with firms at issue, but in matters not related to today's agenda.  The agency has determined, therefore, that they may fully participate in the panel's deliberations.

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

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

            There will be transcripts and videos available for today's meeting, and if there are any presenters to the panel who have not already done so, they should provide FDA with a hard copy of their remarks, including overheads.

            Kathy Daws-Kopp ‑‑ would you please stand? ‑‑ will collect these from you at the podium.  Thank you.

            CHAIRMAN BLANCO:  Thank you.

            It's a pleasure for me to introduce Colin Pollard, Chief of the Obstetrics and Gynecological Devices Branch, who will make some introductory remarks.

            Colin?

            MR. POLLARD:  Thank you, Dr. Blanco.

            First of all, I want to welcome the panel.  We've got a few new faces, and we really appreciate the time and trouble you all went to be here, and especially for the new panel members, the training that you went through yesterday and this morning.

            I would also like to welcome the sponsor, who has had to travel quite a distance to be here today.

            Before we get into today's agenda, I would like to go over a few things with you.  Since our last panel meeting, FDA has approved two PMAs.  In September we approved the Novasure Endometrial Ablation System.  This is the fourth such device of this type.  As we get more experience with that, it's less and less likely that we bring PMAs of this sort before you, unless it's something new and different, a new energy mode or some new safety issue, or something like that.

            Just last month, we approved the Lea's Shield, which is vaginal barrier contraceptive device.

            As Joyce just mentioned, we've got two panel meetings scheduled for the remainder of the year:  July 22nd-23rd and October 21st-22nd.  It's quite possible we will use both of those dates, but we'll see what the agenda holds for us.

            I would like to talk a little bit about post-approval studies.  First of all the purpose of a post-approval study, it needs to be understood that there's already, before you even get into the post-approval study, the determination on the part of FDA that a reasonable assurance of safety and effectiveness has already been established.  However, the post-approval studies may address either long-term events or rare events.  If the PMA was based, if the pivot study supporting the PMA was based on surrogate endpoints that FDA thought were reasonable, a post-approval study might be done to validate those endpoints or a post-approval study might answer specific panel concerns.

            I mention this because at our meeting in July our Office of Post-Market Surveillance is planning to give a presentation, and it is going to go over a number of post-market activities that FDA engages in, including an update on post-approval studies.

            Today we're going to bring before you a PMA, which you have had a chance to review already, for the STAN S21 Fetal Heart Monitor submitted by Neoventa Medical.  I just wanted to give a few remarks about that.

            First of all, this falls generally in the category of electronic fetal monitors, and I would just like to highlight that most electronic fetal monitors are Class II devices and get to the market through the 510(k) pre-market notification process where the sponsor would show that they're substantially equivalent to a monitor that's already on the market.

            However, if there is a new feature for that fetal monitor, one that we believe has significant implications, in that case we would require a PMA.  In the case of the PMA before you today, we felt that the use of the fetal ECG waveform in this monitor, the special analysis to identify ST events, and the claims that it would improve assessment, we felt that, when all taken together, that this was something that required a PMA.

            Just real briefly ‑‑ you have this in your PMA ‑‑ the intended use of the STAN monitor is to use the fetal ECG analysis with the cardiotocography to obtain information on the impact of labor on the fetus, and that is intended to improve the assessment of fetal condition during labor.  I'm not going to go through all the indications for use that are listed in the PMA, but it represents a fairly broad approach.  It's intended for use on a singleton fetus in the vertex presentation with greater than or equal to 36 weeks gestation.

            Real briefly, I know we went over this yesterday for the new panel members, and I just remind the more experienced panel members, the regulatory framework for PMA reviews.  I want to highlight three elements:  valid scientific evidence, safety, and effectiveness.

            With respect to valid scientific evidence, FDA is given quite a bit of latitude has to what to recognize for a given product, for a given intended use, everything from well-controlled clinical studies, partially-controlled studies, all the way to even reports of significant human history.

            Our assessment of safety ‑‑ in fact, this must be based on valid scientific evidence.  In the case of safety, we must make a determination that the benefits outweigh the probable risks, and for effectiveness ultimately that the studies must have shown a clinically-significant result when used for the intended population and used the way it is supposed to be used.

            Joyce and Jorge are going to go over this with you later, but your recommendation, which we consider very seriously in the context of our making our decision, should come in the form of a recommendation, a panel recommendation, that would take the form of one of those three aspects, either an approval, a full approval, an approval with conditions, or not approvable.

            So, finally, our agenda, we are going to hear first from the sponsor.  After the open public hearing, FDA reviewers will present their findings, and you will have an opportunity to discuss this, hear more from the sponsor and audience before you engage in the questions that we've prepared for you.

            Thank you, Dr. Blanco.

            CHAIRMAN BLANCO:  Thank you, Mr. Pollard.

            The next item on our agenda is the open public hearing.  We have one speaker signed up to present before the panel, Dr. Raul Artal, Vice Chairman, ACOG OB Practice Committee.

            Is there anyone, before Dr. Artal addresses the panel, is there anyone else who would like to participate in the open public forum at this time?

            (No response.)

            All right, thank you.

            Dr. Artal?

            DR. ARTAL:  Good morning.  Dr. Blanco, members of the panel, ladies and gentlemen, thank you for giving me the opportunity to represent ‑‑

            CHAIRMAN BLANCO:  I'm sorry, Dr. Artel.  Can we start by stating your name, your relationship, and whether you have any conflict of interest, and then go ahead with your presentation?  I apologize for interrupting you.

            DR. ARTAL:  No, thank you.  I appreciate this.  Thank you very much.

            My name is Raul Artal.  I'm currently the Vice Chair of ACOG's Committee on Obstetric Practice.  I'm also the Professor and Chairman of the Department of Obstetrics, Gynecology and Women's Health at St. Louis University.  I do not currently have any conflicts of interest with this device or any current device that is being considered by this panel.

            CHAIRMAN BLANCO:  I think you have a conflict of interest with a computer, though, the way they're working on them.

            (Laughter.)

            DR. ARTAL:  Yes, well, I'm all prepared, Mr. Chairman, for this eventuality, too.

            (Laughter.)

            CHAIRMAN BLANCO:  Thank you.  Go ahead.

            DR. ARTAL:  Thank you.  Again, thank you for giving me the opportunity to represent ACOG here today.

            The charge of the ACOG's Committee on Obstetric Practice is to provide the ACOG membership with objective, valid, scientific evidence and guidelines for current practice, as well as existing and emerging technology.

            As to the device that is being considered today by the panel, we had the opportunity to review the following documents:  The Summary of Safety and Effectiveness Data of Neoventa Medical, the publication listed in the summary, the detailed analysis of the Amer-Wahlin, et al., the Swedish randomized control trial, as published in the Lancet, August 2001, and we also had the opportunity to review a presentation of the lead author, Professor Rosen, at the International Perinatal Society in Barcelona last October.

            The ACOG Committee on Obstetric Practice bases its opinions on the fundamental principles of evidence-based medicine.  The extent of the evidence, the benefits versus risks, patients' inconvenience, cost, and patients' values.

            As to the study presented and the adequacy of the study design, the experimental study design was considered to be appropriate for evaluating the efficacy of cardiotocography plus ST analysis versus cardiotocography only for intrapartum fetal monitoring.  However, questions remain about the device's ability to detect either metabolic acidosis or hypoxia, and certain improvements in future validation studies will be required.

            One concern about the study design is that, while the main outcome variable was metabolic acidosis at birth, very few contemporaneous fetal blood scalp sampling determinations were conducted on fetuses affected by metabolic acidosis.  The overall rate of fetal acidosis acidemia reported in these studies was significantly ‑‑ significantly ‑- lower than the one reported in our country.

            In terms of adequacy of the materials, methods, and procedures, the methods for identifying eligible study participants, randomizing participants, training clinicians to use the CTG or CTG plus ST, gathering baseline, and follow-up measurements, and analyzing the data are appropriate and clearly described in the materials submitted.

            However, I have to again emphasize the overall very low incidence or rate of fetal acidemia, significantly lower than in this country.

            No data is provided to determine false positives and false negatives for the proposed method to detect metabolic acidosis/hypoxia, such as physiologic events, environmental factors, fetal positioning, and so on.  None of this has been reported in the analyses.

            ST segment deviations are known to be caused by the following events:  injury to the cardiac muscle or ischemia, changes in the synchronization of ventricular muscle depolarization, overload or strain on ventricular muscle, drug or electrolyte influences, and certainly could be observed also as a normal physiological variant.  It is this particular thing that we have not seen any data reported in these reports.  Particularly, it's known that nested segment deviation can appear as a normal physiological variant during stressful events.  In the studies completed and data presented, have these factors been considered?

            Other things of concern to the ACOG Practice Committee was that, as reported on page 536 of the Lancet article, second column, first paragraph, the interim analysis revealed protocol violations in which the recommendations to intervene were disregarded, and babies with cord arterial metabolic acidosis were born.  Indeed, the protocol violation occurred in both arms of the trial, it appears.  What percentage of births were affected in each arm?

            In terms of accuracy of interpretation of the results, the results as presented in the tables and text material reviewed are presented very clearly.  However, what is the sensitivity and specificity of the CRT ST for detecting cord artery metabolic acidosis or hypoxia?  And more important, what is the likelihood ratio?

            Nowadays, for an evaluation of a diagnostic or a screening test, likelihood ratio is the appropriate effect size to describe the performance of the test.  So what is the likelihood ratio?

            In summary, we urge the panel to clarify with the sponsors of the device the following issues: What are the false positives and false negative results for the proposed method in detecting metabolic acidosis/hypoxia?  Do other events influence the ST waveforms in the fetus?  Clarifications on protocol violations.  What is the sensitivity and specificity of the CRT plus ST for detecting fetal metabolic acidosis/hypoxia?  What is the likelihood ratio?

            The ACOG Committee on Obstetric Practice recommendation is that, although this new technology appears very promising, additional information is required.  Currently, the Committee cannot endorse the adoption of this device in clinical practice.  ACOG is particularly concerned with the introduction of new technology to clinical practice, which could further escalate the cost of medical care without necessarily improving clinical outcome.  For example, how does CTG plus ST analysis improve outcome when compared with a simple and less expensive test like fetal stimulation testing?

            Thank you very much.

            CHAIRMAN BLANCO:  Thank you, Dr. Artal.

            We'll open it up once again, if there's any other members of the public who would like to address the panel at this point.

            (No response.)

            Okay, not seeing anyone, we will move on to the next item on the agenda.  The next item on the agenda is the presentation by the sponsor, and I believe Professor Karl Rosen is going to begin the presentation.

            DR. ROSEN:  Mr. Chairman, members of the panel, ladies and gentlemen, my name is Professor Karl Rosen.  I'm a pediatrician/perinatal physiologist.  I am currently Medical Director of Neoventa Medical, where I hold shares.

            What I would like to start is to present the way we can detect hypoxia during labor, but also touch on the dissemination of knowledge as we see it with the research we are doing currently in Europe.  I will be followed by Professor Ingemar Kjellmar, who will present the current perinatal level of care in Sweden, and then Professor Karel Marsal will draw parallel observations with regard to the U.S. and Swedish obstetric care, and Professor Lawrence Devoe will follow with comparing or addressing the current issue in the U.S.

            Now to start this, we could look at a case which is a term delivery where there's been ruptured membranes for three days, induction.  On the upper part you see the heart rate trace, uterine activity, and here you have something called the T/QRS ratio.  This is now the height of the T wave related to that of the QRS.  These are ECG complexes recorded from the fetal scalp electrode.  We could have a look at events as they emerged in the heart rate trace as well as in the ST with this case.

            So we move forward.  A baseline heart rate of roughly 160 beats per minute.  The mother has an increased temperature.  She's been given Parasetimol. Still no ST events, and heart rate now in still first-stage labor is constant with a tachycardia reactivity.

            As we now start to enter the end of first stage and approaching active pushing in the second stage, deceleration starts to emerge, and with that we have a first ST event, which is now the baseline rise of the T/QRS ratio, and there are now more T waves that you can see.

            Now the midwife in charge, she also recognizes that heart rate ratio is now becoming what we call abnormal.  So there is an indication that this is ongoing hypoxia.  She reports to the clinician.  Over this time period, there are further indications from the STAN monitor that there is an ongoing ST elevation.

            However, the clinician thinks the heart rate rates may not look that bad.  There are signs of accelerations, and while they are discussing matters, the decision is made to perform a fetal scalp sample, which, as you see, is low, 6.93, and obviously, then, emergency intervention is required.  There is failed vacuum extraction, followed by an emergency Caesarean section and a baby delivered approximately one hour after the onset of hypoxia in second-stage labor.

            As you may see, the Apgar scores are very low.  There is cord vein metabolic acidosis.  There simply wasn't any blood in the artery.  The baby needs active resuscitation and needs time on the respirator, but the outcome seems to be good.

            And this you are here to discuss is really what happens when things like these emerge, and where we know that there is sometimes some uncertainty about how to interpret the heart rate rates, and where we are looking for additional information now contained in the ST wave form of the fetal ECG.  But, at the same time, to have a system that presents to everyone what actually is taking place, so that delay factors are avoided as much as possible.

            So what we aim for is to identify fetuses as risks of an adverse outcome, based on our understanding of the path of physiology involved.  We know labor is a stress, and we know also that Nature has provided the fetus with very good resources to manage labor.  Most fetuses are untroubled, although we may find a lot of heart rate changes.  Some are troubled somewhat but managing and compensating fully, in no immediate danger.  But then there are a few that's more troubled and forced to utilize key resources in an attempt to compensate, and some may not even be able to fully compensate, and we need to identify those.

            If we look at the history of fetal monitoring heart rate, clearly, the basis on which we stand for more than 100 years.  When electronic fetal monitoring was introduced in the sixties, it was thought that we needed additional information.  Scalp pH was introduced already at that point in time.  The last couple of years we have seen the introduction of posimetry, and now we have the STAN monitor based on qualified assessment of CTG velocity ST waveform changes.

            What is, then, the problem?  Well, we have seen the emerging, rising Caesarian section rates, and they vary across the world, with the U.K. and the U.S. running rather high figures, and in Sweden we have somewhat lower figures.  You may wonder what the reason is for this.  Well, one of those is probably the uncertainty about how we assess the condition of the fetus during the labor, in which there is uncertainty.  Clearly, there is a need for intervention, because we don't want to see babies being affected.

            So what are the alternatives?  Well, we have blood-based assessment, oxygen saturation, scalp pH, but maybe additional and more significant information could be then gained by identifying specific organ-tissue functions, such as metabolic acidosis, ST analysis, and also the fetal heart rate.

            Where I started, well, that was 30 years ago, with a by chance observation, more or less, in the mature guinea pig fetus exposed to hypoxia, noticing on a strip chart recorder these marked ST events with enormous T wave amplitude changes prior to the bradycardia.  Obviously, at that point in time we had been using the electrocardiogram in clinical medicine, and we know that in a stress test in the adult there are good data to use to assess the condition of that myocardium we are now looking at, meaning that in case of an ST segment depression or changes in T wave amplitude, that would signify a fetal or a myocardium, rather, not fully compensating, and finding it sometimes hard and sometimes there would be coronary insufficiency.

            Why would we see these changes?  Well, we know that the ST interval represents the time of repolarization of the myocardium.  This is now where energy is required.  The ion membrane pumps are working hard, and metabolic events occur.  So in case of hypoxia, well, this is where we expect changes to be seen.

            Now then we need to quantify these changes.  As you noticed, there were marked peak T waves.  So we could look at the amplitude of the T wave and relate it to that of the QRS.  QRS, well, that's a passive electrophysiological event, and its QRS amplitude basically stays the same from the first through the last heart beat.  So by taking the ratio of T/QRS, in this case in the normal, 5 versus 50, a ratio of .10; then during hypoxia, 10 versus 50, so the ratio increases to .20.  We have a measure that we can record and use.

            Now the animal work that we have conducted, starting in the seventies, initially were observations on hemodynamic cerebral metabolic function during experimental hypoxia, but also in spontaneous labor in association with intrauterine deaths in predominantly the fetal sheep preparation.

            We also did studies on myocardium metabolism, showing that the ST rise was associated with the breakdown of myocardial glycogen, in particular.  Now this turned out to be regulated by beta receptors and, in particular, the catecholamine surge that goes with hypoxia.  We also studied gestational age and the impact of growth retardation.  In these fetuses we saw biphasic ST segment depressions emerging whereas we didn't see in the appropriately grown fetuses.

            So, to summarize, the experimental work, the normal ST, anaerobic metabolism, positive energy balance, with hypoxia, we see a rise in T wave amplitude that is caused by the surge of adrenalin activating beta receptors and then activating glycolysis.  So, with that, we have the liberation of potassium ions, and we have the close linear relationship between the rate of rise in T wave amplitude and the rate of breakdown of myocardial glycogen.  Thus, we are looking at anaerobic metabolism, which we know from all physiological studies being of the most important defense systems for the fetus exposed to hypoxia.

            But there are also other fetuses, other signs within the ECG.  This we call biphasic ST, which is now where the ST segment is leaning downward, cutting the baseline.  Grade them, grade 2 and 3 being those being significant, and where we are now looking at the pumping performance of the myocardium.  So recognizing a reduction in the ability of the myocardium to respond to the stress of labor.

            We know that prematurity infections will have an impact on that capacity, and we find these associated with biphasic ST events, overall demand.  If the heart has to pump beyond its capacity, we will sometimes see biphasicity, rare instances of myocardial dystrophy, but also chronic hypoxia and in the initial phase of acute hypoxia these changes could emerge.  They're not associated with metabolic acidosis initially, normal acid base, but they signify that the pump is not working at full performance.

            What are we recording?  Thirty beats.  Put them together in an average, thereby getting a better signal, allowing us to assess the specific components of the fetal ECG.

            So then the need for extractive clinical work.  In the eighties, observation started and technological developments.  In the nineties, we saw the Plymouth randomized trial, fostering further technological developments, verified in the EU Observational Multi-Center Study, Nordic Observational Multi-Center Study, with sensitivity/specificity figures calculated, and you have then the Swedish randomized trial, and currently the ongoing EU project looking at the clinical implementation phase, being equally important.

            So if we look at the Plymouth randomized trial, that was targeted to show safe reduction in operative interventions for fetal distress, which was achieved with almost 50 percent reduction for operative deliveries for non-reassuring fetal status with no increase for failure to progress.

            If we look at the outcome, comparing the heart rate only with the heart rate plus ST, meaning in the STAN, the baby's outcome, whatever parameter we choose, showed signs of improvements.  If you take metabolic acidosis, this didn't become significant almost, but there were three cases in the STAN arm.  They had ST events, but they were not recognized.  So we saw the need to refine the technology with further digital signal processing and automatic ST analysis.

            Because we had the basic physiological findings now verified in a large, randomized, controlled trial, meaning that an ST rise, that is a fetus responding to hypoxia, and biphasic ST, meaning a fetus not fully capable of responding or has not had the time to respond.

            So we didn't want to change this approach with a standard fetal scalp electrode and maternal thigh electrode.  We wanted to have a clear, nice signal to work with.  So we needed to develop the computer, the digital signal processing, the software engineering required, and, in particular, establish what we call the ST log event log, automatic identification of ST events, like in this case for its 1 centimeter a minute, but it is the only one showing that, where you have a heart rate with a varietal deceleration.  Prior to that, there are some biphasic events now being identified on the log, followed by an episodic rise in T/QRS reaching a level, which means that in this case there is a need to intervene.

            What this recording is telling us, that in the first stage of labor that fetus could not meet with the challenge of a contraction, so there would be a decrease in fetal placental function, maybe perhaps now metabolic acidosis at that point in time.  But I can show you that that baby entered second-stage labor, and it developed marked metabolic acidosis and meconium aspiration, and required substantial neonatal care.

            So the guidelines:  Well, we are combining heart rate with ST waveform, recognizing that the heart rate, when it's normal, is a very good marker of a healthy fetus fully capable of adapting.  Then we have the situation of a pre-terminal heart rate trace, which is a trace which shows no variability, no reactivity.  They would say, that is sufficient information, but in between we have a large number of cases where there are non-reassuring fetal heart rate traces and where we now need additional information based on the qualified assessment of ST, meaning sometimes short-lasting, episodic ST events, sometimes baseline rises lasting more than 10 minutes, and the appearance of biphasic ST.  And the computer would guide where one would be with regard to ST events within these blocks.

            Then what to do?  Well, if you're in first-stage labor, obviously, sometimes time is on your side.  So you can intervene.  But if you are in second-stage labor, we learned that clearly sometimes there is an urgent need to do something, and with these guidelines, immediate delivery is recommended in second-stage labor.

            Clearly, when we use a computer, we've got to give it a bit of time to assess the baseline because we are largely looking at changes.  Twenty minutes is the standard that we have established.

            But why this is ongoing?  The user had been taught to recognize these dots on the graph, on the screen of the computer, and also in cases when there is somewhat less signal quality, then we say, well, you had better be able to trust the data, then perhaps the computer.  So the user is taught how to use the data also, regardless of what the computer is saying.

            Looking at the specificity issue here, I want to use different ways to assess that, but the positive predictive value is sometimes useful.  We can now look at how the STAN versus heart rate performs with regard to different cutoffs in cord artery pH.  If we take the cutoff of 7.15, well, in 80 percent of the cases where there are ST indications to intervene, the cord artery pH was less than 7.15, whereas we had a heart rate only in 43 percent of the cases.  And there were significant improvements, regardless of what cutoff you choose.  When it comes to sensitivity, well, if we have metabolic acidosis, in this study of 573 cases, we did identify all of this.

            So we take this forward now to the second randomized trial, this one being multi-center in Sweden, Gothenburg, Lund, and Malmo.  These are all large maternity unit, university departments with between 3,000 and 4,000 deliveries.  We have among us here today Hakan Noren, Ingemar Kjellmer, and Karel Marsal, being part of this study.

            As you may notice, perinatal mortality figures are reasonably low.  Obviously, these are tertiary referral centers, so they would have a little bit higher perinatal mortality figures.  If you look at Caesarian section rates, you will find those vary, as they would do between different centers, depending on the culture of electronic fetal monitoring.

            The objective of the study was to show a reduction in number of newborns born with cord artery metabolic acidosis by at least 50 percent.  We identified pH of less than 7.05 base deficit, extracellular fluid greater than 12.  Now this is not a situation where the baby is severely affected.  Eighty percent of those had perfect normal outcome innately.  But this is a marker of an ongoing anaerobic metabolism, and these are cases that need to be discussed and identified.

            While doing that we did not want to see an increase in the rate of operative deliveries, but rather a significant reduction without any increase for other causes of operative deliveries.

            Design of the STAN monitor, where randomization occurred at power on, which allowed us to collect ST information in those cases being monitored with heart rate only.  So, retrospectively, we can look back to see what ST might have added.

            Training period for two months and an interim analysis of 1800 cases which occurred.  Busy units, a lot of deliveries, but also in this trial the cases were managed by regular labor ward staff comprising more than 300 people that had to be trained and meet and understand the clinical protocol.

            Inclusion criteria:  term, active labor, singleton, cephalic, and the decision to apply a fetal scalp electrode.  Clearly, then there was a need for more continuous detailed information.

            You see here a sample of a STAN recording with a heart rate showing marked decelerations with contractions.  You have a normal ST and a normal outcome with a normal vagina delivery, whereas this is a case showing ST events in association with this deceleration, initially, followed by baseline rise, forcing an intervention, emergency C‑section with a normal outcome.

            So we look at the clinical management of these.  The control group, they were managed according to the standard practice, FIGO Guidelines for interpreting the heart rate, allowing for fetal blood sampling, and that practice varied between the different centers somewhat.  Obviously, if there was a low cord arterial fetal blood sample, intervention was required, as recommended.

            You've seen the STAN guidelines, and just to summarize those once again, in a case of pre-terminal fetal heart rate, immediate delivery regardless of ST; normal fetal heart rate, no intervention regardless of ST.

            In the interim analysis we found that in the STAN arm there were deviations from the clinical management guidelines, in that six babies had metabolic acidosis with STAN changes and clear indications to do something.  When you discuss these cases, well, the clinicians thought perhaps that wasn't that much of a heart rate change, so they still carried on as they were used to.

            That focused an issue of repeated training, which meant continuing case discussions, because now we had samples from their own labor ward settings.  So we can bring those forward and show how the ST information would affect the clinical assessment in a neutral fashion.

            Just to illustrate one of those cases, you have a heart rate showing a somewhat increased baseline, and then the deceleration, but good maintained heart rate variability, and you find an ST rise with a lot here stating a need for intervention, but that wasn't a lot for a normal vagina delivery, and you have a baby born 30 minutes later with a cord artery metabolic acidosis.

            So if we look at the characteristics of this child, we find that there was more prima gravida, more post-term cases.  The epidural analgesia varies a bit, but a lot with oxytocin augmentation.  So, clearly, a need for further information during active labor.

            Now when you develop a technology, you set certain guidelines for how you use it.  You need at least 20 minutes in order to obtain information.  In some cases a recording was less than 20 minutes.  Okay, that was not a recording made according to the specifications of the technology.

            At the same time as we wanted to assess metabolic acidosis, we said there is a need to have data within 20 minutes of delivery, and also in case one should intervene on the basis of intrapartum asphyxia, you need to intervene rather urgently.  So this was agreed.

            But also, if he had severe fetal malformation, as we know that these cases will have an increased risk of encephalopathy, and so on, so there would be some skewness if they were included.

            If we look at the inadequately recorded cases, there were no difference between the two arms of the trial.  With regard to metabolic acidosis, we find that in the control arm four cases where the STAN recording was discontinued.  That would happen because of a failure with the printer, for instance, and then the case would still be monitored, but with a standard electronic monitoring device, and the same would go for those cases in the STAN arm.

            Operative deliveries, well, these cases here with the discontinued recording or shortage recording, they would have obviously been managed on the standard electronic fetal monitoring procedures, but you may say that, yes, with a short or more than 20 minutes to delivery, clearly, the indications were based on the STAN monitor to intervene.  So they should be still within the group being assessed for reduction in operative interventions.

            So we look at the primary outcome, what metabolic acidosis, intention-to-treat analysis, a significant reduction, if we look at it from adequate recording's point of view, then there's further improvement in the figures.

            Now we can also compare the situation before and after, just looking at the impact of increased experience, and we find that with regard to metabolic acidosis, there is a reduction, not becoming significant before, but becoming significant after.

            On the secondary outcome, operative deliveries for non-reassuring fetal status, overall intention-to-treat, significant reduction with no increase in operative intervention indications.  And if we compared before, once again, before, no difference, whereas after retraining intention-to-treat analysis, the difference becomes significant.

            So we can summarize the overall outcome being a 54 percent reduction in metabolic acidosis, 19 percent reduction operative interventions for non-reassuring fetal status.  This result is largely based on what was the outcome of the retraining and past experience of the technology and confidence in the technology.

            Now we are interested in what happens with the neonates.  In particular, if you want to sort of look at the risk here, while you may choose different pH cutoffs ‑‑ 7.00, for instance, is one being used ‑‑ but we know that there is sometimes a problem in obtaining core data, and we wanted to make sure no case was missed.  So we took that went into special care and an experienced neonatalogist went through all these notes ‑‑ this is Professor Ingemar Kjellmer ‑‑ not knowing which arm of the trial the case belonged, and identified 29 cases that he said these are neonates affected by labor.

            We had three cases that died, one in the control arm because of asphyxia, one in the STAND arm because of septicemia, Streptococci Group B, and one because of asphyxia.

            We can have a look at that case where you have ‑‑ once again, we don't set up active pushing in second stage, and ST rise coming with an ST event, and as this now continues, heart rates become much more alarming.  At this point in time the trace is discontinued because of vacuum extraction.

            Then there is a delay for 24 minutes before the baby is delivered.  In that time the baby is monitored by auscultation, and it's claimed that there were normal heart rate sounds, and the baby is born with Apgars 0-0-0 and does not respond to resuscitation.

            But looking now at that enter special care, we had three cases on neonatal seizures, all in the control arm; four cases of increased neuromuscular tone and metabolic acidosis, low Apgar scores, all in the control arm, and then we got some cases of jitteriness or irritability under no metabolic acidosis, and there's one in the control and three in the STAN arm before training.  These we know are babies that are not at risk of long-term sequelae.

            Then you have a group of babies who have metabolic acidosis and predominantly respiratory symptoms requiring special neonatal care, and in this case we had one in the second stage of the trial, which was a case requiring three hours in special care for observation after delivery, followed the mother, but later on developed signs.

            So we saw a significant reduction in the second phase of the trial of the number of babies having an adverse outcome, but we also saw, with regard to cases, moderate to severe neonatal encephalopathy, increased neuromuscular tone, innate diseases, a reduction with an incidence in the control arm of 3.3 per thousand to .4 per thousand in the STAN arm of the trial.

            Clearly, when we look at a new technology, we should look at deficiencies of that technology as well.  Among those 46 cases of metabolic acidosis there were 15, now also including the control and the STAN arm, where while there weren't any ST log statements, 6 of those, yes, there was some reduction in signal quality, but the ST changes were clearly visually there.  Sometimes they just emerge 10 minutes, 15 minutes, before delivering, in which case there's not enough time for the computer to tell.

            Then there are some in the control arm, in particular, where there is also poor signal quality that you couldn't do any ST analysis.  But then there are some where, yes, there are heart rate changes, but they either regard them as not significant second-stage events, perfectly healthy, vigorous newborn, and we say, well, perhaps these are not that significant, and sometimes you may not have data for the last 10 minutes, for instance, in which case, particularly on metabolic acidosis, could emerge.  But then in one case, which was a case that died, Streptococci septicemia, that had a pre-terminal heart rate trace, and we know that these cases they are rare, but they will not display ST events.

            Another way of analyzing data here, looking at potential deviations:  5,000 cases managed by 300 obstetricians/midwives.  So there are 46 cases of metabolic acidosis, 8 in the STAN arm, you might say potential deviations, 8 in the control arm, and we can look at these.  So we have in the STAN arm cases where we say, well, in this case no deviations.  Because there were ST events, an action was done, normal neonatal outcome, a bit of metabolic acidosis, but that we would expect.

            Then there are those where deviations, probably ST events, no, but at the same time not that much of a heart rate change, normal neonatal outcome.  But then, clearly, there are those where deviations occur where we have ST events that were missed, and these are the ones in the initial phase of the trial, apart from this last trace that I showed you, the baby that died with a vacuum and lack of data over the last 24 minutes.

            If we look the situation in the control arm, these are eight cases.  There were some cases where we can say, well, on the basis of heart rate interpretation, there might not have been any deviations.  There were intermediary fetal heart rate patterns, but then these got to ST events, and I believe that is a significant observation, in that if we add ST, we focus on the heart rate change that emerged.  Then we become more accurate in assessing the heart rate and reducing the risk of an adverse outcome.  Clearly, there were those where there were deviations from the standards of clinical practice and a case, obviously, of no signals, then that's clearly another cause of deviation.

            So we can compare two randomized trials, the Plymouth, the Swedish, and you'll find similarities, perhaps a little bit more aggressive on the reduction of operative deliveries, fetal distress in the Plymouth trial.  But if you look at metabolic acidosis, the figures look identical.

            So what to do next?  Well, we have learned a lot of experience, so we want to disseminate this knowledge, focusing on education, training, use and certification, quality control, using cord acid base as routine, neonatal data, index cases, metabolic acidosis.  Take this up for discussion, provide feedback, expert assessment.

            This is now an EU-supported project, initially with 10 Centers of Excellence.  These are all academic units across Europe, currently being 22, who are becoming expert learning their own experience and providing service to their regions in training, education, and followup.

            The aim is obviously to improve the understanding during labor and to document consistent improvements, based on the STAN monitor and a dedicated teaching/training package.  This, I believe, is important, if we have in our database more than 10,000 cases that we can use in a sort of similar environment to our STAN use, to our own experience, before being exposed to real-life cases.

            So now we can look at the experience from STAN being used in routine obstetrics care.  This is now the City of Gothenburg, 16 months' experience.  In September eight STAN monitors were introduced, and these are the rates of usage per month, covering this 16-month period, finishing January this year.  You will see initial rise and then a continuing rise here, approaching more than 30 percent of all term deliveries and active labor being monitored with STAN.

            So let's look at the outcome.  Now we can compare the initial four months period, where the technology was reasonably new, with one year later, October-January, more than 2,000 deliveries over both periods.  We see significant improvements in operative delivery for fetal distress, in the reduction in Caesarian sections, reduction in cord artery metabolic acidosis, coming down to very low levels, and also with neonates having moderate-severe encephalopathy, over the last 2,256 cases, only one baby that had short-lasting neonatal seizures.

            What we also noticed, that in the STAN, in the CTG arm, where heart rate monitoring is taking place, we see extraordinary improvements.  I think this is not that surprising.  If you focus on the monitoring issue, if you bring physiology into this, you improve also the understanding of heart rate reactions.  We see no risk, in fact considerable improvements, in those cases where heart rate monitoring is applied; as you remember, 34 percent being monitored on the STAN.

            We can summarize on the issue of encephalopathy, finding that in the CTG-monitored group over this time period of 11,000 cases with term deliveries, eight cases, three perinatal deaths, all post-term emergency C-section for fetal distress, severe asphyxia and the baby dying, three of intrapartum hypoxia with acidosis, and two post-natally.

            In the STAN arm there are two cases of hypoxia, one that died two hours after birth, associated with an E. coli septicemia, another case with listeriosis.  Yes, there is one case where still ST events were missed and second-stage hypoxia and then natal seizures, and then post-natal, and then some cases with a more unclear origin.

            So we can summarize now where we start with experimental work, engineering developments, clinical research.  Guidelines were developed, tested in randomized, controlled trials.  Then we learned clinical experience, and the need for expert system support, the Swedish randomized trial, and now we can verify in routine clinical care the situation of improved fetal monitoring.

            Intended use.  Well, if approved, the assessment of fetal condition in labor.  Is it a singleton fetus, vertex, at term, and the indication.  Well, as you noticed, though, the indication in the Swedish randomized trial was based on the need for close surveillance, the need for more accurate assessment.  I believe this is what we're trying to bring forward.

            Thank you very much for your attention.

            CHAIRMAN BLANCO:  Thank you, Professor Rosen.  I think you have several other individuals who will be presenting?

            DR. ROSEN:  We will start with Professor Ingemar Kjellmar.

            CHAIRMAN BLANCO:  Please, for the record, remember to introduce yourself, your affiliation, and any conflict of interest.

            DR. KJELLMAR:  Thank you, Mr. Chairman, members of the panel, ladies and gentlemen.  I'm Ingemar Kjellmar.  I recently stepped down as Professor of Pediatrics, Neonatology, at the University of Gothenburg, Sweden.  I've been involved in the ST development for very many years.  I'm acting as a non-paid scientific consultant for Neoventa Medical.  They have paid my fare and my accommodation here, but I have no economic interest.

            I briefly reviewed some quality markers of perinatal care in Sweden.  First, just to give you an idea of the country we are talking about, I have on the scale included Scandinavian peninsula, Sweden in yellow, in comparison to the United States.  You see areawise Sweden is not a very small country, but it is sparsely populated, particularly in the northern part of the country.  We have only 8.5 million inhabitants.  We have roughly 100,000 births a year.

            It's not only the rate of the cord acidemia that is lower in our birth population; it's also the prematurity rate, about 5 percent, and the perinatal mortality rate, that is close to 5 to 1,000; the infant mortality rate, that is below 5 per 1,000 in the national statistics.  About 12 percent of the mothers are immigrant mothers.  In the three cities taking part in the randomized control trial, Gothenburg, Malmo, and Lund, the immigrant mothers are around 20 percent, and they come mainly from the Middle East, from Africa, and from former Yugoslavia.

            Since we are talking about fetal surveillance in term fetuses, the term perinatal mortality rate, based on the two-year material, is 2.6 per thousand with very low rates of intrapartum death.

            We have, since 1972, national guidelines for resuscitation for asphyxiated babies.  This is the latest pamphlet from 1997 that is widely disseminated among midwives, obstetricians, neonatalogists.

            Now with a country where part is very sparsely populated, one needs to have a program of regionalization.  This slide I have included to demonstrate that we do have such a program.  The delivery hospital level refers to the equipment and the staffing of the hospital.  The hospital in the Category 1 are the very small, rural hospitals with obstetrics service only, no neonatal service.  The Category 2, the district hospitals with daytime neonatal service.  The Category 4 are the big regional centers, and the Category 5 are the big university centers that are fully equipped.

            As you see from the first recordings, this is material covering 12 years of deliveries in the country, about 1.5 million deliveries, about 9,000 stillborn and neonatally dead babies.  You see the expected progression of the risk for cases accumulated in the grade 4 and 5 hospitals.  These are odds ratios.  You see that this is particularly effective in cases where you can foresee the events such as maternal disease and fetal disease, where there is clear-cut gradient from the poorly equipped to the very well-equipped hospitals.

            But for conditions that are difficult to foresee, such as obstetrical complications and asphyxia in labor, the regionalization program apparently is not enough to rid the smaller hospital of these problems.

            This slide demonstrates one often-discussed final outcome variable when it comes to intrapartum asphyxia, namely, cerebral palsy.  It demonstrates the rate of cerebral palsy in the population of west Sweden following in four-year cohorts, all the way from 1954 up to 1994, and followed, moreover, by the same team of investigators.

            As you see, there is a rather big fluctuation in the total incidence of cerebral palsy with a very promising decline during the years when perinatal care was first instituted in my country, but then a gradual increase that took place from the middle of the sixties up to the mid-eighties, and after that at least a tendency toward a drop.

            If this material is split crudely into term and pre-term deliveries, and both of them are counted on the basis of the total population, so these are not gestation age-specific incidence rates, you see that most of the fluctuations are explained by changes in the pre-term population, which is not very surprising.  This is what has been demonstrated from many other countries.

            But the CP incidence in term babies stays, has very little changes over this 40-year period.  Mind you, this is the period when electronic fetal monitoring was introduced during the seventies in west Sweden, where modern neonatology gradually was introduced from the beginning of the sixties.  We have been, like many others, quite successful in effecting the outcome rates in the premature baby, but when it comes to the term baby, very little has been achieved.

            The latest cohort, where brain neuro-imaging is available, demonstrates that about 28 percent of all the term babies with cerebral palsy might be ascribed to intrapartum asphyxia.

            Thank you.

            CHAIRMAN BLANCO:  Thank you very much.

            DR. MARSAL:  Mr. Chairman, members of the panel, ladies and gentlemen, my name is Karel Marsal.  I am Professor and Chairman of the Department of Obstetrics and Gynecology, the University of Lund, in Sweden.  I am an unpaid scientific advisor of Neoventa Medical, and the company covers my travel and accommodation expenses.

            In the next few minutes I would like to address the question of similarities and differences between the use of fetal monitoring in labor in the United States and in Sweden.  I never worked as an obstetrician in the United States, so I based my information on the published statements and recommendations from the American College of Obstetrics and Gynecology and on my contacts with American colleagues.

            If you look at the staff providing during labor in the United States, this is done primarily by obstetricians.  In uncomplicated pregnancies and labors, the midwives are also involved.  In Sweden, by recommendation of the Ministry of Health, the uncomplicated labors and deliveries are conducted independently by midwives.  It's their task to recognize deviation from normality and to call upon obstetricians who in that case will take over responsibility for these patients.  The complicated pregnancies and the complicated labors and deliveries are being conducted by obstetricians.

            I mentioned the technical bulletin published by the American College of Obstetricians and Gynecology, the recommendation on the fetal monitoring during labor.  In Sweden the Swedish National Board of Health and Welfare published in their database of medical facts recommendations for the clinical management of uncomplicated labors.  It includes the use of fetal monitoring.  In addition, there are two reference books which are used by all obstetricians and midwives in Sweden, and there is rather big consensus with regard to the use of the fetal monitoring in labor.

            Looking at the definitions of hypoxemia, hypoxia, and asphyxia, I have found that these are identical in both countries.  The fetal heart rate evaluation during labor is recommended for all patients both in the United States and in Sweden.

            In pregnancies with absence of risk factors, auscultation is accepted as the only way of fetal monitoring both in the United States and Sweden.  However, more and more often the electronic fetal monitoring is applied also in the low-risk pregnancies.

            In the United States, one-to-one nurse-to-patient ratio is recommended in the cases of auscultation.  In Sweden the ratio is about one to two or three patients.

            During the active phase of the first stage, the control of fetal heart rate is recommended in this stage at least every 30 minutes after a contraction, and during the second stage at least every 15 minutes.  In cases with risk factors present, the frequency of recommended controls is higher, risk control every 15 minutes during the first stage and every 5 minutes during the second stage.

            In Sweden, during the active stage of the first stage of uncomplicated pregnancy and labor, auscultation for a minimum of 15 seconds every 15 minutes after a contraction is recommended, and during the second stage with active pushing after each contraction.

            The electronic fetal heart rate monitoring is performed using the commercial monitors of similar makes both in the United States and Sweden.  There is a difference in the speed, which is 3 centimeters per minute in the United States and 1 centimeter per minute in Sweden.  Also, the heart rate scale differs, being 30 to 240 beats per minute in the United States and 50 to 210 beats per minute.

            The recommendations for the use of scalp electrodes as been mentioned in Professor Rosen's description of the Swedish study, and these recommendations are practical identical in both countries.  The documentation, either as a paper strip or electronically, is also similar in both countries.

            The absence of risk factors, when using electronic fetal monitoring, the elevation of the heart rate is recommended in the United States every 15 minutes during this first stage and every 5 minutes during the second stage.  In Sweden, practically all pregnant women admitted to the labor ward are using the admission test, the recording of the heart rate for 15 to 20 minutes.  Afterwards, in those uncomplicated cases, intermittent recording is done every two to three hours and, alternately, auscultation may be used.  During the second stage a continuous recording of the heart rate is recommended during the pushing stage.

            In cases with risk factors present, the recommendation to use the electronic fetal heart rate monitoring is identical, either continuously or intermittently, depending on the degree of risk, and this is up to the obstetrician to decide.

            Looking at the definition of the fetal heart rate patterns with regard to the baseline variability accelerations and decelerations, these definitions are identical in both countries.  The overall definition of the fetal heart rate patterns is in the United States, according to the guidelines of American College of Obstetrics and Gynecology, divided into reassuring or non-reassuring fetal heart rate patterns.  In Sweden we are using the fetal classification, classifying the records into four classes:  the normal one, suspect, pathological or abnormal one, and pre-terminal.  These three ‑‑ suspect, pathological, pre-terminal ‑‑ would be classified as non-reassuring in the United States.

            The management of such non-reassuring patterns is very similar in both countries four steps recommended to find the etiology of the abnormality, attempt to correct the pattern, to analyze the fetal blood sample for the pH, and if indicated, to perform operative intervention.  The operative interventions used are emergency Caesarian section or the instrumental vaginal delivery, comprising both forceps and/or vacuum extraction.  These indications are identical in both countries.

            So, to summarize, after looking on these similarities or differences in the use of fetal monitoring during labor, I came to the conclusion that the United States and Swedish national guidelines are very similar, including even liberal use of electronic monitoring in normal pregnancies.

            Thank you for your attention.

            CHAIRMAN BLANCO:  Thank you.

            DR. DEVOE:  By way of introduction, I am Dr. Lawrence Devoe.  I'm Professor and Chairman of Obstetrics and Gynecology at the Medical College of Georgia.  I'm also a consultant for Neoventa, who paid for my expenses and travel to attend this meeting.

            I would like to kindly thank the panel and its Chairman for their attention for the remarks that I am about to make, which deal with the issue of where we are with intrapartum fetal monitoring in the United States at the present time and the case for new technologies such as the STAN S21 system.

            My objectives are to give you a snapshot of the current situation of affairs with monitoring the U.S., where the current problems and limitation are, and possibly where the STAN system might fit in to be a useful adjunct.

            As we know, back in the 1960s, when fetal monitoring first arrived on the scene, it had some very discrete and laudable goals.  That was to screen for intrapartum fetal asphyxia, to detect fetuses at risk or in early stages of compromise, to prevent fetal neurologic injury or death without, on the other hand, an excessive maternal risk from obstetric interventions.

            But this is a summation of about 90,000 patients that were collected in the first observational, nonrandomized reports in the U.S. literature that emanated from the late seventies and early eighties.  Looking at these, at that point in time, painted a rather sanguine picture for the use of electronic fetal monitoring with substantial reduction in intrapartal stillbirths and neonatal deaths compared to the lack of monitoring or auscultation.

            As recently as 2002, going to the Cochrane Library, Thacker has updated the meta-analysis of the randomized control trials, which currently number 13, that met their criteria and assume about 60,000 pregnancies comparing electronic monitoring with or without scalp sampling for pH with auscultation.  Looking at the outcomes, both maternal interventions and infant outcomes of Apgar ‑‑ NIC admissions, seizures, and death ‑‑ to cut to the chase, this meta-analysis direction hasn't changed very much since it was first published in 1995 and suggests that the use of electronic monitoring is associated with a higher rate of Caesarian and operative deliveries, similar perinatal death rates, and the only group in which there apparently was benefit and reduction of neonatal seizures was a group in which EFM was backed up with fetal scalp sampling.

            The question one might ask is, why were the randomized control trials not as good as the early nonrandomized trials?  The answers to that I think are very obvious.  The early nonrandomized control trials had less antepartal screening and, consequently, fewer compromised babies were excluded from labor.

            The care in randomized control groups is very intense, featuring to one-to-one bedside nursing care, and in turn, during the period in which these randomized control trials evolved, the neonatal intensive care unit performance improved.  Consequently, there were fewer neonatal deaths and much larger groups than those included in almost all the trials would have been needed.

            There are three issues that really affect the performance of electronic fetal monitoring today in the United States.  These are well-known.  They are the issues of observer reliability and reproducibility, the issues of fetal heart rate interpretation per se and its clinical correlates, and the use of ancillary assessment methods.

            As far as the first issue, many studies have already shown wide ranges of intra- and inter-observer agreement, and the lack of observer reliability applies to visual and auditory identification of fetal heart rate patterns.  This is a summation of best-case studies which is inter-observer studies simply featuring the same tracings being presented to individual observers at different times and showing that cap levels of agreement vary widely, but definitely fall far short of the gold standard of 1.0.

            Things are clearly much worse when one gets to the inter-observer agreement, and the parameters of fetal heart rate baseline, such as rate, accelerations, decelerations, and variability, all degrade as more observers are brought into the mix.

            The reliability of fetal heart rate assessment also decreases as you ask the observer to discriminate more different types of patterns or categories.  In fact, even applying a standard criteria with standard terminology to visual inspection may not improve reliability or agreement.

            The targets for electronic fetal monitoring are to make some assessment of fetal oxygenation and acid-base balance, as well as try to predict cases in which these are disturbed, and endpoints have been used such as scalp or cord blood gases in neonatal outcomes, as you have already heard.

            Going back, in fact, as early as the late 1960s, it became very obvious that certain patterns of fetal heart rate monitoring tended to be more closely associated with lower pH values, and these particularly were persistent, severe, variable, or late decelerations.

            But we look at two retrospective studies which were published in the 1980s.  Both of these identified that the main significant alteration in the fetal heart rate tracing, when examined visually, was the occurrence of late decelerations which was associated with an increased risk for metabolic acidosis in the study published by Low, and also an earlier occurrence of decreased fetal pH in a study published by Fleischer in the same time period.

            As far as pathophysiologic correlates of intrapartal fetal heart rate patterns, there have been numerous retrospective and few prospective studies that have ensued in the last 20 years with limited sample sizes.  In the main, the fetal heart rate pattern assessment has been visual.  There have been various studied conditions as well as interpretative criteria, and few fetal heart rate features specifically stand out, that is, late decelerations and the absence of variability, that correlate well with current fetal acid-based status.

            There are current standards of care that have been published, and you have already heard some of these bulletins referred to, both in the United States and elsewhere.  I want to call your attention to what we currently refer to in the ACOG Technical Publication of 1995, and specifically direct your attention to the issue at hand of the management of persistent non-reassuring fetal heart rate patterns.  Professor Marsal has already indicated the four-step algorithm for attempting to determine the etiology of the pattern, to attempt to correct the conditions that produce this pattern, and if this fails, consideration of a fetal scalp blood pH, and finally to determine if and how urgently operative intervention is needed.

            If the fetal heart rate pattern persists, intervention depends on the clinician's assessment of the likelihood of severe hypoxia and metabolic acidosis, as well as the estimated time to spontaneous delivery.

            Now we have accepted that persistent, late, or severe variable decelerations with absent variability always are non-reassuring and require prompt intervention unless they spontaneously resolve or are rapidly corrected.  Spontaneous accelerations virtually assure the absence of fetal acidosis.

            Ancillary methods, such as fetal scalp stimulation or vibrio acoustic stimulation, can be used to induce accelerations, and also, when positive, indicate the absence of acidosis, but in nonresponsive fetuses, assessment of scalp blood pH may be used to clarify acid-based status.  Frankly, this technique, while occasionally helpful, is used uncommonly in most current obstetric units, including tertiary centers.

            When fetal heart rate patterns are persistently non-reassuring and acidosis is present or, as importantly, cannot be ruled out, the fetus should be prompt delivered by the most expeditious route.  This current status I believe places clinicians in a dilemma.  Most non-reassuring fetal heart rate patterns lack precise correlation with current fetal status or prognosis.

            Secondly, there are limited adjunctive tools to evaluate such patterns, not all of which are either practicable or available.

            Third, the fetal heart rate feature best correlated with local myocardial hypoxia and, ultimately, metabolic acidosis is the evolution of repetitive late decelerations.

            The width of the window for optimal intervention for this pattern, that is, late decelerations, may be unclear in any given fetus.

            Fifth, the education and experience of bedside caregivers on fetal heart rate interpretation, technology, and intrapartum management of abnormal fetal heart rate patterns varies widely across this country.

            Sixth, the clinician is left to determine the best management with limited supportive data, and supported by better analyses, these results in a probable excess of risky obstetric interventions without demonstrable perinatal benefits.

            I was privileged to be part of an NIH research planning workshop which was conducted between the years 1995 and 1996.  The thing that spurred this workshop on ostensibly was the lack of consensus on definitions and nomenclature of electronic fetal monitoring, but really behind all this was the knowledge that the meta-analysis and the results of the randomized control trials were disappointing to those of us that had been proponents of electronic monitoring for two decades.

            This workshop put together 22 clinical experts and about half a millennium of clinical experience in electronic fetal monitoring, and arrived at standardized unambiguous definitions for visual interpretation of fetal heart rate.  It also made recommendations for further research.

            The study areas that came out of this were to address these very issues of reliability of visual interpretation of fetal heart rate pattern, the validity of pattern interpretation when viewed in its linkage with adverse CNS outcomes, and to determine whether the causal relationship between fetal heart rate outcomes could, in fact, be affected by interventions using fetal heart rate interpretation as a clinical guide.

            It came to four main conclusions:  that no consensus on strict guidelines for management using fetal heart rate patterns alone could really be recommended until evidence-based algorithms could be developed.

            Secondly, that normal fetal heart rate patterns do exist, that they confer high likelihoods of well-oxygenated fetuses at the time of observation.

            Third, there are several patterns that predict current or impending fetal asphyxia with increased risk of injury to the neurologic system or death.

            Fourth, there are many fetuses, possibly as many as 30 to 40 percent of those being monitored, that fall between these extremes, and this group would experience the most benefit from the results of the recommended research.

            So that brings us to what I think we have to look at to improve the status of electronic monitoring today in the United States.  With very few exceptions, the kind of research to move the subject along has been lacking since the 1980s.  In fact, there have been new technologies that have arrived.  We obviously used from time to time vibrio acoustic stimulation, but this particular intervention is not useful if the fetus fails to respond.

            Biophysical profile is cumbersome and has not been shown to be helpful during labor.  Doppler has simply been inadequately evaluated in labor.  There have been three more recent approaches using computerization of fetal heart rate, fetal oximetry, and ECG-waveform analysis.

            Going to computerized online evaluation of fetal heart rate, there have been a number of studies that have been published in the last decade.  These have all been somewhat limited in number.  Although computerization may yield a more reliable and precise analysis of fetal heart rate compared to visual assessment, thus taking the observer problem out of the equation, it has not been shown to improve the detection of acidemia or fetal compromise.

            Fetal oximetry has also recently appeared.  The FDA trial, as you are well aware, reduced intervention for non-reassuring fetal heart rate cases, but did not reduce the overall Caesarian section rate.  There are still documented signal capture failures.  The best targeted use for this technology has not been fully identified.  The cost-benefit ratio was unclear, and products are in play now that are adaptable to current electronic monitoring systems.

            I summarize the ACOG Committee opinion on oximetry simply in this single paragraph and emphasize the points that Dr. Artal already articulated in reference to the material that you've heard this morning.

            What is the target for electronic monitoring?  I think it is really unrealistic to think that a single tool, unaided, is going to make any difference in eliminating hypoxia ischemic encephalopathy or cerebral palsy.  This is still, fortunately, a rather infrequent event, and the sample sizes to really clarify this need to be humongous.

            The reduction in profound acidemia or asphyxia with a pH falling below 7.10 or 7.05, we don't know how often this occurs in the United States, but this reflects a 2.5 percentile of births, somewhere between 10,000 and 100,000 births per year.  This was achieved in the Swedish STAN trial.  The reduction in newborn seizures, a rather limited surrogate for later performance in neonatal and childhood outcome.  Again, this was the only group that was shown to be affected when using electronic monitoring backed up with some ancillary adjunctive guide.

            Fourth, the avoidance of unnecessary intervention, Caesarian deliveries or operative deliveries, and this was also achieved in this latest STAN trial.

            STAN S21 addresses the problem areas I think of acidosis and OB interventions.  It has soundly founded in a scientific basis which has been established for higher-level FECG analysis.  There now is real-world experience and high-quality clinical obstetric units that show that this technology mated, importantly, with responsible oversight and education, is applicable and practical.

            Other advantages is that it comes into an arena where there is always already some familiarity, that is, using a standard fetal heart rate monitoring device and using signals that are inherent in the fetal ECG with a familiar probe.  There is concurrent display with real-time events, making alerting to clinicians possible in time to change the natural history of a labor.  It is objective.  It is reproducible data display which deals with a thorny and still unresolved issue of observer problems, resulting in visual estimation alone.

            It is an adjunct for problematic fetal heart rate patterns, which is a most desirable target and goal.  It reflect myocardial events that are affected by hypoxia and acidosis.  It is a conservative system, and that is, a conservative system in computer terms requires both good input and requires some time.  Both of these are important prerequisites.

            The educational context which embraces this system is added valued.  I think as you saw from the Gothenburg study, in a real-world situation in a busy clinical unit, users got better at doing monitoring overall, whether they used the STAN system or conventional monitoring, largely I think boosted by the education process that is required for the use of this system.  As importantly in busy obstetric units, it is adaptable to labor units with different provider levels of expertise and experience.

            I just want to conclude by making four statements:

            Auscultation, I think we really have to bite the bullet and accept the fact that it is an impractical approach in today's environment.  It is labor-intensive.  It requires one-to-one nursing.  This is a level that is hard to achieve in almost any obstetric unit in the United States today.

            I have alluded to, and I'll iterate, that unaided visual assessment of fetal heart rate tracings remains problematic for a variety of variance issues.

            Computerized processing of raw fetal heart rate data might be able to improve point No. 2, but still by itself gives a limited picture of fetal adaptation.

            And, finally, FECG analysis moves one step closer to the intrapartum fetal adaptive or maladaptive responses when hypoxia is present.

            Thank you very much.

            CHAIRMAN BLANCO:  Thank you, Dr. Devoe.

            Anything else, Dr. Rosen?

            DR. ROSEN:  I'm fine with this presentation, and thank you very much for your attention.  Thank you.

            CHAIRMAN BLANCO:  Thank you very much.  Thank you to all of the speakers.

            It is now, by the official clock, 9:45.  We're going to take a 15-minute break, and we will reconvene promptly at 10 o'clock.  Thank you.

            (Whereupon, the foregoing matter went off the record at 9:45 a.m. and went back on the record at 10:01 a.m.)

            CHAIRMAN BLANCO:  All right, if we could go ahead and reconvene?

            I think that we will start at this point with presentation by the FDA, and I believe that Ms. Daws-Kopp will be starting the presentation.

            MS. DAWS-KOPP:  Hi.  Good morning, ladies and gentlemen, distinguished panel members, and guests.  I'm Kathy Daws-Kopp, the lead reviewer for FDA on this PMA, and I'm here to give you a brief overview of the review process we've gone through on the PMA.

            This is a PMA for a specialized perinatal monitor that includes a feature for ST analysis of the fetal ECG.

            First, I would like to acknowledge the review team.  As you can see, a number of people have been involved in the review of this PMA application in the areas of clinical, statistical, epidemiology, software, hardware, fire research, monitoring, and manufacturing.

            Before I go over our efforts in review of this file, I am going to give a brief overview of our interactions with the company leading up to the PMA submission.  We met with the company in 1999, when they briefed us on the Plymouth trial and the Swedish randomized controlled clinical trial.  At that time the Plymouth trial was completed, and the Swedish RCT was ongoing.

            In December of 1999, the sponsor submitted a plan to start a modular PMA review where the company is allowed to submit portions of the PMA information early in an attempt to address some requirements of PMA submission prior to submission of clinical data.

            In this slide you will notice that one modular for general information is closed, a term we use with modular submissions to indicate that the issues addressed within the module have been addressed to our satisfaction.  However, the prior clinical studies and device description and non-clinical testing were not closed, and those reviews were rolled into the PMA review.

            I'm going to give a brief overview of the device design, and I'll discuss what we have looked at in the course of our review.  I will finish with an outline of ongoing issues that we're still working with the company to resolve.

            For design, I'm going to talk about both the components of the device and the mechanism of action.  As previously mentioned, this device is a perinatal monitor with ST analysis capabilities.  It has many of the features of a typical perinatal monitor.  The STAN system is made up of the following basic components:  base unit, monitor, software, and sensors.

            The base unit has as a main electrical component that serves as a conduit for input and output connections such as those for data storage or access and accessories to the device, including the sensors.

            The monitor is the display.  Software includes all the software for the standard monitoring functions as well as the ST analysis features.

            The next slide I will address the sensors.  Spiral electrode is used for the ST analysis feature and for the fetal heart rate.  The STAN system does not have external fetal heart rate doppler ultrasound capabilities.  This device does have the capability for both internal and external uterine activity monitoring.  The event marker is used by the patient to mark fetal movement.

            As previously mentioned, the device uses the fetal spiral electrode to obtain a fetal ECG signal.  This signal is an analog electrical signal that is converted to digital by the device.  The ST analysis software manipulates the signal to create a smooth wave form.  The software evaluates this wave form for the three types of events the device is looking for:  episodic T/QRS rise, baseline T/QRS rise, and biphasic ST.

            As you have seen, the device displays the T/QRS ratio with "X's" below the uterine activity on the strip chart, with events annotated next to the "X's."  The smooth ECG complex is also displayed.  It should be noted that the device will not identify an event in the first 20 minutes of operation, so that the device can set baseline.

            This slide lists the things that we look at during our review.  For software and hardware, we look at safety and effectiveness.  Examples of safety issues for software and hardware include electric shock, EMI shielding, and explosion hazard.  Examples of effectiveness are adequate identification of the ECG waveform and ST events and poor signal quality.

            We specifically look at requirements in testing.  We check to see that the device is designed to do what the sponsor and manufacturer says it will do, and we look to see that the tests check to see that the device works the way it is supposed to.

            We also look at biocompatibility and sterilization or disinfection of the materials that contact the patient.  For bioresearch monitoring, we look at study execution, including recordkeeping, informed consent administration, et cetera.  For manufacturing, we look at compliance with design controls.  Both include an inspection.

            Bioresearch monitoring inspects clinical sites as well as any records related to the conduct of the trial at the sponsor's facility.  Manufacturing conducts an inspection at manufacturing facilities.  A bioresearch monitoring inspection is common for clinical trials supporting the PMA.  The manufacturing inspection is required.

            Dr. Corrado, whose presentation follows mine, will address clinical and statistical reviews during her presentation.

            One of the review issues we're addressing is that of signal quality.  Here the device monitors signal quality, providing an indication on the display when signal quality is poor.  The sponsor provided the information regarding an ongoing retrospective visual assessment of signal quality being conducted on the strip charts from the Swedish randomized controlled clinical trial.

            This partial assessment identifies the percentages of low, medium, and high quality signals in the study as defined by the size of data gaps.  That is, for ST information, signals with data gaps larger than four minutes are considered low quality, and signals with maximum data gaps of two minutes or less are considered high quality.  It is not clear whether the device uses a similar criteria for its signal quality indicator.

            The labeling and training provide some guidance on management of data gaps such as advising the user to reapply the fetal ECG electrode or to do a manual assessment of the ECG complex, that is, to estimate the T/QRS ratio by eye.  We haven't finished this review and will work with the sponsor and expect that we will obtain the descriptive information and test data we need.

            I would also like to mention another aspect of signal quality because it relates to the analysis of the data for primary and secondary endpoints.  Analysis II, which Dr. Corrado will discuss in her talk, excluded cases that did not meet the sponsor's requirement for adequate recordings.  The database includes reasons why specific cases did not meet these clinically-based criteria.  A large portion of these excluded cases, about 42 percent, were to attributable to signal or technical problems.

            As mentioned in my last example, our hardware and software reviews are ongoing.  We currently have some information from the sponsor that describes the device design requirements and verification/validation.  As I outlined in my previous slide, we don't have as many details about the design of components, features, and algorithms for the device as we would like to answer all of our questions about how the device works.  Likewise, we would like further information on verification and/or validation testing that was done to confirm performance issues that may not come out in the clinical testing.  We are still in the process of describing our needs in this area to the sponsor.

            Bioresearch monitoring and manufacturing reviews are also still ongoing.  The inspections in Sweden have not yet been scheduled.  We will continue to work with the sponsor on these issues.

            Now I will turn the floor over to Julia to discuss clinical and statistical findings of our review.

            CHAIRMAN BLANCO:  Thank you.

            DR. CORRADO:  Thank you, Kathy.  Good morning, everybody.

            As we all know by now, the subject of this PMA is the STAN fetal ECG monitor, and its use is that of fetal ECG analysis and that includes standard fetal heart rate plus ST waveform data to obtain information on the impact of labor on the fetus, to evaluate the status of the fetus during labor.

            The sponsor's proposed indication for use is for a term fetus in vertex presentation who is a singleton, and there are additional criteria, all of which contribute to an indication for fetal scalp electrode monitoring.

            I'm going to cover the following subjects during my presentation.  I'm going to emphasize the results of the Swedish randomized controlled trial.  I'm going to talk about clinical issues that arose during our review of this trial, and then I'm going to summarize some other clinical experience with the STAN monitor in European countries, and then try to bring us to focus on the questions FDA has identified for the panel.

            Very briefly, as Dr. Rosen indicated, there have been multiple studies of the STAN monitor in European countries.  The Plymouth study was in England.  There was a multi-country trial called the Fetal ECG Analysis During Labor Trial that was relatively small.  There was then the larger Nordic Observation Multi-Center Trial.  The Swedish RCT is what we're going to be talking about, and then there are data from the City of Gothenburg, which she has summarized and I will summarize very briefly as well.

            I am afraid that I accidentally hit the "end" button.  Sorry.  I'm sorry.  I'll be real careful not to press that "end" button there.

            Finally, I wanted to highlight that there's an ongoing study that is sponsored the European Union.  It involves obstetrical units in 10 different countries, and the primary focus is the training and clinical results from a Center of Excellence and dissemination of knowledge program.

            So the point I've just tried to make with the last couple of slides is, although we're going to be focusing on the results of the Swedish RCT, there is a very large body of data from this device that has been accumulated over the last 10 years, clinical data.

            One of the reasons that we focused on the Swedish RCT is that we felt that it met the criteria that we look for in establishing relatively more value to certain types of clinical studies.  This was the prospect of randomized multi-center controlled study.

            It took place during a time period of December 1998 through June of 2000.  Forty-nine hundred and sixty-six subjects were enrolled at three university hospitals, relatively equally throughout those three centers.  It involved a comparison of the STAN monitor, which, again, includes both fetal heart rate plus the analysis of the fetal ECG versus fetal heart rate only data.

            I would like to point out, though, that in the control arm the STAN monitor was used.  It is just that the ECG data was not available to the clinician while they were taking care of the patient.  So they just had essentially fetal heart rate plus uterine contraction pattern.

            The objectives of that study were to reduce perinatal morbidity as identified by cord artery metabolic acidosis; secondarily, to evaluate the use of STAN protocols and clinical guidelines in practice.  That will be the focus of several of our slides that we will get to.  Then, finally, to reduce operative interventions.

            The endpoints, slightly different from the objectives, were several.  The study endpoints that I'm going to focus on are the primary outcome for the study, which was metabolic acidosis.  It is important to keep in mind that that was defined by a cord artery pH of less than 7.05 with a base deficit of greater than 12.  In order to calculate and analyze this data, both cord arterial and venous samples were required to ensure that the data, the arterial data, were reliable.  Neonatal morbidity and frequency of operative interventions, including Caesarian sections, were also evaluated.

            I would like to just give an overview in terms of safety.  FDA looks at safety and efficacy, as everyone here knows.  In terms of safety, I wanted to make a couple of kind of broad statements.

            The safety of the device with respect to the specific components actually did not bring up any new issues because these components are already in use in other devices.  However, the safety implications of interpretation of the STAN data and resulting clinical management are part of our safety evaluation of this device.

            For efficacy, the sponsor did two analyses, referred to as Analysis I.  That was an intent-to-treat analysis.  They also did an Analysis II in which they only looked at what they considered to be adequate recordings.  I will take us through a patient tree and, hopefully, make it clear to you all what these two analyses, what patients these two analyses involved.

            This is the Analysis I.  This is the intent-to-treat analysis.  Forty-nine hundred and sixty-six subjects were enrolled and randomized more or less equally to the two study arms that I describe.  As you see, 2,519 were enrolled in the STAN arm, 2,447 in the control arm.

            For the primary endpoint, which was metabolic acidosis, it is important to keep in mind that around 700 subjects total could not be evaluated, and the reason was that the cord blood sample, for some reason, was considered inadequate in some way for analysis.  Again, recall that you had to have both umbilical cord artery plus umbilical cord vein blood, and they had to meet certain criteria for acceptability.  Therefore, for the primary endpoint of metabolic acidosis, somewhat fewer than the enrolled numbers were actually evaluated:  2,159 in the STAN arm and 2,079 in the control arm.

            This is just a quick slide that shows the Analysis I.  This is an important slide because this is essentially the primary endpoint, the results of the analysis for the primary endpoint for the study.  As you see, for metabolic acidosis, for the intent-to-treat population minus those whose blood samples were considered inadequate, 15 out of 2,159 subjects in the STAN arm met the definition that was prospectively identified as constituting metabolic acidosis, whereas 31 of 2,079 in the control arm did.  This boiled down to 0.7 percent in the STAN arm and 1.5 percent in the control arm, and this did meet the test for statistical significance.

            As we move down the list and we look at operative deliveries for fetal distress, now that includes vacuum, forceps, as well as C-section deliveries, and you can take a look at the numbers there.  There was a reduction in numbers of operative deliveries for fetal distress in the STAN arm of the study, and that also met statistical significance for the intent-to-treat population.

            We looked at Caesarian sections for fetal distress for the intent-to-treat population, and although there were fewer relative numbers in the STAN group, this did not meet statistical significance for the intent-to-treat analysis.

            I would like to return to the patient tree, and moving through, I would like to just point out that sort of the lower half, the middle and lower half of the diagram represents Analysis II, and that was the analysis that was done for what were considered adequate recordings.

            As you can see, approximately the same number of subjects were excluded in each arm of the study.  I will go into the reasons for that in a subsequent slide.

            The point here is that, for Analysis II adequate records, the numbers of subjects evaluated are less by around 280 or 290 in each group, approximately, again, the same numbers on both sides of the study.

            The reasons for inadequate recordings were as follows, and we've given percentages, so you get an idea, of those subjects who were excluded, about what percent fell into different categories of explanations.  A very, very small number of congenital malformations, only 1.3 percent.  Thirty-seven and a half percent were excluded because the tracing, the labor didn't meet the requirement for at least 20 minutes on the STAN monitor.

            More, a larger percentage, 56 percent, were excluded because more than 20 minutes had elapsed between removal of the scalp electrode and delivery.  Then there was a small category of other reasons for excluding cases for inadequate recordings.

            So these are the results.  We're familiar with this.  These were the primary and secondary endpoints that were looked at.  Metabolic acidosis became even more statistically significant after the subjects were excluded.  The P value decreased.  Operative deliveries for fetal distress also improved in statistical significance.  For Analysis II there was a statistically-significant reduction in Caesarian sections for fetal distress in the STAN group.

            I guess what I would like to do right now is very briefly ‑‑ the sponsor didn't speak to this issue of Caesarian section for fetal distress.  We did an analysis of this because we were very interested in why that became statistically significant after cases were excluded.

            Very briefly, on this slide what you see is that although the total numbers of Caesarian sections for fetal distress in the intent-to-treat population, and even after excluding inadequate recordings, was relatively small and uniform across both arms of the study.  However, among the population of patients who were excluded due to inadequate recordings, that population seemed to be enriched in subjects who ultimately went on to Caesarian section for fetal distress, and there were more in the STAN arm, which might help explain the outcome.

            We don't know the cause of this.  We, nevertheless, thought it was interesting to note that for some reason that excluded population seemed to relatively concentrate those subjects who went on to C‑section for fetal distress.

            We, at this time, would just like to point out possible significance of the definition of metabolic acidosis in this study.  The sponsor's prospectively defined criteria were a pH less than 7.05 and a base deficit of greater than 12.  This was established prior to the carrying out of the study.  So this was an entirely prospective definition.

            We thought it would be interesting to stratify the results for different pH cutoffs, for example, ACOG in a 1995 technical bulletin described the pH of less than 7.00 as possibly being significant with respect to fetal hypoxia.  When you look at the differences between the STAN arm and the control arm at that pH level, 7.00, you see that the differences between the two arms are less impressive.  For the 7.05 cutoff, it was 15 and 30 versus 31 in the control arm; for the pH cutoff of 7.00 the numbers were 11 and 15.

            Let me also add that, in all fairness, the Swedish RCT was not powered to detect the statistical difference between the two groups at a pH cutoff of 7.00.  So in some ways it is unfair to apply such a criterion.  Nevertheless, we think it's just worthwhile noting.

            With respect to adverse events in the study, there were 29 adverse outcomes.  Now this excludes one death in each arm in fetuses who had severe anomalies.  Therefore, excluding those two cases, there were three deaths in the study, either intrapartum or early perinatal deaths.  There were 11 babies with mild, moderate, or severe encephalopathy.  Now that includes the babies with and without metabolic acidosis, and I'll talk about that a little bit more as I go on.  Fifteen other babies were admitted to the Sick Care Baby Unit with metabolic acidosis and other symptoms.

            I'm just going to focus on moderate and severe encephalopathy because I think that those are probably, I believe that those are the ones that are most clinically-relevant here.  It is very easy to see that there were no cases of modern encephalopathy in the STAN arm of the study.

            I again want to point out that I'm making a distinction here between babies who had confirmed metabolic acidosis and babies who either did not have metabolic acidosis or the cord blood data just weren't adequate to confirm metabolic acidosis.

            In the control arm there were three infants who had confirmed metabolic acidosis plus moderate encephalopathy.  There was one additional infant in the control arm who had modern encephalopathy, but the cord artery data were not available.  So we don't know whether or not that encephalopathy occurred in conjunction with metabolic acidosis.  So there is a difference in these two arms of the study for modern encephalopathy.  It was zero in the STAN arm and 4 in the control arm.

            Regarding severe encephalopathy, there were three cases in the control arm of severe encephalopathy.  There were none in the STAN arm.  I think it's important to note, with regard to severe encephalopathy, that none of those cases in the control arm had confirmed metabolic acidosis.  As a matter of fact, two of the cases had evaluable cord blood and those infants did not have metabolic acidosis.

            One of those infants underwent a mid-cavity vacuum delivery.  This was, I believe, a prima gravida subject who ‑‑ and the baby turned out to be a little bit over 4,000 grams.  Unfortunately, a shoulder dystocia was encountered, and about 8 to 10 minutes elapsed before delivery of that infant.

            In another one of the cases of severe encephalopathy in the control group the infant also had an operative vaginal delivery.  I believe it was a vacuum delivery, and there was evidence of trauma, possible trauma, from that delivery.

            The third case is one I believe where the cord blood data was not available to say whether or not metabolic acidosis was present.

            With regard to intrapartum and perinatal deaths, there was one perinatal death in the control arm of the study.  Unfortunately, that was a subject for whom about only 20 minutes of the tracing is available.  The STAN monitor was disconnected from the fetus around two-and-a-half hours prior to delivery.  So we don't have data to know how that labor really ‑‑ what took place during that labor and what the tracing looked like.  Therefore, we can't make any conclusions as to the cause or the circumstances surrounding that perinatal death.

            We have more information on the two deaths in the STAN arm.  One occurred before retraining and one occurred after retraining.  One is described in the PMA as an intrapartum death, one as a perinatal death.  I've included them both under intrapartum.  The reason being is that both of them had Apgars of 0‑0‑0 at 1, 5, and 10 minutes.  One of these infants was resuscitated approximately 12 minutes after birth, placed on life support, and then expired or had life support removed at approximately 36 hours of age.  Because of the Apgars at 1, 5, and 10 minutes, I included those under intrapartum death, although that might be controversial.

            Some of the other issues that arose out of our review, you will find all of you have in your package today a list of questions that the FDA staff put together for the panel.  One of these is the lack of automatic ST event signals.  The second is deviations during the Swedish RCT from the patient management protocol.  Another issue has to do with retraining of clinicians during the trial, and the last is intercountry differences in clinical practice between Sweden and the U.S., and whether or not that has any implications in our review of the PMA.

            These are cases where there was not an automatic ST event signal.  Now it's important to note here that, in all fairness, we only looked at the 46 cases of metabolic acidosis as well as a few other cases of bad outcomes.  We definitely did not look at every tracing from this study of 4,966.  It wasn't practicable.

            So, again, these cases of lack of ST event signal really only apply to the 46 total cases of metabolic acidosis in the study.  So there's no way we can draw any conclusion as to whether or not this occurred in the other 4920-or-so subjects.

            There were seven cases in the STAN arm, 7 out of 15, and 8 out of 31 in the control arm.  What you see on the lefthand side of the slide is just apparent or confirmed reasons why there was no automatic event signal.  In several cases the STAN monitor hadn't been on for the required 20 minutes.  In other cases poor signal quality probably contributed to a failure to register an automatic event.

            The last two categories are more difficult to confirm.  Nevertheless, it appears possible that some sort of acute hypoxic event took place sometime after the scalp electrode was removed and actual delivery.  We believe that there is one case where pre-existing severe hypoxia was responsible for failure to record an automatic ST event because a myocardium that has been exposed to severe chronic hypoxia may not respond similarly to hypoxia during labor, as an infant, a fetus, who is not.

            What else did I want to say about this slide?  I think we can go on to the next slide.  Oh, one last thing.  I'm sorry.

            There's one other point that I wanted to make.  That is that the fact that there was no automatic ST event signal doesn't mean that you can't figure out whether there was an episodic or a baseline or a biphasic change.  That is a skill that is intended as part of the training for use of this device, but it's important to understand that there is a difference between registering an automatic signal that's very obvious to the clinician and the need for the clinician to look closely and try to look for ST events that might not be registering automatically.

            Now I would like to just talk about deviations from the patient management protocol during the Swedish randomized trial.  Primarily these deviations fall under a category of lack of intervention when it is indicated by the patient guidelines, and there are different combinations of situations of intermediate, abnormal fetal heart rate combined with different ST T/QRS changes or ST segment changes that, taken together, lead the clinician to intervene.  So, without going into specific cases of different combinations, nevertheless, I can again summarize and say that these deviations from the clinical management protocol primarily involve failure to intervene.

            I included a category for the STAN subjects in one case of failure to replace the scalp electrode.  Although this was not explicitly stated in the simplified clinical guidelines, it was in the protocol for the study that in the second stage that, if the signal quality was poor, the clinician was supposed to replace it, and in at least one ‑‑ well, in a number of cases this didn't happen.  But that is a protocol deviation that doesn't satisfy the strict definition that you see on the slide, which is the Swedish RCT clinical patient management guidelines.

            For the control arm, similarly, lack of intervention constituted the general category of protocol deviations.  We're only talking again about the patients who fell into that metabolic acidosis group.  That was only 46 out of 4,966.  So we can't make any generalizations regarding frequency or rates of protocol deviations across the entire study.

            One observation I would just like to make at this time is that, with respect to deviations from clinical management guidelines, we looked at the numbers of deviations in the two arms before and after retraining.  It was interesting to note that, of the protocol deviations that we have come to agreement with with the sponsor, that in the STAN arm four occurred before retraining; only two occurred after retraining.  In the control arm all of the protocol deviations occurred after retraining.

            The purpose of this slide is just to summarize or highlight some differences in clinical practice between Sweden and the U.S., and our guests here have done a very thorough job of this already.  So I'm not going to belabor it.

            There are some subtle differences between FIGO and ACOG terminology and definitions.  Certainly there is a difference in rates of Caesarian delivery between Sweden and the U.S.  The average rates at the three centers in the Swedish RCT were, I believe, around 11 and 12 percent, whereas our national rate in the U.S., I believe, is somewhere between 20 and 22 percent.

            The management of normal labor by midwives is one characteristic of OB management in Sweden that's different from the U.S.  As it has been pointed out, midwives manage normal uncomplicated deliveries in Sweden.

            Mid-cavity operative vaginal deliveries were interesting in this study in that in the STAN arm about 50 percent of the operative vaginal deliveries were mid-cavity deliveries.  That's probably a higher percentage than we would see in the U.S., but, in all honesty, I don't have statistics on that.  In the control arm it was slightly lower; around 40 percent of those operative vaginal deliveries were mid-cavity deliveries.

            In looking through the cases, the tracings for the cases of metabolic acidosis and other adverse outcomes, we noticed a few examples of what might be differences in labor management, including management of chorioamnionitis, duration of the second-stage, and management of uterine hyperstimulation, the presence of a non-reassuring fetal heart rate, but I can't make generalizations here.  I could point to specific cases where I believe that patients might have managed differently in the U.S.

            With respect to biostatistical issues, there are no FDA biostatistical issues regarding the presentation of the efficacy data for the primary and secondary endpoints.  The numbers and the percentages and the statistical significance are not being contested.

            As I said earlier, I just wanted to summarize some outcomes of the other clinical studies using the STAN monitor, in part to give credit for what is a vast body of clinical experience with this device in Europe.  The Plymouth study enrolled 2,400 subjects.  These were high-risk labors.  It was not designed or powered to evaluate metabolic acidosis.  There was a significant reduction in operative delivery for fetal distress and Caesarian section for fetal distress in the arm of the study that used the STAN device.

            At that time ‑‑ that was in the early 1990s ‑‑ there were some differences between the device used then and the one used in the Swedish RCT; for example, in automatically identifying biphasic events.  There was also in that study a trend toward reduction of metabolic acidosis in the STAN arm, but that was not statistically-significant.  As you see, the cases of asphyxia were approximately the same in each arm.

            The European multi-center trial was a relatively small trial which, as I understand, prospectively recruited subjects.  However, the ST data was blinded, and the subjects, the tracings were evaluated retrospectively.  Eleven out of 12 of those cases with evidence of hypoxia or asphyxia did have ST changes when the tracing was unblinded for that data.

            The Nordic observational study was larger than the EC multi-center study.  Management was based only on the fetal heart rate data, although ST data was available, and the results showed that following a retrospective evaluation, the tracings blinded to clinical outcome of 100 percent sensitivity for the STAN guidelines to recommend intervention for cases with neurological symptoms and/or metabolic acidosis.

            The City of Gothenburg experience was what we would term an observational study.  We have looked at some of those data.  Primarily, we have looked at the improvement that occurred from the first four months of the study to the last four-month interval of the study and the labors monitored with the STAN device.  We would note that this, as an observational study, one of the sites I believe was a participant in the Swedish RCT.  So that site probably had greater experience and expertise with the device than the other two.  If we were designing an observational study, we might have tried to include sites with similar levels of experience.

            Currently, the European Union is sponsoring a large study of 10 different sites, evaluating the dissemination of knowledge and training in the use of the STAN monitor, and this is significant because there is clinical outcome data for part of the study, and the sponsor would plan to establish a similar Centers of Excellence program in the United States for this device.

            At this time, if you'll just bear with me briefly, I'm going to try to show you some examples of tracings of some cases using the STAN monitor.  The first three cases I would like to show, for those of you who might not have gotten your disks working, actually, they're from a training CD, a collection of cases that are used to train clinicians using the STAN monitor.  The purpose of showing these case studies is just to illustrate different aspects of the STAN monitor, but it is not to comment on management per se.

            The first case that I'm going to show you isa VBEC patient at 41 weeks ‑‑ I'm sorry, 42 weeks, one day, who was dilated 5 centimeters at the time that the monitor was applied.

            These tracings go relatively quickly.  So I'm just going to back up very briefly.

            Again, what you see on the bottom of the screen is the T/QRS data here.  So this is the line that we would be looking at during cases recorded with the STAN monitor.  For those of you in the audience who might not be obstetricians, this line up here is the fetal heart rate, and this line down here reflects the uterine contraction pattern.

            So what I'm going to ask you to do is mentally be evaluating how you think the fetal heart rate tracing is going, and bear in mind what's going on in the bottom of the screen, where we are getting our T/QRS data.

            I'll try to speed this up to what I think is the punch line.

            There's reasonable variability here with some reactivity, and then we're going to evolve into a slightly different pattern.

            So what we had here ‑‑ I'm just going to back that up for a second ‑‑ drop in baseline, and again some drops in baseline that look to be ‑‑ it's difficult to evaluate the fetal contraction pattern.

            We're losing some variability.  We're losing reactivity.

            And what I am looking for here is an ST signal event marker.

            (Member of the audience begins to speak.)

            CHAIRMAN BLANCO:  I'm sorry, please, no comments from the audience.  If you need to make a comment, you need to identify yourself, okay?

            DR. CORRADO:  What might be happening here is that the point that I was trying to illustrate from this case is that there was an ST event, an automatic ST event, that occurred approximately 40 minutes prior to delivery.  It would have signaled an intervention.  The infant was born with Apgars of 1-5-5 and a cord arterial pH of 6.82 and a base deficit of 17.

            The reason I selected this case was to illustrate that if the ST event data had been available and the case intervened according to the STAN monitor, we very likely would have gotten a better outcome.

            Can I just interrupt this for a second, please, and ask if maybe the sponsor can tell me, are we not going to see the actual ST event automatic signals here?  Because that was what I was trying to illustrate.

            (Sponsor representative speaks privately to Dr. Corrado.)

            Well, given that I need a slightly different directory, what I will do here is, regarding these training cases, what I would like to do is just tell the panel that that is a case where there actually was an automatic ST signal.  There were a couple of other cases where there were no automatic ST events, and I think that, because there are no automatic ST events, I'm going to go ahead and look at those.

            So, again, we're getting into an area here where there's relatively little variability, extremely low variability, really a poor tracing.  It's hard to tell whether these decels, the pattern of the decels here.

            Okay, the lesson in this case is as follows:  We saw there was a non-reassuring fetal heart rate during this tracing.  There were no automatic ST events recorded, however.  This baby was born with Apgars of 9-10-10, and the point here is that this was a case where we could have relied on the STAN monitor and not intervened and essentially tolerated the non-reassuring fetal heart rate.  Because this baby had a normal vaginal delivery and was in very good shape.

            Okay, third case, we want to talk about, again, this is a case where there was no ST event, but I would like you all to think about, looking at this case, whether or not you would have intervened.  I'll try to slow down just a little bit.

            So, again, the exercise is to think what might I have done here.

            Okay, no automatic ST events.  There was a normal vaginal delivery.  In this case the baby was born with Apgars of 7-10-10.  I think that this is illustrative of the importance of the training because I think that that time of tracing, given the clinical guidelines, will require a lot of training and experience for us to manage a case like that according to the STAN guidelines.

            The last case that I would like to show is a case that occurred in the RCT, the Swedish RCT.  This was a case of an infant whose mother had a fever of 39.3 degrees C., was on antibiotics.  This is an infant who died, who was born with Apgars of 0-0-0.  I believe this is the case Dr. Rosen referred to as the cervix having been colonized with Group B strep.  Let's just take a look at this case.

            We have a relatively short record on the STAN monitor from this case.  I'm just going to try to slow this down.  Extremely almost absent variability.

            This recording takes place after a prolonged rupture of membranes.  The patient is on antibiotics.  So, again, this would have constituted a pre-terminal tracing, according to the STAN guidelines, and is a case that would have been intervened in earlier according to the STAN guidelines.

            The point here, though, is, again, this is a case of a stillbirth, and there was no automatic ST event signal.  So sometimes when there's no automatic ST event signal, everything turns out great, and sometimes it doesn't.

            I'm going to just conclude now with a wrapup of FDA's review to date that focuses on the panel questions.  Everybody has a copy of the questions that we posed to the panel.

            Question 1 has to do with the study design and the endpoints.  Question 1A is, was the study endpoint appropriate?  And Question 1B was, was the definition of metabolic acidosis as prospectively identified by the sponsor clinically meaningful?

            Question No. 2 has to do with the outcomes in the study.  I showed a couple of slides just presenting the outcomes.  We would ask the panel to discuss the clinical significance of these results, first for the intent-to-treat group ‑‑ that was Analysis I ‑‑ and then for the adequate recordings group.

            As I just summarized ‑‑ I apologize, this is very difficult to see on the screen.  It's virtually impossible to see.  So I just, again, point out for everybody that you've got these questions in the handout for the meeting.

            This is panel Question 3, and we would like the panel to discuss the implications of a number of issues in relation to the clinical significance of the results that were described under Question 2.

            First, deviations from the patient management protocol.  Second, no registration of an automatic ST event in certain cases.  Third, exclusions of subjects in the study based on inadequate recordings.  Fourth, intercountry population and management differences.  Then, lastly, whether or not there's any significance to the retraining that occurred in the middle of the study.

            We would like the panel to weigh the other body, the large body of clinical data outside of the Swedish RCT and determine to what extent do the results from those studies support the safety and the efficacy of the monitor.

            We haven't said much about labeling and training, but it's one of the most important things that we do here at FDA.  We would like the panel to comment on the appropriateness of the indication and whether or not the PMA data support this indication for use, and whether or not the professional labeling and training materials are sufficient to ensure appropriate use of the STAN monitor.

            If the panel votes to recommend approval of the monitor, we would like the panel to comment on whether or not there's a need for post-approval studies.

            With that, I'm going to retire here, unless there are any questions for me specifically regarding the FDA review, and let the panel begin its deliberation.

            CHAIRMAN BLANCO:  Thank you very much.  We'll go ahead and begin the deliberation portion of the meeting.

            We've already read the discussion questions, and, Panel Members, they are in your handout folder, so you can look at that.

            What I would like to begin with, and we started doing this, is we would like to try to address if the panel members have any questions of fact that they might like for the company, the sponsor, or FDA to address, if we could bring those up during the time that we have before lunch, so that would give them some time over lunch and prior to the last open public hearing portion to try to address those specific issues.  Is that kind of clear with everybody?

            All right, I'm going to ask Dr. Ramin, as the lead reviewer, to go ahead and start us off on the discussion.

            DR. RAMIN:  I'm Susan Ramin.  I'm the lead reviewer of this PMA.

            I have a question to begin with, and that is, looking at the Plymouth study, the intervention there was that it should be performed, to be done, if the T/QRS ratio rise exceeded a certain value or a certain level, whereas in the Swedish randomized clinical trial the intervention should be if there is a non-reassuring fetal heart rate tracing and if there is a rise in the T/QRS ratio that occurred.  I was wondering if we could have clarification regarding the use of a certain value for the rise in the T/QRS ratio versus the occurrence of a rise in the T/QRS ratio.

            CHAIRMAN BLANCO:  Go ahead.  Keep on with the questions.  What we would like to do, give them a chance to write them down, think about it, and then come back later on in the afternoon, so that they don't have to do immediately.  So let's just keep going if you have other questions for them.

            DR. RAMIN:  Certainly.  In the report published in the American Journal of OB/GYN in 2001 by Westgate, they looked at near-term fetal sheep, looking at ST waveforms that occurred during repeated umbilical cord occlusions, trying to simulate the dynamic changes that occurred during labor.

            What the authors concluded is that monitoring must include the T/QRS height as well as the waveform analysis.  I was wondering if you could clarify, does the STAN monitor evaluate the height of the T/QRS as well as the waveform analysis?

            There were also approximately 14 percent of the cases, roughly about 600 cases, where there was no cord gas data available.  I know that part of the trial requirement was to have both an artery and a venous sampling.  I was wondering if you have the numbers of how many cases had just an arterial sample and how many cases had a venous sample, and whether or not there were any cases of metabolic acidosis based on one artery or venous sample.

            I was wondering if the sponsors could provide their definition of mid-pelvic operative delivery, and whether or not ‑‑ what is utilized for that definition in Sweden and what was utilized for the trial.

            CHAIRMAN BLANCO:  Your implication there being a comparison to what may be the definition in the United States, correct?

            DR. RAMIN:  Correct.

            There were the 46 cases of metabolic acidosis out of the entire population.  In the randomized clinical trial, as was brought out by the FDA, the STAN monitor did not register an ST event in 7 of the 15 cases in the STAN arm and in 8 of the 31 cases from the control arm.  My question would be, what is the sensitivity and specificity of this device in detecting metabolic acidosis and hypoxia?

            Likewise, in the randomized controlled trial, there were inadequate tracings or inadequate recordings in almost 12 percent of the cases.  My question is, is the 11.6 percent inadequate recording rate satisfactory?

            In the United States there is a tendency toward less operative vaginal deliveries, marked increase in Caesarian delivery rates, and there's a limited use of fetal scalp blood pH.  Given the lower Caesarian delivery rate in Sweden, roughly 13 to 15 percent, compared to the United States, which is over 20 percent, is there any evidence that the use of the information from the ST event or the ST waveform analysis during the second stage of labor would result in obstetricians not to intervene when it's appropriate?

            CHAIRMAN BLANCO:  I'm sorry, could you clarify that a little bit more?

            DR. RAMIN:  Sure.

            CHAIRMAN BLANCO:  Because I'm not sure I understood that one.

            DR. RAMIN:  You gave the example of that recording where there was good variability, beat variability in the fetal heart rate tracings, but there was also an ST event, and a Caesarian delivery was performed.  The cord gas was normal, and thus intervention was done based on the ST event recording.  And the question is, how many times did that occur where an intervention was done based on ST events even though the fetal heart rate tracing was reassuring and the cord gas status was normal?

            Yes, you have a question?

            DR. ROSEN:  Could I ask for clarification?

            CHAIRMAN BLANCO:  I'm sorry, are you going to ask her to clarify the question, so you understand how to approach it?  Yes, go ahead.  Identify yourself first, please.

            DR. ROSEN:  Karl Rosen, a member of the sponsor team.

            So you are looking for cases where intervention was made on the basis of ST information only ‑‑

            DR. RAMIN:  Correct.

            DR. ROSEN:  -- with reassuring fetal heart rate ‑‑

            DR. RAMIN:  -- fetal heart rate tracing.

            CHAIRMAN BLANCO:  Any other questions?

            (No response.)

            All right, does anybody else on the panel have some questions?  Well, let's just start from this end.

            DR. RINGEL:  Richard Ringel, pediatric cardiology.

            I had questions or clarifications, or I would like a clarification on the Gothenburg reports.  I understand that it's observational, but I was, I guess, curious about whether these were commercially-acquired devices by the hospitals that then did the study or were they provided by the company?  What was the training like?  How much was this like a study as opposed to how much was this like clinical practice?  Because I think that has a lot of implications for how the device would be used once sold in the United States.

            And I don't think it was reviewed in his presentation by Dr. Rosen, but I was curious about the results in the Gothenburg trials.  It appeared that the decrease in metabolic acidosis was not as good in the STAN group as in the CTG group in the latter half of the trial, October to January, 2001 to 2002.  So it appeared as if the greater improvement was in the patients monitored by traditional technique rather than by the STAN, I think.

            Then my final question was regarding alarms.  Does the system come with any alarms?  Are there alarms that can be set, or is it all totally dependent on observation of the treating physician or nurse?

            CHAIRMAN BLANCO:  Thank you.

            MS. TOLEDANO:  Alicia Toledano from Brown.

            I had the same question about audible alarms, and I had a question about how much experience you have with breach presentation.

            My main question is, how much do you think you would have an effect on labor management if you just retrained people on the current ‑‑ you know what I'm trying to say.  If you retrained people on using the current fetal monitoring techniques, would you be able to effect the same change as retraining them including the STAN?

            I'm done.

            CHAIRMAN BLANCO:  Okay, thank you.

            DR. WOLFSON:  I'm Bob Wolfson.

            My question is actually a couple of questions.  One, why the 20-minute interval?  What is the statistical basis of that?  I presume it has to do with signal-to-noise ratio, but could you please explain why that was derived?

            Also, the issue of mid-forceps, my impression is that is a relatively rare event in the United States.  Therefore, one could at least crudely move the mid-forceps rate either by vacuum or by forceps formally that's reported in the Swedish study into the Caesarian section column for assessment in the U.S.

            I'm interested if you have any specific information on IUGR.  The low birth weight rate clearly in the United States is substantially higher.  Granted, a significant portion of that is due to pre-term births because our pre-term birth rate obviously is approximately twice that of Sweden, typically running around 10 percent.

            But, specifically, because IUGR represents a fetus at risk for metabolic acidosis, and you've demonstrated, at least my understanding of the physiology, you're reporting that this is specifically an area in which the biphasic T wave is important.  I would like to know if you can tell us a little bit more about the predictive value specifically in IUGR, since I think that's a major impact in the population differences.

            Oh, and could you please clarify for me, I still remain confused as to which is the primary drive in this study or in the way of using this device.  When you're monitoring an individual, are you using the electronic fetal monitoring information that would be, I'll say, conventional information as your primary tool, where the ST segment information now becomes secondary in making a decision?  Or are you really utilizing the ST segment information, and when that is of question, you then look to the electronic monitor?  So which is the primary tool here?

            CHAIRMAN BLANCO:  But what you're asking, by primary tool, you mean, what was supposed to be the trigger that caused, that would have caused an intervention in the patient?  Is that what you mean?

            DR. WOLFSON:  Well, Jorge, it's not just the intervention.  The thing that I find myself doing, and just in looking at the images that were put up on the screen, my mind always looks for more information.  I found myself rapidly watching the T wave information, and when I saw the segment rise or become biphasic, I merely looked to the EFM part of it, the electronic fetal monitoring part.

            So I'm wondering, in the actual clinical practice, where is the emphasis in making the decision to intervene?  Because my impression was that this was originally done as an electronic fetal monitoring tool.  That's what drives the decision.  Yes, that's what drives the decision, and the ST segment information is the secondary piece of information.  So that if it's reassuring, you're cool; if it's not, you move toward delivery.  And, therefore, in a non-reassuring electronic fetal monitoring tracing, the ST segment information is the tie-breaker as to whether you're going to intervene or whether you're not.  Is that the correct impression?

            CHAIRMAN BLANCO:  Anything else?

            DR. WOLFSON:  That's it.  Thank you.

            CHAIRMAN BLANCO:  We're going to go back to Dr. Neuman over here.

            DR. NEUMAN:  I have several questions related to the technology, and let me begin these questions with a clinical issue; namely, the machine is essentially, if we put it in very simple terms, a traffic signal with a green light and a red light that's made by a decision in the machine some way or another.  I'm not sure I understand what that decision is, and it worries me a bit.

            So I would like to ask the proposers and the FDA if they could comment a little more about that, particularly the issue that Dr. Wolfson brought up, namely, about the 20-minute period where we don't know what's going on.  Yet, in one of the presentations the mention was made that 30 cardiac cycles were averaged to do the analysis, and I get that to be about 15 seconds or so.  So I'm wondering, what else is going on?

            The second thing is the robustness of the analysis routine in terms of signal distortions, either arising from the electrode itself or from some of the technical issues in terms of signal matching, which I don't think we should get into here.  But at least we should hear a little more about how robust it is and what constitutes the gaps in the data that we see.  What does the monitor decide as non-analyzable data?

            I believe somewhere in the documentation that was a comment that this data, in fact, can still be visually analyzed.  I would like some more information on that.

            The final question I have is with regard to the lead configuration, and probably my colleague on my left could better ask the question than I can, but it is my memory from the distant past that the configuration of the T wave is very much dependent on the particular lead that you're using to monitor it.  I'm curious if anyone knows what the spiral electrode on the fetal scalp is in terms of cardiographic lead, and even more important, is there the possibility that ‑‑ and I have to use a technical term here ‑‑ the lead vector changes during labor as the position of the fetus changes, and that this could result in some ST changes?

            CHAIRMAN BLANCO:  Thank you.

            Dr. Sharts-Hopko?

            DR. SHARTS-HOPKO:  Okay, Nancy Sharts-Hopko.

            I was also concerned about lead placement and diagnostic tracings.  So thank you.

            I am also concerned about the comparability of how women are attended during labor in Sweden versus the United States.  We can typically two registered nurses to six laboring women, thanks to central monitoring.  So there has to be a person there watching for these ST events, so that intervention can happen within 20 minutes, and I'm pretty worried about that.

            I am also concerned about in Book 1, Section 2, page S-4, we have a list of complaints from people who purchased 96 units.  There were 57 complaints.  I am interested to know how those have been addressed, and I'm assuming that these complaints the data that we've been given.

            Thank you.

            CHAIRMAN BLANCO:  Thank you.

            Jay?

            DR. IAMS:  Jay Iams, Ohio State.

            I have a couple of study questions that can probably be best found by looking at Dr. Corrado's presentation on pages 8 and 9 regarding infants who were excluded in Analysis II.  Is there any reason to think that infants who were excluded due to inadequate cord blood might have a different incidence of metabolic acidosis than infants who remained in the study?  Were they so profoundly depressed and had such poor cardiac output that the cord samples preferentially perhaps would have revealed babies who were sicker?  And if so, or regardless, do you have any data on those babies regarding their Apgars scores?

            The next one was on the next page, Dr. Corrado's slide No. 17, on page 9.  The percent of removals, she listed them as removals according to what percentage of removals were 1.3 percent, 37 percent.  Those are percentages of the removals.  I was wondering more about in each group, the STAN versus the CTG alone, if those percentages within each group were approximately the same.

            Let's see, we also learned that the STAN is not completely sensitive in detecting pre-terminal or being coincident or in agreement with pre-terminal tracings.  What is the degree of concordance when there's a pre-terminal tracing?  Does the STAN virtually find that always?

            And then the last question is regarding ‑‑ it's somewhat theoretical.  When you're looking at myocardial hypoxia, myocardial acidosis, how reflective of that, in animal studies perhaps, how much does that reflect hypoxia and acidosis elsewhere?  Is it so sensitive to the heart that, in fact, it's not necessarily applicable to the systemic circulation?

            CHAIRMAN BLANCO:  Thank you.

            Gary?

            DR. EGLINTON:  I'm still thinking about mid-pelvic operative deliveries and wondering about the definitional difference.  It's probably in here and I just can't remember it.  I would like to know what the frequency of mid-pelvic operative delivery is in this sequence of studies, if it's relatively uniform in all these studies or especially in the Swedish RCT.

            Since I've been a chairman for three years and been looking at credentials packages, a third of my staff that has joined my staff in the last three years, almost none of them apply for privileges for mid-pelvic operative delivery.  They just skip that box on the credentials sheet.  I think there's probably a major cultural difference here between the European Union and the U.S.

            Thank you.

            DR. O'SULLIVAN:  Mary Jo O'Sullivan, Miami.

            I have a question regarding the training/retraining.  I would like to know (a) what were the differences between the initial training and, after the first interim analysis, the retraining, if there were any differences at all and how that training was done, other than I realize that individual hospital cases were used for the retraining purposes, and they were, therefore, retrained on their cases.  Was that the only difference that occurred in the whole retraining process?  If it was not, then what were the characteristics of the original versus the retraining?  I ask that because of the differences between the interim analysis and the subsequent analysis.

            DR. BROWN:  Carol Brown, New York.  I have a couple of questions.

            One was about the correlation between cardiac ischemia and specifically the effects on the central nervous system.  Is that a correlation, again, in an animal model, specifically to metabolic acidosis and encephalopathy, any type of fetal model?

            Another question about the training/retraining:  Was there any assessment of the baseline experience and familiarity with using CTG at your centers?  In Sweden I obviously you have, I would assume, a much higher proportion of nurse midwives for managing most labors.  Was there an assessment of what their baseline experience and interpretation of CTG was before retraining, before training and retraining?  And specific numbers in terms of the percentages of these traces that were managed by nurse midwives versus OB/GYNs, proportionately, because I would suspect there would be a big difference in the United States in terms of the number of nurse midwives proportionately who would be managing these cases.

            One other question was, it was mentioned in some of the background information about ST segment depression specifically.  I just wanted to clarify how ST segment depression, is that incorporated in the biphasic assessment or is it possible that that could be a separate thing that would or would not be picked up by this system?

            CHAIRMAN BLANCO:  Thank you.

            DR. SEIFER:  David Seifer, New Jersey.

            Two questions:  One is the inadequate recordings of 12 percent, is a number that one would expect with a new technology, looking at these endpoints?  Also, so would one in actual practice, as opposed to a study, expect to see 12 percent inadequate recordings?

            I would imagine that part of the answer to that question might include what clinical information investigators have to understand if this greater than 50 percent of greater than 20 minutes between the device and delivery, if we have an explanation of why that might have occurred and whether or not that would be preventable in a clinical setting?  Also, with regard to the greater than 35 percent with the less than 20 minutes of recording, again, is that something that is preventable?

            My second question has to do with just basic statistics with regard to positive and negative predictive value of the control group versus the STAN group with respect to metabolic acidosis and Caesarian section rates.

            MS. MOONEY:  I have no questions.

            CHAIRMAN BLANCO:  Thank you.

            MS. LUCKNER:  I've just got one.  Kleia Luckner, Toledo, Ohio.

            We've had some discussion here about the comparability of the labor management between Sweden and the United States.  My question is, specifically, in Sweden is the nurse midwife physically present in the labor suite, for the Swedish colleagues?  And then ACOG or FDA have any data on the United States?  Where labor is managed by obstetricians, does that mean physical presence at the labor suite, since most of us around the table know that that's not what happens?

            CHAIRMAN BLANCO:  Anything else?

            MS. LUCKNER:  Thank you.  That's it.

            CHAIRMAN BLANCO:  Thank you.

            I have one question I would like them to address and clarify, the utilization of 7.05 as the definition of metabolic acidosis, how was that arrived at, and why the decision was made to use that value?  It's not, I believe, the common value used in the United States.  That was the one question that I had.

            What we're going to do now ‑‑ there are a lot of questions that folks have.  I would like to encourage the sponsor and FDA to be very brief.  Most of these are actually addressed to the sponsor.  Be concise and to the point because we need to not be spending all the time listening to the answers.  We need to discuss what we think.  So we would like clarifications, and many of the questions are similar, so you might want to join them together on there.

            DR. WOLFSON:  Jorge?

            CHAIRMAN BLANCO:  Yes, go ahead.

            DR. WOLFSON:  Can I add an additional question?

            CHAIRMAN BLANCO:  Sure.

            DR. WOLFSON:  Could you clarify the basis by which the actual management guidelines were derived?  There were specific differences in the rise of the ST segment based on the CTG information.  It wasn't clear to me, though I think you had one slide that suggested, but I didn't understand the differences as to why in one case it's greater than .1 where in the other case it's greater than .15.  Is there actually specific statistical information that demonstrates the value of those specific thresholds or was this based on clinical grounds and then simply prospectively moved forward?

            The other question that occurred to me in the midst of the discussion was, could you clarify again the use of fetal monitoring in Sweden compared to the U.S.?  My impression certainly in my own venue is that electronic fetal monitoring will be used somewhere in 80 to 90 percent of labors.  Granted, typically, this is external monitoring through the bulk of the labor and not an internal monitoring?

            My impression was that your data suggests that the internal monitoring takes place in about 40 percent of your labor.  So you have different criterias.  I would just appreciate a clarification.  Thank you.

            CHAIRMAN BLANCO:  All right, so let's be concise, and let's try to put some of the questions that are fairly similar together, so that we can get some discussion of the questions that FDA would like for us to address.

            The other thing that we're doing, I think we only have 30 minutes under the current agenda.  I'm going to take the Chairman's prerogative and go ahead and close the meeting now for lunch and reconvene a bit early, so that we don't have to break the discussion up into a small amount of time, if that's all right with the panel.  Okay?

            So it is right now, by the official clock, 11:30.  We would reconvene at 12:30 and begin the discussions at that point.  Thank you very much.

            (Whereupon, the foregoing matter went off the record at 11:31 a.m. for lunch and went back on the record at 10:54 a.m.)


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

                                        12:34 p.m.

            CHAIRMAN BLANCO:  All right, let's go ahead and see if we can get started, please.  All right, let's go ahead and begin the afternoon session.

            The way that we're going to try to do this is we would like to, first of all, address the discussion questions that the FDA has posed to the panel.  Panel Members, you should have a copy of these in your panel booklet.

            If some of the questions that were asked by the panelists of the sponsor are included in this particular questions for the FDA, then we'll ask them to come up and answer it at that time, but if not, then we will wait until the time that we have allotted for the open public hearing later this afternoon again, and at that time we'll bring up the company, the sponsor, to answer some of the questions that the panelists posed.

            All right, well, let's go ahead and begin then.  Let me just read the first question, and it's also up on the screen for everyone.

            Discussion question, safety and effectiveness.  No. 1:  "The pivotal clinical study supporting this PMA is a large multi-center randomized controlled trial conducted in Sweden.  The Swedish RCT was designed to compare several fetal and maternal outcome measures between women managed with STAN monitor technology and women managed by conventional monitoring technology."

            A:  "The primary endpoint for the study is metabolic acidosis.  Is this endpoint appropriate?"

            And I guess I'll go ahead and turn to Dr. Ramin to start off the discussion on 1A.

            DR. RAMIN:  I think the primary endpoint using metabolic acidosis is appropriate and reasonable to utilize, given the pathophysiology behind this device.

            CHAIRMAN BLANCO:  Okay.  Any other comments?  Anybody else want to address the issue of the metabolic acidosis?

            DR. IAMS:  I'll just make one comment.  It has to be recognized; it's the best surrogate endpoint you can use, but metabolic acidosis is not, in and of itself, an endpoint.  It's a marker for the ultimate endpoint, which is neurologic injury, presumably.  We all know the difficulties inherent in using that as an endpoint.  So it's an appropriate choice for a surrogate endpoint.

            CHAIRMAN BLANCO:  Well, let me throw out some things on this.  I mean, they may be an appropriate endpoint as surrogate, but in looking at how it was defined in this particular study, there seems to be a significant number of neonates that, although had the definition, really didn't have very much of a significant, were not affected in any way that was reported.  So I don't know if part of that might have been what FDA had in mind, but to me I had some concern about whether the definition of that really reflects, and is a good surrogate for, something bad happening in the fetus and neonate.

            DR. RINGEL:  Our markers of a bad outcome, if you just look at the neonate, are perhaps weak.  So that stress that occurred during delivery might not show up for a year, two years, five years.  So that you have to use some marker; otherwise, the studies would never get completed.  Then you won't even be able to link it to later events, learning disabilities, and so forth.  So I think that metabolic acidosis is a reasonable marker.

            DR. IAMS:  I would just add I think the best marker, and I certainly defer to the pediatric colleagues around the table, but the best clinical actual marker of long-term performance that's measurable in the neonatal period is seizures.  If you have seizures in the newborn period, I think that's still the best clinical marker, not laboratory marker.

            CHAIRMAN BLANCO:  I'm sorry, was there a comment over here?

            DR. O'SULLIVAN:  Yes, I was just going to say, a lot of kids with seizures, Jay, don't have any problems at all.

            DR. IAMS:  Well, true, they don't, but if you're looking for a marker that's not a laboratory event, subject to the vagaries of where you set the threshold, et cetera, seizures I think are the one that's been used in most trials trying to predict who's going to have neurologic injury.  That's not perfect by a long shot, I agree, but I think that's the one that's in most common use.

            DR. RAMIN:  I have a comment just about your statement about metabolic acidosis and a fair number of these babies didn't have any adverse outcome.  We have seen that in a lot of studies where two-thirds of babies with metabolic acidosis go to the newborn nursery and do perfectly fine.  But I think we have to have some endpoint in this study, and it would be an immediate marker.  I mean, ultimately, what we need is long-term neurological outcome, but that's not the purpose of this study.

            CHAIRMAN BLANCO:  Gary, you had a comment?

            DR. EGLINTON:  That's okay.

            DR. O'SULLIVAN:  Maybe the difference should be in terms of what is metabolic acidosis.

            CHAIRMAN BLANCO:  Well, that's part of the next question, so that's why I was leaving for that.  I just wanted to make sure.  I think it's important to point it out, as Jay did, that it is a surrogate endpoint for what you eventually want to do.

            DR. WOLFSON:  My only comment was that I think that we all would agree that metabolic acidosis is not a natural state and, therefore, it's a state of physiology that we want to avoid.  While we may not have all the information on its actual impact, both short-term and long-term, on an individual, I think it's still a worthwhile endpoint to choose in a study of this nature because it's a space we don't want to be in for a fetus or, for that matter, the mother.

            CHAIRMAN BLANCO:  All right.  Any other comments on 1A?

            (No response.)

            All right, well, let's move on to 1B, brought up as well by Dr. O'Sullivan.  "The definition of metabolic acidosis in this study was an umbilical cord arterial pH less than 7.05 and a base deficit greater than 12 millimeters per liter.  Is this definition of metabolic acidosis clinically meaningful?"

            Would you like to go ahead and start, Dr. O'Sullivan?

            DR. O'SULLIVAN:  I think that the data that was pointed out by the FDA showing the differences between the 7.05 versus the 7.00, and the equivalent base deficits, is a little bit more convincing in some ways; plus, it's also what we use in this country, which I think is much more what we're used to.  That would be my only point.

            CHAIRMAN BLANCO:  Any other comments?

            DR. NEUMAN:  I'd like to just make a comment that it seems to me we're looking at a continuous variable and trying to make it a categoric variable.  Perhaps there should be some way of analyzing this data where severity enters in rather than just yes or no.

            CHAIRMAN BLANCO:  So what you're saying is some sort of analysis where you're looking not just at a cutoff 7.05, but you're looking at, you know, how well did it perform as the pH was lower?  Is that what you're saying?

            DR. NEUMAN:  Yes, and, of course, base deficit as well, but I think the data from the FDA demonstrated that, depending on where you take your cut point, you can get different results.

            CHAIRMAN BLANCO:  All right.  Well, I think it may be appropriate at this point maybe to have Dr. Rosen, whoever is going to answer the question about, why was the definition of 7.05 utilized, if somebody could come up and address that?  Maybe to the podium might be better.

            DR. ROSS:  I'm Michael Ross, Professor of OB/GYN at Harbor UCLA, and I'm a consultant to the sponsor.  My travel has been reimbursed for today.

            The issue of the 7.05 and the base excess is a great question.  Firstly, although, Dr. Neuman, as you suggested, pH and base excess are continuous variables, it appears from really the best literature, and that's largely derived from the extensive work of Jim Low in Canada, indicates that really there are threshold levels, and that base excess is probably the best single individual level to assess the risk of newborn neurologic disease.

            So I would argue that, rather than looking at either pH or base excess as a continuum, it appears that babies have a minimal to negligible risk of neurologic disease from an acute event unless their base deficit is 12 or greater.  And at that point still, Dr. Ramin, as you say, two-thirds of those babies will do well, but up to a third will have a significant neurologic deficit, again depending upon the degree.

            CHAIRMAN BLANCO:  So what you're saying is that the base excess of 12 was utilized on the basis of the study that you quoted with having correlation with neurological sequelae?

            DR. ROSS:  That study as well as really the extensive human literature largely produced by Dr. Low.  So the key issue in this, rather than being a pH, which is both a log function and has its problems, and a mixture of respiratory and metabolic issues, is the base deficit.  That is the key number.

            CHAIRMAN BLANCO:  Yes, but, nevertheless, the study was designed and the PMA is being produced utilizing both of those numbers.

            DR. ROSS:  Right, and both of those are important, if I may, because a base deficit of 12 is associated with a pH of 7.05, if the PCO2 is less than 60.  If you get above a 60 value with the same pH, then your base deficit is going to drop.  We felt that invariably these babies have some mixture of metabolic and respiratory acidosis.

            One could move to a value of 7.1 and then you only require a PCO2 of 50.  Anything above that, your base deficit is less.  So this, in fact, is a really stringent requirement of both the pH and the base deficit ‑‑

            CHAIRMAN BLANCO:  Well, you haven't addressed, or at least I'm sorry if I missed it, but you haven't addressed how you came up with 7.05.

            DR. ROSS:  7.05 is based on the issue that if the base deficit, being the primary issue, being 12 or greater, that denotes a high-risk infant.  If the PCO2 is less than or equal to 60, then the pH will be less than or equal to 7.05.  So it's a mixture, plugging in the CO2 value.  If you raise the pH ‑‑ or lower the pH threshold to 7.0, you may turn out to have largely just a component of respiratory acidosis in that.

            DR. RINGEL:  Can I ask, it's almost a lay question?  Does everyone obtain the umbilical cord arterial sample the same way?  Are we talking about the same thing here?  I really don't know.  I mean the OBs in the room would be able to tell me.

            DR. O'SULLIVAN:  Well, they had a specified way of doing it in their studies.  They did.

            DR. RINGEL:  And is it done the same way here?

            DR. O'SULLIVAN:  No.

            DR. RINGEL:  Oh, okay.

            DR. WOLFSON:  It's not?

            DR. RINGEL:  How's it done here?

            DR. O'SULLIVAN:  Well, first of all, I mean I can talk about one place, and perhaps Susan will talk about another.  But the cord is double-clamped and cut for one thing.  The specimens are not sent right away because they're kept at room temperature and they're felt to be valid for 60 minutes.  Here they were much more rigid about that.  Then if we decide to send them off, we collect them and send them off.  We usually collect them with the heparinized needle, as they suggested.  We do not ice them at all.

            DR. RINGEL:  And they're taken, then, from the harvested umbilical cord?

            DR. O'SULLIVAN:  They're taken from the umbilical cord ‑‑

            DR. RINGEL:  But the CO2 is changing then?

            DR. O'SULLIVAN:  Hum?

            DR. RINGEL:  The CO2 is changing as ‑‑

            DR. O'SULLIVAN:  Not significantly.

            DR. RINGEL:  Not significantly?

            DR. O'SULLIVAN:  No.

            DR. RINGEL:  Because it's clamped on either end.

            DR. RAMIN:  There have been studies that have shown that it doesn't change dramatically.

            DR. RINGEL:  Thank you.

            DR. ROSS:  Right.  So I think that the minor differences in procedures or rapidity of doing samples are not going to change the value significantly.

            CHAIRMAN BLANCO:  All right, thank you.

            All right, any other comments on the definition, and we do we think the definition is clinically meaningful?

            (No response.)

            I guess there is no comment on that.

            DR. EGLINTON:  Jorge?  Jorge, let me be the naysayer then.  I mean, everybody is familiar with probably 25 publications by Dr. Low on base deficit.  That's fairly straightforward, fairly clean.  I don't think anybody, answering Michael's original question, I don't think anybody has performed a receiver operator characteristic curve analysis of pH, which is probably what we would like to see.  But I think that, based on clinical outcomes, a bunch of the things that have been published, and a lot of it by Larry Gilstrap, has focused on a pH under 7 as being more clinically important in terms of longer-term outcome.

            If we look at the slide that Dr. Corrado provided for us, slide 21, at a pH of less than 7, there was no difference between the STAN and control arms.  They're really identical proportions.

            So it seems subtle, 7.0 versus 7.05, but it may actually be clinically important.  The 7.05 may not front-load the population with enough risk to be clinically useful.  It may require a pH lower than that in order to have enough risk in the resultant population, enough risk of long-term injury.

            CHAIRMAN BLANCO:  Well, I think it goes back to the original issue of, is the definition ‑‑ you know, we said metabolic acidosis as a term is a good surrogate endpoint for damage, but, yet, you know, that's part of the problem:  Is 7.05 enough of a definition that it a surrogate marker or should a lower number have been used to have been a better marker of actual damage to the neonate or the fetus or neonate?  I think that's the question that needs to be addressed.

            Dr. Wolfson?

            DR. WOLFSON:  Two things:  One, if we're looking back at other studies, we've got to be consistent in terms of what the outcome is.  What's the gold standard here that the pH is being compared to, and what level of damage are you requiring in those studies?  We also know that many of those older studies didn't take into account that a lot of the damage that was observed was not based on metabolic issues.  It was anatomic disorders in the individual.

            So I don't think that you can ‑‑ I'm not sure how to use that information.  It is all going to depend on what your endpoint is, what your gold standard is, as people say.

            Second of all, I think we have to be very careful.  This study was not designed ‑‑ we cannot retrospectively go back and expect a different analysis.  The power was not designed on that.  As the FDA pointed out to us, we have to be very careful in saying that, gee, it out to be a 7.0 based on some of the information that was presented.  If the study had been designed based on a 7.0, the study would probably have been larger and it would have been different, and maybe the same outcomes would occur, but we don't know that.

            So I think we have to accept what we've got as a good measure of metabolic acidosis that is reasonable in its calculation, as I understand the response, and its resource, and go with it.

            CHAIRMAN BLANCO:  Well, let me play devil's advocate a little bit and say, well, you know, I recognize what you're saying, that when you design a study, you use a certain number, and that's really what you're trying to achieve, and that you can't change the numbers after the fact to fit what you would like to see.

            But, at the same time, the issue, as we have already said, is that we're using metabolic acidosis as a surrogate for what you're really interested in, which is damage to the fetus and neonate.  Okay?

            DR. WOLFSON:  Right.

            CHAIRMAN BLANCO:  So the question that I was asking was not, let's change the definition and, oh, it doesn't work at 7.0, so that invalidates it.  It goes back to the issue of, with the numbers used, they show difference, but was the definition stringent enough to have that difference be clinically meaningful and significant for the application of other patients that this is going to be applied to, if it's approved?  That's the question.  It's not changing the statistics or changing the analysis.  It's, is that definition, was that definition appropriate to show what this surrogate endpoint needs to show?

            MS. TOLEDANO:  Let me focus for a second on damage to a neonate and what's clinically significant.  We have on Dr. Corrado's slides on page 12, her slides 23 and 24, and those show us the babies with moderate and severe encephalopathy.  There are seven of those babies.  What was their core blood pH?  Was it all less than 7 or were there some of them that fell between the 7.0 and the 7.05 that within the United States we would not have considered to have metabolic acidosis?

            So maybe if people are interested to find that out, how many of those babies fell below 7.0 versus within 7.0 and 7.05 ‑‑

            CHAIRMAN BLANCO:  Well, why don't we see if they're able to address that.  I think along with that, Dr. Ramin had asked a question about the 46 cases that were defined as having metabolic acidosis and what was the sensitivity and specificity of this device in identifying those particular cases.  Maybe we can have someone from the sponsor try to address those two questions at this point.

            Dr. Rosen, come to the podium, please.

            DR. ROSEN:  So the issue is very much, what babies are affected by the process of being born?  And we did an attempt in analyzing those cases by using the neonatal records, looking at what happened.  Did they have increased neuromuscular tone?  Did they have neonatal seizures?  What we found was that some of them had metabolic acidosis and others were exposed to the forces of labor applied by vacuum/forceps deliveries.

            I think the issue is very much on how we manage, learn to manage labor more appropriately.  As Dr. Corrado demonstrated, there were no cases in the STAN arm that suffered severe or moderate encephalopathy, whereas there were seven cases in the control arm that did this.  I believe that monitoring is very much a managerial issue and how you relate to the changes.

            To what extent there were metabolic acidosis of greater than 7, I have to look at the data.  You have all the data presented to you in your folder.  We have a detailed analysis of all these cases.

            CHAIRMAN BLANCO:  Well, that's why I would ask you if you could have someone maybe in the next 30 minutes or so ‑‑ I mean the seven cases with encephalopathy were either moderate or severe, and see what their pHs were.

            DR. ROSEN:  Well, I know that in some of these the pH was quite reasonable.

            CHAIRMAN BLANCO:  Well, let's see if we can figure that one out.

            DR. ROSEN:  Yes.

            CHAIRMAN BLANCO:  Okay?  And then how about the issue of the sensitivity and specificity in the 46 cases?

            DR. ROSEN:  Well, we haven't done that analysis simply because in those 46 cases we intervened, obviously.  We did sensitivity/specificity analysis on the Nordic observational data, where no intervention was done, according to the STAN protocol.

            So, clearly, whether one should go for automatic ST assessment or whether we should use what the educational program says, and the way the user is taught to handle the information by visual analysis as well ‑‑ and as I showed you, yes, there were cases where metabolic acidosis did emerge with no ST events, in one case with a pre-terminal heart rates where the guidelines said there was a need to do something, but there will be the occasional case where there may not be that dominant heart rate change and there will be no ST, or there may be a period of time, of 10 minutes, where we don't have data available to us.

            CHAIRMAN BLANCO:  All right, well, thank you for your answer.  I think we ought to add the question of, if you could see if you can have your data and have someone look to see what the pH of the seven cases with encephalopathy were.  We would appreciate it.

            All right, any other comments on 1B?  No other comments?  Dr. Wolfson?

            DR. WOLFSON:  Just one more question:  If part of the outcome issue is neonatal outcome with respect to neonatal encephalopathy, then can we not consider that the outcome that was created, since that's the long-term outcome, was one that was statistically-significant, and therefore, it correlates with the definition of acid-base?

            In other words, what I am really saying is that we just said that the metabolic status is a surrogate marker for neonatal outcome.  We have neonatal data across both the two groups and the total analysis that demonstrates the statistical level of .02.  That's on page 15 of the original book that we received.  So it seems to me that the endpoint we just said is statistically-significant in and of itself.

            MS. TOLEDANO:  I just wanted to say something about sensitivity and specificity calculations.  We need to make sure that when you're calculating your sensitivity, it's for the protocol guidelines.  Because, for instance, when you have a pre-terminal CTG trace, independent of whether or not there's an ST event, you would still catch it because that's your guidelines.  So I want to make sure that people realize that.

            CHAIRMAN BLANCO:  Any other comments on this issue?

            (No response.)

            Okay, well, let's move on to Question 2 then:  "List the outcomes from the Swedish randomized control trial.  Please discuss the clinical significance of these results, essentially looking at the primary endpoint as metabolic acidosis.  In the Analysis I, which is the intent-to-treat group, you have a significant difference from 0.69 percent to 1.5 percent with a P value of 0.02, and in the Analysis II, the adequate recording group, you have a 0.57 percent versus a 1.4 percent with a P value of 0.01."

            Any comments on those results and the clinical significance?

            I think it's kind of what we've been discussing, which is the definition, whether you accept the definition of metabolic acidosis, as was done in the study, and then show the difference.

            Anybody else want to say anything else?  Jay, do you want to add something?

            DR. IAMS:  I was thinking of something else when you were asking that question, but the bottom line comes down to the number of individuals with an adverse event.  Although the almost 5,000 patients is a lot, by the time you get down to the final endpoint, you're dealing with numbers that, for other tests, other populations, would perhaps be considered simply not enough.

            That's a concern I think that I have about the ‑‑ I'm not sure what the right word is, but the durability of these relationships is evident.  As Bob said, the P values are fine.  The design of the study was appropriate.  I don't disagree with the choice of endpoints.  I think the response that Dr. Ross gave was appropriate.  But when you get all finished with that, you end up with not very many patients in those cells, which is troubling.  If you add to that other issues that we'll talk about regarding trans-Atlantic transplantation, it just gets a little concerning.

            So it's hard to say I'm really concerned about this question or this question so much, but in the end the number of babies who had problems is a genuine concern.

            CHAIRMAN BLANCO:  I thank you.

            Gary?

            DR. EGLINTON:  I wanted to talk just a little bit more about people's concerns about sensitivity and specificity.  I think what we're talking about here with the RCT is really an interventional trial.  So we're not really interested in sensitivity and specificity.  We're interested in whether the intervention reduced the outcome of interest, and it did.  There was an outcome of interest prospectively assigned; it reduced the outcome of interest.

            In terms of sensitivity and specificity, we really shouldn't talk about that.  We should talk ‑‑ if you want to talk about that concept, what you really want to talk about is likelihood ratio.  This is a screening test, and in that regard, then, you would get that information from the multi-center trial and the Nordic Observational Trial, where it's not really an interventional trial.

            That will tell you, based on these 11 or 12 cases with evidence of hypoxia or asphyxia in the multi-center trial, and in looking at the Nordic Observational Trial, how good was the test, and the measure of that is the likelihood ratio, but it wouldn't come out of this trial, the RCT.

            CHAIRMAN BLANCO:  Any other comments from any panel member?

            DR. BROWN:  Yes, I just had kind of an informational question for the maternal/fetal medicine people around the table in terms of translating this data.  Someone made the comment earlier that the incidence of metabolic acidosis, I guess you all were saying, defined as less than 7.0, is much higher in the United States.

            Can you comment on that in terms of the numbers, if you were to sort of extrapolate this data, in terms of the numbers of patients you would expect to see if you were using this intervention in the U.S. population?  Or if you're going to use the cutoff of 7.05, which is what the indication is, do you have any sense or can you give us a sense of what comparable expectations of percentage of babies born, say, at Jackson Memorial who have less than 7.05 and base excess of greater than 12 are?

            DR. O'SULLIVAN:  Very low.  Very low.

            DR. BROWN:  So it's still very low?

            DR. O'SULLIVAN:  Uh-hum.

            CHAIRMAN BLANCO:  Any other comments?

            MS. TOLEDANO:  I've got my hand up, as usual.

            CHAIRMAN BLANCO:  Any other comments?

            MS. TOLEDANO:  I've got my hand up.

            I'm used to a screening setting in radiology usually where we look at very rare outcomes.  When I sit here, I see us talking about, are there too few of these babies?  Is this too rare of an outcome to really worry about?

            I think the decision has already been made by clinical practice where we are worried about that outcome because we go after operative deliveries and any indication of fetal distress to try to avoid the metabolic acidosis, to try to avoid the neurological damage.  So every single one of those 45 babies or 50 babies or 90 babies in the United States that could be saved by this, I think it's worth considering.  And, yes, it's a low number, but it's an important number.  That's just my personal opinion.

            CHAIRMAN BLANCO:  Dr. Ross, do you have the pH use?  Is that what you've got?  All right, come on up and let us know.

            DR. ROSS:  My appreciation to Dr. Corrado for helping me with these cases here.  Again, there were three cases in the moderate encephalopathy group, all of whom were from the control arm.  Their values were 6.78 with a base deficit of 12.8, 6.73 with a base deficit of 21.8, and 6.87 with a base deficit of 16.9.  Clearly, those three cases would meet anyone's definition.

            In the severe group there were two cases, again severe encephalopathy, again two cases, both from the control group, and those both have a little interesting twist.  One was born and a cord gas obtained only from one vessel, and it's not noted which vessel.  That baby was born after 8 to 10 minutes of a shoulder dystocia.  You're aware that that sort of marks your point in time and then you have 10 more minutes of now acidosis.  So that baby's single cord value was 7.17 with a base deficit of 3.7.  However, the immediate newborn blood, which was done at 13 minutes of age, essentially just following the shoulder dystocia, was 6.79 with a base deficit of 16.5, consistent with encephalopathy.

            Then the final case speaks to this issue of the importance of base deficit rather than pH, and that is the severe encephalopathy case, the second one from the control group, had a pH at birth of 7.1 but a base deficit of 13.7.  That likely was possibly due to some loss of carbon dioxide in the sample or some absence of respiratory acidosis.  But, again, it speaks to the base deficit being the key issue.

            So I don't think there's any real difference in this prediction whether you use 7.05 or 7.0.  Thank you.

            CHAIRMAN BLANCO:  Thank you.  All right, how about the secondary endpoints and other measures in the intent-to-treat group?  We haven't addressed the issue of operative intervention and C-section.  We've been mainly on the metabolic acidosis.  Anyone care to make some comments on the data for the secondary endpoints and other measures?

            (No response.)

            Not particularly overwhelmed by the panel's response here.

            (Laughter.)

            DR. SEIFER:  Only to reiterate Dr. Iams' concern about the small margin of difference here.  It's even slimmer, and it becomes more of an issue when we get to the next question.

            DR. O'SULLIVAN:  I think that the only thing I would add here is that, since mid-cavity delivery is whether they're forceps or vertex ‑‑ I'm sorry, forceps or ventouse ‑‑ are extremely rare in this country.  Now it's not that they don't happen, but it is extremely rare.

            CHAIRMAN BLANCO:  It's extremely rare, so you think it's not that applicable to the United States situation?  Is that what you're saying?

            DR. O'SULLIVAN:  Yes.

            MS. TOLEDANO:  But what would have happened to them?  Would they have ended up being operative abdominal?

            DR. O'SULLIVAN:  They would have fallen into the all operative interventions, which includes Caesarian section.

            MS. TOLEDANO:  Okay, so they still would have been C-section?

            DR. O'SULLIVAN:  Right.

            CHAIRMAN BLANCO:  So I guess the question is, you know, later on to think about in the vote, so I'll go ahead and bring it up, eventually, if the device is approved, the company will have indications for which the device has been proven to be clinically useful.  Does the fact that P value is somewhat small, and you have the issue of the difference in operative delivery rates in terms of vaginal versus Caesarian section in Europe versus in the United States, do you think that that makes any difference?  Would someone like to make a comment about the validity of operative intervention differences for this particular device?

            DR. RINGEL:  Just before we get to the issue of how they're delivered, I'm just thinking this through.  The STAN doesn't fix anything.  All it does is indicate that you've got to get the baby out sooner.  So in a certain group of patients that the STAN identifies as being at risk, you will increase the number of C-sections, or whatever, you know, operative interventions.

            So if we have accepted the metabolic acidosis as the endpoint, the small difference in operative interventions shouldn't bother us because, hopefully, the ones that are being intervened on are more appropriate candidates, so that we are swapping.  We are taking jeopardized or at-risk babies and doing operative intervention, and taking healthy babies out of that group.  So we wouldn't expect a huge difference in C-section rate, just in the metabolic acidosis rate.

            CHAIRMAN BLANCO:  Well, actually, I was going in a totally different area.  I was going exactly the opposite.  One of the things that's always of interest is being able to say that you've got a fetal heart rate monitoring tracing that may look non-reassuring, but you have a STAN tracing, if you will, that is reassuring, so you can keep going and maybe end up with a non-intervention vaginal delivery.

            If you look at the numbers down here in the first part, in the bottom part of the first page, the rate of operative delivery was actually lower in the STAN group.

            DR. RINGEL:  Right, I agree, yes.

            CHAIRMAN BLANCO:  What I was really addressing was just exactly the opposite of what you said, which is, you know, if the company is interested in saying that they can lower Caesarian section rates by identifying the non-reassuring fetal heart rate tracings, that the STAN monitor says it's okay to keep watching.  Is that data that we have before sufficient to be able to allow that statement to be made.  That's what I was trying to get at.  Okay?  Do you understand?

            DR. RINGEL:  Yes.

            CHAIRMAN BLANCO:  Okay.  Anybody want to address that?

            DR. BROWN:  Well, I just think it's problematic because, although you would have to then make the assumption that all the patients that had mid-forceps in Sweden would be patients who would have C‑sections here, is that really a valid assumption without actually knowing?  I mean, you're making a jump in the data to make that statement.  So I would personally have some concerns about making that statement, because you don't actually have the data in the labor management culture that you're saying the indication would apply for.

            DR. IAMS:  That's a reasonably comfortable jump, I think, for most of us doing obstetrics in the United States.  We don't do mid-forceps very often, as has been discussed.  So that's a concern ‑‑ I mean that is not as much a concern for me as the issue of we operate with nearly almost twice the rate of Caesarian sections in the first place, many of which are done for presumed fetal compromise when, in fact, no such thing exists.

            So in this marketplace that's a very appealing endpoint.  It wasn't the primary endpoint in this trial, but it is a very appealing endpoint because we know we do more Caesareans than we would like to do.  So that's a concern.  Our culture takes devices like this, and it's on the next page, I guess, as questions, and we do have a hard time using information like this to reduce the Caesarian section rate.

            I think culturally caregivers in this country are more likely to look at this as another way to find babies, especially given the primary endpoint of the study, babies who wouldn't be identified with traditional cardiotocography.  It might further raise our C‑section rate, maybe appropriately, as you said, but we already have a perhaps even bigger problem with injury to many mothers and complications with many mothers.  So we really are in a very different cultural environment to use this technology than what was done, I think, in Sweden.

            CHAIRMAN BLANCO:  Jay, I've got to kind of call you a little bit on that because you started out sort of saying you were comfortable with the fact that there's a difference in Caesarian section rates in the United States and Sweden, but you kind of ended with sort of implying, or at least to me making me think that you're really not comfortable with trading the data to necessarily have that indication.

            DR. IAMS:  Let me see if I can say it quickly and clearly again.  The simple issue of the data in this study using operative delivery as a descriptive term for operative vaginal and operative abdominal deliveries, that doesn't bother me.  That's nice.  So end of comment.

            I guess the second point was what I was trying to say, maybe not clearly, is that in our culture, marketed and promoted for its primary indication, this device has some significant potential to increase the Caesarian section rate, looking for babies who have got problems.  So we may decrease it; we may increase it, but we're more likely perhaps in this culture to increase it.

            CHAIRMAN BLANCO:  All right.  Any comments from anyone else?

            DR. O'SULLIVAN:  I think it's kind of interesting that we are talking about, on the one hand, decreasing the Caesarian section rate and, on the other hand, now people can come in and ask for a Caesarian section for any reason they want.

            CHAIRMAN BLANCO:  Yes?

            DR. WOLFSON:  I don't think we have the data, but I think that we have to be very careful in making the comparisons to look at, again, obstetrical management because, as I recall in some of the early work that was done on one of the NIH panels, one of the principal reasons or one of the areas of concern was that Caesarian section rates were increasing because of, for example, breach delivery.

            I don't know ‑‑ when we're looking here, we're looking specifically at delivery for fetal distress.  I don't have a way of comparing raw Caesarian section data or operative delivery data between the two.  My anticipation is that the overall incidence of Caesarian delivery for fetal distress is going to probably be uniform among most populations, given that the individuals come relatively healthy into the laboring process.

            We talk about the translation between countries.  That may be the key point, that in many of our academic centers, many of our urban facilities, that the individual coming in in terms of the level of pre-natal care, the knowledge of what's going on with the infant, or I should say the fetus, are not going to be comparable to the Swedish system, where they may have much more information and have their patients better defined than we do.

            So I think that for fetal distress I think we're okay.  I don't think that the differences in section rates is going to make a big difference.

            CHAIRMAN BLANCO:  Let me bring it back, because I really started this as an adjunct to, in other words, what the sponsor and company can claim that their product does.  Let me get not necessarily the definition, but if you're comfortable with the lumping together and if you're comfortable with the differences between Sweden and the United States in terms of operative deliveries, is this level of data, this level of significance sufficient to make the panel comfortable that they can make a claim that they've lowered the rate of operative interventions for fetal distress?

            Did I make that specific and clear?

            DR. RAMIN:  It's only 1.6 percent difference.

            Can you hear me?

            CHAIRMAN BLANCO:  No.

            DR. RAMIN:  When you look at it, it's only about 1.6 percent difference.  So it's questionable if it really ultimately decreases, at least clinically significant.

            DR. WOLFSON:  Okay, but you have to look at it from a proportional standpoint.  It's 1.6 percent out of 7.7.  So, proportionately speaking, it's a relatively large number if you then project it into a general population.  I know we're sort of maybe splitting hairs here, but even though it looks like a small number, the fact that it's statistically-significant says that you should be able to expand it to a larger portion.

            CHAIRMAN BLANCO:  All right.  If there are no other comments, I think we probably beat that one to death.  So let's move on.

            Let's go on to No. 3.  Hang on, we're going to No. 3.

            "Several issues identified in the FDA review may affect the results.  Please discuss the implication of each issue in relation to the clinically significance of the results presented in Question 2."

            You can see that there are five items, so let's go ahead and take it with the first side of 2A, deviations from the randomized clinical trial patient management protocol.  Anybody want to start the discussion?  Maybe Dr. Ramin?

            DR. RAMIN:  Right.

            CHAIRMAN BLANCO:  Do you think that the fact that there were deviations once patients were identified as having an event, but yet the management protocol was not followed, does that create a problem in your understanding or support of the data?  Or do you understand what I'm getting at?

            DR. RAMIN:  My opinion would be that, yes, there would be a significance clinically in the fact that there was deviation in the clinical management protocol, in defining an ST event but not intervening, especially in cases of non-reassuring fetal heart rate tracing, and you take the opposite approach as well.  So I think there are implications.

            CHAIRMAN BLANCO:  Okay.  Anyone else want to make a comment on that?  Agree or disagree?

            DR. EGLINTON:  Yes.

            CHAIRMAN BLANCO:  Gary?

            DR. EGLINTON:  I would like to congratulate the sponsors on in one case at least a 30-year career working on this and translating some basic science research.  This is translational research here now, bringing this into the delivery room.

            But I'm worried about a second translation.  These are services that are very tightly controlled, very high-level excellent care, people who are very bright, coached by zealots who are also very bright, and now we're talking about trying to translate this across an ocean.

            When we have some people who really have a lot of trouble understanding basic fetal heart rate patterns at this point, I don't see this as a question on piece of paper here, but it is a question ‑‑ maybe this is it; this is the best place for it, deviations.  I mean, they had deviations from their management protocol.  I'm just trying to imagine the deviations from this management protocol in my own labor and delivery unit, if I'm trying to introduce this technology.  I don't see that this is going to translate.

            We don't have any evidence that this management protocol is doable in the United States, that this can be implemented in the United States.

            CHAIRMAN BLANCO:  Yes, Dr. O'Sullivan?

            DR. O'SULLIVAN:  Well, I think there are two issues.  First, I think that the investigators pointed out, and they have to answer this question for me, what were the differences in education before and after?  I mean prior to the protocol violations, when they first got started, versus after the protocol violations.

            I would hope that eventually, with the right education, and starting out from ground zero, that we could teach our people in this country to perform adequately.  I think that it will be difficult, but I think it can be done.

            I think the big issue here is, what was the difference in the educational process as it was introduced versus after the first 1600 patients were done and the interim analysis were looked at that made the difference here?  How come there was such a difference?  Or how come it appears there as such a difference?

            MS. TOLEDANO:  I want to be devil's advocate for a moment and just to notice that when we evaluate safety and effectiveness as members of an FDA panel, it's for the device used in accordance with its indications.  While looking at the way the device would be used in our culture is certainly very important to all of us, that is not necessarily what we are supposed to be looking at when we are determining whether the device is safe and effective.

            DR. RINGEL:  Just to comment on the deviations question that you're asking us to comment on, I thought the deviations were equal in the control group and the STAN group, so I actually think it's a moot point.  People deviate from protocols, people deviate from accepted medical practice.  They were just showing that the deviations were equal in the two groups.  It may be a high number of deviations, but they were equal.  So I think it's a wash.

            CHAIRMAN BLANCO:  Yes, I think the question was really aimed, and it's because of what you said, in terms of how the FDA has to evaluate things.  The question was more, does the fact that there were so many deviations invalidate the results?  Your point is they were probably randomly in both groups, or maybe not randomly, but at least they were the same numbers.  So probably it didn't affect it one way or another.  Correct?

            DR. RINGEL:  That's the way I look at it.

            CHAIRMAN BLANCO:  Okay.  Any other comments?  Anyone else want to say something?

            DR. SHARTS-HOPKO:  Well, I have to recall that we've had experience in this panel with devices that were not used according to the written instructions and with adequate training.  The chicken came here home to roost.  So I think that is something that we have to be concerned about, and I'm concerned about that.

            CHAIRMAN BLANCO:  Okay.  Any other comments?

            (No response.)

            All right, what about B, no registration of an automatic ST event?  I think, Dr. Ramin, you brought that up.  What you meant by this ‑‑ let's try to get clarification ‑‑ was that there was obviously an ST event, but that the machine did not record it, or that there was some event happening with the fetus and it wasn't recognized as an ST event?

            DR. RAMIN:  Like, for instance, there was an example of an ominous or pre-terminal tracing where subsequently the infant died or had Apgars of 0-0-0.  So it was an intrapartum death, but there was not registration of an ST event.

            CHAIRMAN BLANCO:  So what you're saying, the machine may not have been accurate enough in the identification of metabolic acidosis or hypoxia in that fetus?

            DR. RAMIN:  Right.  I have some concerns about that.

            DR. IAMS:  That gets back to what Dr. Wolfson said a little bit ago, what is the primary piece of the information displayed that is going to be the action item?  Is it going to be the tracing, which we've all learned to distrust basically?  And the you go and look for the ST event marker and it's not there.  So where does that leave you?

            DR. RAMIN:  Which one do you rely on?

            DR. IAMS:  Yes, and that gets back to one of the questions we asked before the break, about the number of pre-terminal tracings that were also marked as being abnormal by the STAN.

            DR. RINGEL:  But, in fairness to the company, they very clearly, even in their brief summary, state that a pre-terminal ECG, you don't look at the ST analysis.

            DR. IAMS:  Right, that assumes that pr-terminals are all that easy to identify, which is one of the tracings that we were shown this morning reminds me of, sometimes they look pretty good.  It depends on where in the process you come into it, I guess.  If you come in at the end, it can get confused.  So we're just looking to see if there is a correlation there, and then what degree of correlation.

            CHAIRMAN BLANCO:  Well, yes, I think that's a very important point.  I think we ought to spend a little bit more time on it, because I think this is an issue with some other things that have come before this panel as well.

            You wonder, going back to A, if you'll allow me, how many of the deviations went, because it wasn't clear in terms of the indications, if you look at the little box thing where you act on it, I mean it wasn't totally clear to me, you know, exactly where you needed to go.

            I wonder if some of the deviations might have been the STAN monitor was identifying an abnormality or an event, but, oh, hey, the tracing looks great, so I'm just going to keep on trucking and keep laboring the baby, even though I should be delivering it, according to the protocol.

            So I think this is an important point in trying to give some guidance to the sponsor as to how they need to specifically word the actions.  I mean, what are we trying to do here?  Are we saying that this particular monitor is going to be useful when you have a non-reassuring fetal heart rate pattern, and then you use the fact that you have an ST event as your point at which you say, well, this baby's in trouble, so we've got to deliver this baby pretty much immediately, or are we going to say, or is the sponsor going to say, well, you know, I've got a reassuring fetal heart rate pattern, but we know that may not mean that much, so you've got an event, an ST event, being flashed by the STAN monitor; you still need to deliver that baby?

            Some comments on that?

            DR. IAMS:  Well, let me follow up on that comment I made about Dr. Wolfson's comment before and guess what the answer will be from the sponsors about the question of which one goes first.

            My sense is you will probably answer that by, with experience, you integrate the two of them together.  I don't know, maybe that's the answer.

            CHAIRMAN BLANCO:  Well, I don't know that that gives them a lot of guidance, Jay, but ‑‑

            DR. IAMS:  I'm not asking for guidance or trying to give them guidance.  I really would like to know the answer to that question before we deliberate a whole lot further, I guess.

            CHAIRMAN BLANCO:  All right, well, Dr. Rosen?  All right, Dr. Ross?

            DR. ROSS:  Michael Ross again.

            That's not the answer, Jay.  It is really a sequence of looking first at the fetal heart rate tracing and, secondarily, at the STAN.  As indicated on the little package insert, if the tracing is entirely reactive and reassuring, there is no need, nor request, to look at the STAN.  If the tracing is pre-terminal, then one has to act upon that.  In fact, because of the physiology of T/QRS in a pre-terminal baby, the T/QRS actually decreasing with ultimate hypotension just prior to cardiovascular collapse, you may not see it.  Again, of course, that's assuming that you start at the STAN relatively late in the process, which was the case in the examples shown.

            For all other heart rate tracings, those that are non-reassuring in our terminology or, according to the FIGO terminology, somewhat the different levels of concern, then the STAN becomes the secondary level.

            DR. WOLFSON:  Dr. Ross, then explain to me:  So if you have a reassuring electronic fetal monitoring tracing, and because the display shows you your ST segment, and you're seeing either ST event, because I'm still not sure which one to focus on more, the rising or the biphasic.  The biphasic certainly sounds more ominous.

            So if you see an ST event, you're going to tell me you're going to ignore the ST event because the electronic fetal monitoring tracing is reassuring ‑‑

            DR. ROSS:  That's correct, and, in fact, it may be an episodic rise.  I mean, let's assume that we have a completely reactive, reassuring tracing.  In looking at hundreds of these tracings, one may have an occasional deceleration with an episodic rise that does or doesn't meet criteria for what would be an abnormal tracing, as indicated on here.  Those would be ignored and viewed as a short-term event.

            So you're correct, look first at the tracing; ignore some short-term events in the ST, if they occur, which is infrequent at best.

            CHAIRMAN BLANCO:  Any other comments from the panel?  Go ahead, Gary.  Go ahead.

            DR. EGLINTON:  I wanted to ask about this card.  The system, I may not understand it.  Obviously, I don't understand it well enough, but it appears on the surface of it to be impossibly complex, because there's an algorithm on this side where there's a normal CTG, an intermediary which is other places labeled suspicious, an abnormal CTG, which other places is labeled pathological, and then pre-terminal.

            So there are six boxes in here that require a great deal of interpretation along the horizontal axis with intermediary and abnormal.  Once one pigeonholes the strip into one of these four lines, then one turns this over here and watches each individual EKG complex that shows up on the right side of the screen, or at least that's what I was trying to do, watching the strips, because you either have a biphasic ST for continuous for greater than five minutes or more than two episodes of coupled BP2 or BP3 in the case of an intermediary CTG, but it also could be an episodic T/QRS rise of greater than .15 or a baseline T/QRS rise greater than .1.

            Is the computer doing this in real-time continuously, and that's what puts the little ST event marker down on the bottom channel?

            DR. ROSS:  That is.  You do not look at the ST waveform.  That's in fact just displayed on the computer.  It's not displayed on the monitor paper.  So the computer will integrate that.

            I could see at first glance that this looks confusing, the grids on both sides, which speaks in part to the need for an education program, but it actually is not that dissimilar from what we use.  We could all agree on a normal definition of CTG, pretty much can agree on pre-terminal CTGs.  What this is attempting to do is define in fairly objective criteria what is the intermediary and the abnormal.  So it is really breaking our suspicious tracing into two categories, one being suspicious or intermediary and one being abnormal.  But the computer does the rest of that calculation.

            CHAIRMAN BLANCO:  I'm glad that Gary found it confusing because I'm always happy when I find something confusing that I'm not the only one.  I find it's very confusing, and I don't think it was clear at all, the episode or the sequence of events that should occur in terms of the assessment, what you've clarified for us at this point.

            I think you need to, I guess addressing the sponsor, I think you need to do a lot of work on clarifying it.  I guess if you forgive my being ‑‑ well, anyway, if you just forgive me ‑‑

            DR. EGLINTON:  Being from east Texas?  Is that what you're trying to say, Jorge, from east Texas?

            CHAIRMAN BLANCO:  Yes, from west Texas.

            (Laughter.)

            But you've got to Americanize it.  This isn't very Americanized, and it's not very much the way practice is done in the U.S.  This is going to be confusing.  This is confusing.

            Thank you, Dr. Ross.

            Any other comments?

            DR. WOLFSON:  Jorge?

            CHAIRMAN BLANCO:  Yes?

            DR. WOLFSON:  I just want to get one thing that still isn't clear to me, going back to this.  What does the machine actually show us?  In other words, just the basic question, when there is an ST event, does an alarm go off audible, visible?  Do we know so that it at least alerts the attendant to the patient to look at the tracing, if nothing else?

            DR. ROSEN:  Karl Rosen.

            The attempt that we are doing is to verify when the fetus is forced to respond to the stress of labor.  We're utilizing the ST and we're asking the computer to verify when there is a change from the situation that occurred prior to this event, and, in particular, with contractions, obviously, coupled contractions with illustration of where you then would see an episodic rise which is identified, and it has strict rules in the software to tell what type of change is required to achieve an episodic as well as a baseline change, as well as a set of biphasic STs.

            CHAIRMAN BLANCO:  I'm sorry, Dr. Rosen, I don't think his question was addressing that.  I think his question was addressing, you've got some bells and whistles when this bad thing happens ‑‑

            DR. ROSEN:  All right, yes.

            CHAIRMAN BLANCO:  -- bells and whistles to call somebody to look at this.  Is that right, Dr. Wolfson?

            DR. WOLFSON:  Yes, that was question one.  Just does it go off?

            DR. ROSEN:  It doesn't go off.  It's there.  It's written.  There is a flag raised on the screen, and you would have to go and verify that, sort of recognize that, that that is a possibility.  We have not included bells and whistles.

            We try to bring knowledge more, but, I mean, it might a cultural thing here that you like bells and whistles, in which case it will have it.

            DR. RINGEL:  We would like a sign that flashes, "C-section now!"

            (Laughter.)

            That would be good.  That would be good.

            CHAIRMAN BLANCO:  All right, thank you, Dr. Rosen.

            DR. WOLFSON:  I would just like to go back to the other questions.  Why 20 minutes?  What is the ‑‑ do you want to wait for that?

            CHAIRMAN BLANCO:  I'll tell you what, I think that's maybe not necessarily in here.  That's one of the questions that you asked that we said we were going to handle later on.

            DR. WOLFSON:  Okay.

            CHAIRMAN BLANCO:  So I don't see the exact applicability to the issue.  So let's make sure we finish with these before we go into the others.  We'll get to that one, I promise.

            DR. WOLFSON:  Well, I asked the question relative to how it made the decisions it did, but I will be glad to wait.

            CHAIRMAN BLANCO:  Okay, thank you.

            All right, anything else on B?

            (No response.)

            If not, let's move on to C.  Any concern about the exclusions based on inadequate recordings?  Remember that the way the protocol was set up, if you had an inadequate recording, you were supposed to reinsert so that you could get an adequate recording.  And there was a significant number, or maybe not significant, but there were large numbers of folks who had exclusions based on inadequate recordings.

            Yes, sir?

            DR. WOLFSON:  I was going to ask a question of my colleagues, since I haven't been in the delivering process for many years.  What actually is the incidence of failure of standard electronic fetal monitoring, of getting an inadequate signal?

            CHAIRMAN BLANCO:  A scalp electrode?

            DR. WOLFSON:  In other words, in a scalp electrode, what proportion of the time does it actually fail when you're just doing regular monitoring?

            CHAIRMAN BLANCO:  I think I'm hearing from my left very low.

            DR. WOLFSON:  Because in this group it would be about 5 percent, if you look at it from a purely technical perspective.  Only 5 percent of failures of where it would have been replaced.  I don't know what it is in the U.S. history.

            CHAIRMAN BLANCO:  Was it 5 percent?  I thought it was more.

            DR. WOLFSON:  It is.  The total number is about 11 percent, but when you actually look at the technical issues of poor signal where the system couldn't operate, that's only about half that.  So it's about 5 percent.  It's around 5 percent.

            DR. SEIFER:  On slide 17 she says less than 20 minutes on the STAN monitor.  Of that, less than 12 percent, 37 percent, and greater than 20 minutes removal of device and delivery at 56 percent.

            CHAIRMAN BLANCO:  Congratulations on getting the 20 minutes back into the discussion.

            (Laughter.)

            Dr. Wolfson?

            DR. WOLFSON:  That was not intended, Jorge.

            CHAIRMAN BLANCO:  That's okay.  We're going to talk about it now.  We're going to deal with it.  So be prepared.

            All right, so the issue is it is overall 12 percent, but then if you break it down, what percentage ‑‑

            DR. SEIFER:  Thirty-seven percent or ‑‑

            CHAIRMAN BLANCO:  Thirty-seven were due to less than 20.

            DR. SEIFER:  Yes.

            CHAIRMAN BLANCO:  And then 56 percent were greater than 20 minutes between removal of the device and delivery.  Maybe we do need to address the 20 minutes and see what that is.

            DR. SEIFER:  But if anyone has any greater clarification or definition of those cases in those two categories that aren't congenital malformations or other, what that's about, why that occurs, under what circumstances, because it accounts for more than 315 patients?

            CHAIRMAN BLANCO:  I think this is probably the appropriate time to talk about why the 20 minutes, what that has to do about, and then the overall 12 percent inadequate tracings, and how you think of that.

            Dr. Rosen, it looks like you're coming up.

            DR. ROSEN:  Karl Rosen.

            These 20 minutes, that is what the time required for the computer to do automatic analysis, and you need a baseline consisting of approximately 10 minutes to verify where you are to start with, and then you start to look for a change.  Obviously, you do a minute-by-minute basis, but after an additional 10 minutes, calculating the median value change.  Then you are significant and you can demonstrate automatically by mathematics.

            DR. WOLFSON:  So it needs about 1400 samples or 1200 samples?

            DR. ROSEN:  Well, in this process we require 10 T/QRS data for a median to be calculated.

            DR. WOLFSON:  Okay.

            DR. ROSEN:  So you could do that in a shorter space of time tentatively, but to be able to secure the accuracy of this data, according to the software mathematics that the machine is using, there is this 20-minute guidelines to secure.

            It is also a function of how the electrode stabilizes, because sometimes a few minutes are required for this calculator to just settle down and provide us with a clean signal.

            Now on the issue of exclusions ‑‑

            CHAIRMAN BLANCO:  Before you go onto that, because I'm still not clear, I mean do you have data that you looked at that said that 20 minutes was what you needed for the mathematical model?  I'm remembering Dr. Neuman's comment, and somewhere in the presentation, somewhere in the data it was that you needed 30 QRS complexes, and then you've got some information.

            DR. ROSEN:  But ‑‑

            CHAIRMAN BLANCO:  Wait a minute.  Wait a minute.  Let me finish.

            Then that 30 QRS complex in a fetus is a fraction, a fraction of a minute.  So could you elaborate a little bit more on the 20 minutes?

            DR. ROSEN:  I mean the 30 beats, now we've got to separate the beats from QRS data.

            CHAIRMAN BLANCO:  Okay.

            DR. ROSEN:  So 30 beats, high-quality beats, are required, and the system looks for that high quality.  So it will disregard those that are poor quality.  Then we say we need at least 50 percent of the time covered with acceptable half-beats to say this is a reasonable quality recording.

            Because the thing I don't want to see is erroneous data being entered into the T/QRS plot.  So our software is very much focused on avoiding that issue and using modern digital signal processing to sort of take noise out of this problem.

            I think from the traces you have seen you might have recognized the quality of the ECGs has actually been displayed, where you can identify a P wave, a QRS, and a T wave very accurately.

            Then we have T/QRS ratios that in a normal case ‑‑ and now I'm talking about 90 percent of all recordings, where we have a continuous; that means up to four QRS data plots per minute, and that is the norm.  We may address that now looking at the Gothenburg experience, where, according to the information I had here, it's only 1 percent now of the cases where we don't obtain an ECG.  It is obviously a reflection of the motivation of the staff to apply the scalp electrode correctly, but also the software engineering is continuing, and we learn and we improve the technology.  It is part of the process.

            CHAIRMAN BLANCO:  Dr. Wolfson, have you gotten your questions on the 20 minutes answered by that?

            DR. WOLFSON:  I think so, yes.

            CHAIRMAN BLANCO:  Now several people brought up the 20 minutes.  So does anybody else have any comments on the 20-minute window?

            DR. NEUMAN:  I still have a question regarding the fact that you used 30 good QRS complexes.  Is there a maximum number of bad QRS complexes in between?  If I can be a bit sarcastic, you wouldn't want to choose just 30 QRS complexes over the entire 20 minutes.

            DR. ROSEN:  Sure.  There are also are rules that in that case would turn the system on.  I mean we have set strict rules, and I don't have those in my pocket today to show you, but they are part of the software engineering, the quality control, the safety aspects.  We have spent 15 years of software engineering time to develop, having had access to the ECG.  As you can imagine, it is a bit of a challenge to record an ECG where you have a reference electrode placed on the maternal thigh.  This is required to obtain the vector to identify the T wave, which is different from the standard ECG monitoring procedure, where we don't need ST; we only rely on QRS.  It's very simple to measure that, but with ST it is becoming more of a challenge.

            So we need a unipolarity configuration.  We need the maternal thigh.  The mother is moving.  That creates a problem.  We've got to spend a lot of energy and time and skills in developing the software.  Today, with the latest software there is, it appears there's only 1 percent of the cases we obtain a signal.

            You should also remember in the Gothenburg database, that's all intention-to-treat.  The only guidelines have been 20 minutes of recording, and it's all intention-to-treat.  There are no cases taken out of that.  So that really shows how the current technology works in clinical practice.

            CHAIRMAN BLANCO:  All right, thank you.

            Go ahead, Mike.

            DR. NEUMAN:  I would just like to get on the record for the FDA ‑‑ I don't think it needs to be discussed here ‑‑ that it will be important to look at the criteria for selecting the "good" QRS complexes to make sure that it doesn't bias the results in any way.

            CHAIRMAN BLANCO:  All right, thank you.

            Any other comments, going back, now that we have the 20-minutes issue?  Go ahead.

            MS. BROGDON:  Dr. Blanco, Dr. Corrado has tallied some data that may help to answer some of your questions.

            CHAIRMAN BLANCO:  Please, Dr. Corrado.

            DR. CORRADO:  We used the spreadsheet of the results from the entire study that the sponsor provided us, and we tried to evaluate what the precise causes for recordings of less than 20 minutes on the monitor, categories of deliveries that occurred more than 20 minutes after disconnection, and I'll certainly defer to the sponsor, but I will share the information with you that we have been working from.

            For recordings that were excluded, patients who were excluded because the STAN was on for less than 20 minutes, we counted about 111 in the STAN arm and 105 in the control arm.  Of those, 94 in the STAN arm and 76 in the control arm were identified with the word "partus," P‑A‑R‑T‑U‑S, which we interpreted to mean perhaps the patient had delivered.  Is that ‑‑ that is the case?

            So for the STAN arm, out of 111, 94 weren't included because they delivered sooner than 20 minutes, and then 76 out of 105 in the other group.  There was something called signal, which was identified as a problem in 14 of the STAN patients and 21 in the control arm.

            Technical problem, question mark, "TOCO" occurred in four patients in the CTG arm.  The rest of the numbers are very small.  They're less than three or four.  So the overwhelming majority of the category of STAN on for less than 20 minutes was due to delivery.

            CHAIRMAN BLANCO:  So, basically, 60 percent, I mean roughly ‑‑ I don't know the exact, but 70 or 80 out of the 100, a little over 100, were because they delivered before the 20-minute window was up?

            DR. CORRADO:  Correct.

            CHAIRMAN BLANCO:  Okay.  So I think that, I don't know, to me at least it sort of puts an end to the exclusions based on inadequate recordings issue.

            DR. CORRADO:  Well, that is just for less than 20 minutes.  There were also a bunch of cases, 133 in the STAN arm, 130 in the control arm, that were excluded because, for some reason, the monitor was disconnected, and more than 20 minutes elapsed between the time it was disconnected and delivery.

            I will just quickly summarize those numbers for you.  Our analysis showed that there was apparently a signal problem in 80 out of 133 in the STAN arm and 85 out of 130 in the control arm.  There was a category identified as technical that included 20 patients out of 133 in the STAN, 15 out of 130 in the control.

            There is a category that "user error," six and seven, respectively, in the STAN and the control arms, and a category that was just labeled "unknown."  I'm not sure if that is our label or the sponsor's.  There were 19 of those out of 133 in the STAN arm and 16 in the control arm.  So in that category, that is, it was disconnected more than 20 minutes before delivery, the overwhelming majority comes under a category of signal problems.

            CHAIRMAN BLANCO:  Okay.

            DR. CORRADO:  I hope that helps.

            CHAIRMAN BLANCO:  Yes.  Thank you very much.

            Yes, go ahead.

            MS. TOLEDANO:  Just a point of clarification:  Were the signal problems at all correlated with the removal of the electrode to do forceps or ventouse deliveries?

            CHAIRMAN BLANCO:  Go ahead, Dr. Rosen.

            DR. ROSEN:  That was not the common situation.  The more common situation was that they would then would not reapply the scalp electrode and they would switch to a standard electronic fetal monitoring device.

            CHAIRMAN BLANCO:  Thank you.

            All right, any other comments?  Questions?  Issues on 3C?

            DR. IAMS:  Jorge, is this a good place to ask about the exclusions based on cord blood?

            CHAIRMAN BLANCO:  Okay, go ahead.

            DR. IAMS:  Because you're talking about exclusions.

            CHAIRMAN BLANCO:  We're talking about exclusions, so go ahead.

            DR. IAMS:  That was a lunchtime question.  Is there any additional data about that?

            CHAIRMAN BLANCO:  This is a question where you did not have both the arterial and venous cord blood or didn't have, or you didn't have any on some.  Dr. Rosen?

            DR. ROSEN:  I did a quick look at these data where there was the cord artery sample and then there were in the STAN arm about 170, and 190 in the control arm, where there were additional cord vein samples.  In the STAN arm there was no case with metabolic acidosis or cord artery pH of less than 7.05, whereas in the control arm there are two cases, a cord vein, metabolic acidosis, and where the cord vein pH are less than 7.00.

            DR. IAMS:  Okay, my concern before was that babies with metabolic acidosis might have been particularly difficult to sample for some reason.  So it sounds like that was not the case.

            DR. ROSEN:  Well, it's probably true.  There were two cases in the control arm where they couldn't obtain a cord artery sample.  So they were actually excluded from the cord metabolic acidosis group.  They obviously are included, because they had the clinical symptoms, so they are included in the adverse outcome cases, one of those, and the other one is not included.  So you have those represented in the material, in the neonatal outcome of the study.

            DR. IAMS:  Thank you.

            CHAIRMAN BLANCO:  Thank you.  Any other comments on the exclusions?

            (No response.)

            All right, let's move onto 3D then, and this has somewhat been alluded to, but maybe we can spend a little bit more time.  3D, "The intercountry population and management differences."  Anybody want to make a statement about that?

            DR. O'SULLIVAN:  I think that, depending upon the facility on the labor floors as to whether there is central monitoring, one, and, two, that somebody actually looks at a central monitor, which is not common, and, three, if there are alarms and the alarms go off and everybody is still sitting at the desk.  All of these are issues when we talk about bells and whistles.  The whole idea of bells and whistles, and whatever you can do, red lights, et cetera, is to get the staff to pay attention to them, but the answer is usually, "It's not my patient."

            I have grave concerns where you have individual labor rooms with individual delivery room, labor and delivery suites, with a nurse usually taking care of anywhere from four to six, although the floor will tell you it's one to two.  When you actually look at it, between lunch breaks, coffee breaks, union breaks, all kinds of breaks, it's one nurse to four to six patients.  So I have a lot of concerns about it from that perspective in terms of practicality.

            Yet, I also wonder if the educational process somehow or another will make a difference, since there was such a difference in what you all showed before with your education.  I come back to that.  I keep coming back to that because of the importance that I think it has in introducing this in the United States at all.  It's going to be a major educational process, and it's going to be desensitization with resensitization.

            CHAIRMAN BLANCO:  Well, let me play a little devil's advocate.  I mean, I totally understand what you're saying in terms of you have a concern.  You know, you've got a 1-to-4, 1-to-6 ratio, and whether somebody's going to pay attention to this, but that's really not a problem with the device.

            DR. O'SULLIVAN:  No.

            CHAIRMAN BLANCO:  That's really a problem with our system, and actually you might be better, you know, if you've got something that records, so that when they do come around and pay attention to the patient, that they will notice that this event occurred.  They may not have gotten to it on a timely basis, but the other option is simply to central monitor and not showing that.

            I'm not sure that that necessarily to me makes me against the device.

            DR. O'SULLIVAN:  No, but I think in terms of, when you're talking about marketing it in this country, and its useful utilization in this country, to do what it did in Sweden, for example, it is going to be important.  Otherwise, it won't do what it is supposed to do.  It's subsequent failure.

            DR. IAMS:  Jorge, my concerns about transportability don't reflect very well on this country.  I was struck by one of the comments Dr. Rosen made in this beginning of his discussion, that there is a general consensus in Sweden about the interpretation of cardiotocography and some other issues about management.

            I'm not so sure that you could make that statement about any manner of fetal surveillance in labor in this country.  It varies tremendously; interpretation varies tremendously.

            The natural inclination of even the least technological American obstetrician, most of us as we talked about this, did you look at these tracings?  We're all looking at that little EKG thing over there.  We're not supposed to look at that, but that's what American obstetricians do.  We're looking ‑‑ I didn't see the ST wave go up or down.  If I were going to redesign the equipment for an American user, I would take out everything except the ST event marker.  Why do I need the rest of it?  Because we will want to analyze and argue with the computer.  That's what we do.

            CHAIRMAN BLANCO:  Don't forget the light and the bells and whistles.

            DR. IAMS:  Well, I don't need a bell and whistle.  Maybe a little bell, but really take out the data that supports when the ST monitor goes off, because the nature of the American physician ‑‑ I won't necessarily just indict obstetricians ‑‑ is to argue with, well, I didn't see anything go on there.  I'm sure I am right, and the computer algorithm that's taken you 15 years to develop is wrong.  That might be a reason maybe in Sweden why you had to re-educate people.  I don't know.  But in this country it would be a big problem, I think, something you could easily fix.

            CHAIRMAN BLANCO:  Well, let me pin you down a little bit, and Dr. O'Sullivan as well.  Is that an issue that's sufficient enough that you are concerned about the introduction of this device into this country?

            DR. O'SULLIVAN:  Absolutely.

            CHAIRMAN BLANCO:  Really?

            DR. O'SULLIVAN:  Because, again, if the device is introduced into this country without that whatever it is secret he has about education, and that education is not done universally, widespread, it is doomed to failure.

            CHAIRMAN BLANCO:  Jay?

            DR. IAMS:  The second issue I brought up about changing what is displayed, I think you could change that fairly easily.  The larger issue is trying to get American physicians to buy into (a) there's a problem with our current system and (b) this is the answer.  I think, unfortunately for you guys, after all your fine work, the answer to that is probably going to be an American study.

            MS. TOLEDANO:  The No. 2 question on my page, from reviewing all these materials last night, was, how will the knowledge be disseminated in the United States?  Because I definitely agree with the importance of the question, but, if I recall correctly, Ms. Daws-Kopp was telling us that the sponsor is prepared to set up a program similar to the Centers of Excellence that they have in Europe, so that they would be coming into harvest, disseminating the knowledge, doing the educational program, reviewing cases, and monitoring the way that this technology is marketed and distributed in the United States.  So, to me, that alleviates some of the concern.

            DR. SHARTS-HOPKO:  I share the issues that Dr. O'Sullivan raised, and we mentioned earlier a concern about, how are women attended?  When a woman is attended by a midwife in the United States, typically, that person stays with her pretty constantly throughout the labor and the delivery.  That's only 1 percent of our births that are midwife attended.  I don't know if it is similar in your situation.

            The other ‑‑ I just forgot the other issue that I was going to raise.  It will come back later.

            DR. RINGEL:  Just a couple of comments:  I agree with a lot of what has been said there, but I think we are being a little bit harsh on all of our colleagues around the country on their adaptability to use computers.  You guys accept that heart rate printout where it shows you a line that's the heart rate.  As a cardiologist, that's anathema to me.  I want to see the QRSs.  I don't want to see something that tells me the heart rate is actually 110.  I want to measure it myself, but you have learned to accept something that tells you the heart rate is 110.

            So my guess is that the younger people coming out would learn how to accept the ST or the T wave QRS ratio is actually what the computer says it is, as we all get more comfortable with computers.  So I'm not sure that that's ‑‑ you know, to say that we're so stuck in our ways in this country that we couldn't accept that, that that's really valid ‑‑ I do happen agree and I feel very uncomfortable with any piece of medical equipment that does not have appropriate alarms, so that when it is reading a value that they consider out of bounds, the fact that there is no alarm I think makes it personally unmarketable in this country, because you can't have a piece of medical equipment attached to a patient without alarm parameters.  So in that way I agree.

            DR. WOLFSON:  Yes, but a lot of our fetal monitors, you can have recurrent late decelerations; nothing goes off.  You're supposed to look at it.

            DR. RINGEL:  Ah, but that's an interpretation.  I'm being told that people interpret this differently.  They are specifying an out-of-bounds value for their ST or T waves ‑‑ excuse me ‑‑ for their T wave QRS ratio.  That is part of their premise.  They have an out-of-bounds ratio, and that they have an automatic detection.  So if we don't want the alarms, then we go back to displaying just the QRS pattern and, like any electrocardiogram, where we would then interpret it ourselves, or if you're going to say there's automatic detection and you set boundaries, and they have set boundaries, then I think you have to have an alarm.

            DR. WOLFSON:  I would agree.  I would like to see some type of alarm because part of what I see the power of this tool is, is that it's another factor that's going to come back and alert people to go back and re-evaluate the status of the patient, the status of the tracing in the context of interpreting it.

            Because the other feature relative to outcome is, what's the time to delivery?  If the time to delivery is felt to be under 5 to 10 minutes, or even some people might even argue as far as 30 minutes, then, clearly, you're going to go for a vaginal delivery, and the patient is not going to fall into an operative category.  Otherwise, you're going to go in the other direction.

            CHAIRMAN BLANCO:  I think we have sort of gotten off 3D here and we're going back to the fact that, again, and the word I was looking for before was to be so provincial, but I think you need to Americanize the device a little bit as to what the practice here is.

            But let's try to go back into 3D and 3E.  In a way, we've also talked about retraining.  I think training is coming out as an extremely important adjunct to this particular device.

            But, Nancy, go ahead.

            DR. SHARTS-HOPKO:  Thank you.  I remembered my point.

            The Swedish study looked at applying the device to women that had some risk indicator, and we monitor everybody pretty much here.  So we don't know what kind of false findings we're going to find in a normal population, a general population.  I'm concerned about that.

            DR. BROWN:  Another point is we don't know, in terms of comparing the populations, the incidence of pre-existing conditions that might be undiagnosed or unknown or even that are known.  Someone mentioned the contrast in prenatal care.  Was there a comparable rate of women with diabetes, IUGR, pregnancy-induced hypertension, smoking, drug use, whatever, in this population?  I mean, I have some concerns about that.

            Then I think, getting back to Dr. Iams' point, if you're talking now about the management differences, I think the question about, can this be marketed to reduce the operative rate is a valid point, because if the indication of this is that your company wants to market as a way to avoid operative deliveries by letting you use this for that gray area of fetal heart rate tracing, and sit on that patient, that needs to be very clear.  Because to me, just reading it, not having done obstetrics in a long time, it is a little confusing.

            Is that really the primary point of this device, to allow you to sit on the questionable tracing, if there is not an ST flag on the screen or an alarm?

            CHAIRMAN BLANCO:  I think that was sort of what was addressed, that that's one of the ways that the fetal heart monitoring is not reassuring, is non-reassuring; you then look at the ST events or ST changes.

            DR. BROWN:  Because that will require a whole complete change of culture and education.

            CHAIRMAN BLANCO:  Well, I think some people think so.

            The other issue that you brought up, and I think it's important to bring up, because it's been talked about immeasurably, is when it applies to Caesarean section rates in terms of comparing, say, Ireland versus the United States, the United States is a very heterogeneous population.  I'm sorry I've never been to Sweden, but my understanding is that it is a very homogeneous population, very much like Ireland.

            So other than your percentage ‑‑ and I don't have my hands on it right now ‑‑ of folks that were from the Middle East and Africa and former Yugoslavia, and most everybody else is pretty homogeneous, and that percentage was 20 percent ‑‑

            DR. O'SULLIVAN:  Twelve percent.

            CHAIRMAN BLANCO:  -- 12 percent.  So I think that that is a concern in trying to extrapolate the data from Sweden into the United States.  I think there really are different populations.

            DR. O'SULLIVAN:  But the other thing to keep in mind, too, when you're looking at that, at least based on our own experience in this country, is the initial immigrant who comes in is generally the healthier individual.  That's why they're here.  And they generally do better than our own U.S.-born individuals do.

            CHAIRMAN BLANCO:  I'm sorry, but I don't understand your point.

            DR. O'SULLIVAN:  I was just saying that because somebody is an immigrant into the country doesn't mean that they are going to perform more, that something's going to happen that's going to be more complicated in terms of the obstetrical outcome for that index pregnancy.

            CHAIRMAN BLANCO:  Okay, so what you're saying is that, even though they have a 12 percent population that is not what would be classified as homogeneous, they may not be as homogeneous as some of the population in the United States?

            DR. O'SULLIVAN:  Oh, absolutely.

            CHAIRMAN BLANCO:  Okay.  Any other comments?  Yes, Ma'am?

            MS. MOONEY:  I think one of the things we should keep in mind with this particular area question is that, if we think the data we have in hand essentially supports the safety and effectiveness, answering these types of questions may be the things we talk about in the post-approval study area.  So to keep cognizant of the fact that we can separate out the inherent safety and effectiveness in the data we have and make that judgment call versus what we could look at entertaining the post-market study.

            CHAIRMAN BLANCO:  While I don't disagree that you may need to look at it in a post-market study, I'm not sure ‑‑ I think the point is that there is some concern that the difference in populations may mean that the safety and effectiveness may not be the same.  We're not talking about something that's showing an 80-versus-20 percent difference.  We're talking about something that shows a very small difference, significant but very small.  So I think it may still be applicable, an applicable point.

            DR. WOLFSON:  Jorge?

            CHAIRMAN BLANCO:  Yes, sir?

            DR. WOLFSON:  Going back to the differences in the populations, the device, it says, "indications for use."  "Is indicated where there is a planned vaginal delivery and there is need for close surveillance during labor or there are maternal disorders and/or uterine placental dysfunction with potential adverse influence on fetal oxygen and nutrition supply, or deviation from the normal course of labor, including induction, augmentation of labor," et cetera.

            So I think that the actual use of the device in terms of who it is going to get applied to, when you turn on the ST segment, let's say, is probably going to be the same in both countries, in that you're using it as an augmentation to standard monitoring.  A great deal of our clinical monitoring that we see ‑‑ I guess I'll speak for my own venue in Colorado Springs.  External fetal monitoring is what is used early in labor, clearly, because the membranes aren't ruptured generally, and even if they are it is still the simpler thing to do, and our labors are generally attended by the nursing staff.  The obstetrician is called in when there is a concern of what is being seen on the strip.  So the nursing staff is making the decision, and they are generally going to utilize external monitoring.  When internal monitoring is required, then they're going to call the obstetrician in.

            So I think that the difference in makeup of the populations relative to whether it is the heterogeneity of the American population, that as long as we are using it where we think there is a specific need for this additional information, it's not going to differ from population to population.

            CHAIRMAN BLANCO:  All right, let's move on to 3E, retraining during the Swedish randomized controlled trial.  I think we've kind of addressed that.  Does anybody want to summarize it or say something?

            DR. IAMS:  I don't want to summarize it.  I have one question about it.

            It's this sort of event that made me wonder why the investigators did not decide, since we didn't execute the interventions according to our original plan, why didn't you restart the study basically and then get your end from that point on?  It's powerful actually that you were able to achieve the P values you got when the first, what, third of the study was not totally compliant with the original protocol.  So why not repower it at that point?  There must have been some thinking along those lines at some point.

            CHAIRMAN BLANCO:  I think there is some disagreement with you on that.  Do you want to make a comment?  I saw you shaking your head.

            MS. TOLEDANO:  You can't do that.  You now have to turn ‑‑

            DR. IAMS:  Yes, you can.  You can do what is basically a run ‑‑

            CHAIRMAN BLANCO:  Wait a minute.  Let her have a chance to finish.  Go ahead.

            MS. TOLEDANO:  That would be so sorely frowned upon, because basically you would have to report the first study, that you terminated it early for the protocol deviations.  Protocol deviations happen in all sorts of studies that are used by any governing body in this country and in the European Union and other areas as well.

            As far as it is ingrained into me with that all the ethics training that we are required to take during studies and holding NIH grants, you can't just stop a study because you don't like the protocol deviations that people make.  If there are protocol deviations, you do your interim look.  You can try to fix it, but you finish up the study.  You don't just stop the study.

            DR. IAMS:  I don't disagree with that, but all I can say is that you can, because we do it all the time in NICHD-funded trials ‑‑

            MS. TOLEDANO:  Well, we don't do it in cancer ‑‑

            DR. IAMS:  -- look at the performance and then decide that you do not have enough in that has followed the protocol.  You just continue for longer; that's all.  You don't really deny the existence of the first 1600.

            MS. TOLEDANO:  Right.

            DR. IAMS:  You report them.  You're completely honest, but you say, "We haven't achieved our end, and so we're going to have to continue this study in order to make sure that we have the results."

            MS. TOLEDANO:  They did that actually.  They recalculated the total sample size.  It wasn't based on protocol deviation.  It was based on the baseline rate of metabolic acidosis.  So they did go at the 1600 and calculate that they needed more than the original 3200.

            DR. IAMS:  Okay, I missed that part, but ‑‑

            MS. TOLEDANO:  Yes.

            CHAIRMAN BLANCO:  And the other point is they showed more of the difference now.  So I think they showed that, with the retraining and re-education, they actually did better off than without it, before with the violations ‑‑

            DR. IAMS:  Yes, that answers my question.

            CHAIRMAN BLANCO:  Any comments on that?  Anybody want to add anything?

            Well, I'll add kind of a summary because  I think it's been said before, but just to make sure FDA hears it, I think this device is going to need a fair amount of retraining before it hits the market, and not after 1600 patients.  I mean, I think there is some education that needs to go with this.  It can't just be sold and expect everybody to use it and know what they're doing.  I think it has some inherent necessities for a significant amount of training.  There's some precedent for that with other devices.  So I think that we need to recommend that, and I see some shaking heads saying yes.  So I think that's the feeling of most of the panel.

            DR. EGLINTON:  Jorge?

            CHAIRMAN BLANCO:  Yes?

            DR. EGLINTON:  Can I get some retraining?

            CHAIRMAN BLANCO:  Yes, please do.

            DR. EGLINTON:  May I ask another question, please, about the card.  I want to know, I have an intermediary CTG, and I want to know now if I have an episodic T/QRS rise greater than 0.15, or a baseline rise greater than 0.10, how do I figure that out?  Looking at the lower scale, the T/QRS scale, that is rather tight.  That's pretty subtle, trying to figure out or move over one column over to abnormal CTG; 0.05, how am I going to figure that out?  Does the machine tell me that somehow other than just visually?

            CHAIRMAN BLANCO:  Gary, for the sake of time and the other questions, and answering some of the questions that we have asked the sponsor, I think it's pretty clear, or should be pretty clear, to FDA and to the sponsor, that that card is confusing and may not be appropriate, and needs to be worked on significantly.  Maybe we can touch base on the details at a time when there is more time.

            DR. EGLINTON:  But this is central.  This is how you make your management decision.  I mean, if it rises, if the baseline rises more than .05 with an abnormal CTG, that is the reason to intervene.  It's really central.

            CHAIRMAN BLANCO:  All right, Dr. Rosen?

            DR. ROSEN:  The computer identifies these changes, and it will tell you in detail what extent there is in change, whether it's more than .05 or it's more than .10, and whether it's episodic of more than .10 or more than .15, and that's written and it's flagged on the screen.  It is written on the paper.  There is an ST log where all these events are recorded.  That log is available to the operator throughout the delivery process.

            So the issue of interpreting the ST change is very much a reflection of being capable of reading the statements made by the computer.

            DR. EGLINTON:  Is that on the review mode?  Is this illustration, figure 2.2.3, is that what you're talking about, on the review mode in the upper right?

            CHAIRMAN BLANCO:  I'm sorry, Dr. Rosen, but we kind of need to hear and see what you're saying for the record.  So if you would go ahead and go back ‑‑

            DR. ROSEN:  So on the user manual there is a graph showing how the user is informed about a significant event, where what's called an event log window is available, where all these events are logged as well as there is a keyboard, so you can key in specific information that you tell, "I recognize this," and so on.

            MS. LUCKNER:  Dr. Blanco?

            CHAIRMAN BLANCO:  Please.

            MS. LUCKNER:  At least four times the question has been raised that the panel participants are unaware of what the machine does automatically versus what the individual practitioner must do observing it.  So my recommendation put on the FDA record is that part of that must be more clearly specified.

            I mean, I read several times, just trying to figure out, do I have to do it or does the machine do it?  I thought that was just me in Toledo, Ohio, and now I'm very comforted to hear that the rest of my panelists have the same questions.

            CHAIRMAN BLANCO:  Okay.  You would like to say something?

            DR. ROSEN:  It is very important that the user is fully informed about how the system operates. I can tell you there has been quite a lot of resources in the education package.  You should remember that we have a certification process which takes the users through all these aspects.  This is very essential.

            I would not like to see the physiology behind this to fail because the user would not understand how to use the equipment and how to interpret these events.  Please remember that those guidelines you have in front of you, they are like a pocket, just a quick memory thing.  You should, I hope some of you have been able ‑‑

            CHAIRMAN BLANCO:  Dr. Rosen, I'm sorry to interrupt you, but you've also got to remember you've got some fairly, supposedly, bright people up here, and you know we're frankly confused, okay?  So let's let is loose on the United States, and you're going to confuse a lot more people.  Okay?

            DR. ROSEN:  Have you consulted the educational material yourself?

            CHAIRMAN BLANCO:  Well, no.  I mean, I haven't been educated by what you have, but we don't have that on here.

            DR. ROSEN:  Yes, you have.

            CHAIRMAN BLANCO:  We're just making suggestions that there is confusion.  Okay?

            DR. ROSEN:  Okay.

            CHAIRMAN BLANCO:  And there are important things about how you're labeling that may be regionally-specific, issues that you need to address.  I think we have made it very clear about the need for significant education process on there.  Okay?

            DR. ROSEN:  For the record ‑‑

            CHAIRMAN BLANCO:  I'm going to break my rule and I'm going to recognize Dr. Devoe because I'm sure that he has something valuable to add at this point that is short in time, because we are starting to run a little late.

            DR. DEVOE:  I'll put on my Yankee hat and be very brief.

            CHAIRMAN BLANCO:  Please state your name.

            DR. DEVOE:  Dr. Lawrence Devoe, Medical College of Georgia, consultant for Neoventa.

            What has probably not been properly emphasized is that out of the box there is a very, very specific, detailed, organized, orderly, and properly-written tutorial that really must be done before you flop up the first power switch on any of these devices, and to presume that you can use any device like we have erroneously done in the fetal monitoring world for the last 30 years without doing this is really missing, I think, a very important concept.

            There is a tremendous amount of education that goes into the use of a device that has this internal degree of sophistication.  It does two things.  It teaches you what this device does that conventional monitoring doesn't do, and, in turn, as you have seen by some of the clinical experiences, it makes the user better at the basics of electronic fetal monitoring, which is something we have been derelict in since fetal monitoring was first perpetrated in the United States.

            CHAIRMAN BLANCO:  Thank you.

            Gary, any other comments you want to make?  If not, can we move on?

            DR. EGLINTON:  No, that answers it.  Thank  you.

            CHAIRMAN BLANCO:  All right, let's move on.  Let's try to see if we can do No. 4 before we take a break.

            No. 4:  "In addition to a Swedish RCT, the sponsor has conducted the following clinical studies using the STAN monitor.  To what extent do results from these studies support the safety and efficacy of the STAN monitor?  The Plymouth randomized controlled trial, the EC study, and the City of Gothenburg Observational Study?"

            Anybody want to take a stab at that to start off with?  Go ahead.  Dr. Ramin?

            DR. RAMIN:  With the Plymouth randomized clinical trial, it was not designed or powered to look at the primary endpoint that was used in the Swedish randomized control trial, that being metabolic acidosis, but it did show, at least in that trial, a reduction in operative delivery, both vaginal delivery and Caesarean delivery for fetal distress.

            For the EC study, here there were in that trial 12 cases, 11 of whom had evidence of hypoxia or asphyxia associated with an ST change.  There was one case out of the 12 that had low Apgars scores and encephalopathy and cerebral palsy, that had a normal fetal heart rate tracing and a normal ST waveform.

            The Observational Study for the City of Gothenburg, again, with experience, or with experience with this device, it did show reduction in the cases with metabolic acidosis as well as operative delivery.

            DR. RINGEL:  It was with the City of Gothenburg that I raised the question before.  Maybe I am misreading the slide.  This is on page 29 of Dr. Rosen's presentation.  That's why I asked the question before.

            Maybe I've got this wrong, but it looks like, with experience, the obstetricians in Gothenburg did better with plain CTG than they did with STAN in detecting metabolic acidosis, with experience.  Do I have that wrong?

            CHAIRMAN BLANCO:  Dr. Rosen, can you come up?

            DR. RINGEL:  The rates of metabolic acidosis looked much lower than with STAN.  In fact, it's only for the total and for the CTG that there is a statistical difference demonstrated and none with the STAN.

            CHAIRMAN BLANCO:  Anybody?

            DR. DEVOE:  Devoe again.

            One of the things that should be appreciated is that the selection of cases in the CTG arm or risk stratification were much less likely to be high-risk cases than in the STAN arm.  In other words, these weren't cases that were randomly assigned one or the other in general.  One of the obstetricians here can probably speak to that, if there are further questions later on.

            But the standard CTG arm represented the lower echelon of risk patients.  One would empirically expect to have a higher rate of complications in the STAN arm in terms of metabolic acidosis

            DR. RINGEL:  But you're presenting this data in support of the efficacy of STAN in an observational trial, and the rates of metabolic acidosis in the first half of the trial are essentially the same, and then they go down substantially in the CTG arm, but not in the STAN arm.

            DR. DEVOE:  They declined in both arms, but also remember that there is also an educational impact in the area of fetal monitoring in general that would have affected both arms.

            DR. RINGEL:  But there's no little asterisks over the STAN arm to indicate statistical significance while there are asterisks over the total in the CTG arm.

            DR. DEVOE:  Right, because they got better.  They got better as time went on.  I mean, that's part of the educational impact.

            DR. RINGEL:  But the STAN didn't get better.  So then you could have just educated people, continued to use CTG, and you would have dropped your metabolic acidosis rate to .1 percent.  Am I reading this wrong?

            CHAIRMAN BLANCO:  No, you're reading it right, except the number you said is not right.  I think these are fractions of a percent.  It went from, in the CTG arm, from 1 percent to .01 percent.

            DR. RINGEL:  Right.

            CHAIRMAN BLANCO:  Okay.

            DR. RINGEL:  But the STAN arm didn't statistically ‑‑

            CHAIRMAN BLANCO:  And the STAN arm went from a little under 1 percent to about .6 percent.

            DR. RINGEL:  Yes.  So all you had to do was wait a year and things would get better without STAN, right?

            DR. DEVOE:  No.

            DR. RINGEL:  No?

            DR. DEVOE:  They're still training on STAN.

            DR. RINGEL:  Okay.  So it's the paying attention to fetal ECG that's making the difference whether it's STAN or not?

            CHAIRMAN BLANCO:  Go ahead.

            DR. ROSEN:  Could I address this here?  I think it's one important aspect here, and that is to say, what happens if you've got obstetricians and midwives used to having ST information available, and then there is a group where it is not available?  Clearly, that group, according to our practice of using the fetal scalp electrode, would be a low-risk group, because the high-risk cases, they would have a scalp electrode applied:  those with inductions, those with augmentation of labor, those with any antenatal history.

            So then you have a remaining group now being monitored with heart rate only.  I believe what we have seen is the impact of more accurate assessment of the heart rate trace.  We can ask the midwife here to explain a little bit about the process.

            CHAIRMAN BLANCO:  Well, but I think that is what he is saying.

            DR. RINGEL:  Don't you see what I am saying ‑‑

            CHAIRMAN BLANCO:  That's what he's saying.

            DR. RINGEL:  -- is your Gothenburg trial would suggest that just by educating your care providers, you can achieve the results you want to achieve with an expensive apparatus.

            DR. ROSEN:  On the other hand, this is done taking the high-risk cases out of the group being monitored with heart rate only.  So, basically, you would expect a very low incidence of interventions ‑‑

            CHAIRMAN BLANCO:  But there's still a change, and let's go ahead.  I think you have made the point that there is some data from the Gothenburg study that shows that there may be a decrease just by reading the fetal heart rate monitor.

            DR. ROSEN:  Right.  Oh, yes.

            CHAIRMAN BLANCO:  Okay, thank you.

            MS. TOLEDANO:  I just wanted to clarify some numbers on this same graph.  So I'm seeing with the STAN, in the first half of the graph, an "N" of 668 and approximately 1 percent having the metabolic acidosis.  So that would be seven babies.  Then on the righthand side of the graph, an "N" of 764 with a .6 percent.  So that would be about five babies.  Is that what we're looking at, going from seven children to five children?  There is no way you can get statistical significance there.  So that's why there are no little stars.

            CHAIRMAN BLANCO:  But I think the point is still that ‑‑

            MS. TOLEDANO:  They can't.

            CHAIRMAN BLANCO:  -- in the other arm they did.

            MS. TOLEDANO:  Right, in the other arm they do decrease it by the better education ‑‑

            CHAIRMAN BLANCO:  They do have larger numbers.

            MS. TOLEDANO:  And we would benefit from that.

            CHAIRMAN BLANCO:  Okay, we're going to move on, Dr. Rosen.  Okay, we're going to move on.

            So I think the question was asked, actually, in a broader issue, which is, is the data of these other studies that were not specifically aimed at testing the hypothesis that the sponsor was putting forth, does it support or not support, and can it be used in some way to improve our decisionmaking in terms of approval of the device or lack of approval?  Bring it back to that.

            Yes, Ma'am?

            MS. BROGDON:  Dr. Blanco, I need to point out that in the written copy of your questions we left off the Nordic study.  It's on the slide up there.

            CHAIRMAN BLANCO:  Okay, thank you.

            All right, so there is some differences in the Gothenburg study that may not be supported, is what you're saying?

            DR. RINGEL:  That's all I'm suggesting, yes.

            CHAIRMAN BLANCO:  Okay.  All right, any other comments on this?

            (No response.)

            All right, we're going to ‑‑ it's 2:22 right now by the official clock.  We have a 15-minute break.  So we will start promptly at 2:37.  Thank you.

            (Whereupon, the foregoing matter went off the record at 2:24 p.m. and went back on the record at 2:39 p.m.)

            CHAIRMAN BLANCO:  All right, let's go ahead and begin.  We're still going through the discussion process and answering the FDA questions.

            We are at Question 5.   It's labeling and training.  It's:  "The sponsor proposed the following indication:  Use of the STAN system is indicated when there is a planned vagina delivery and there is a full-term singleton fetus in a vertex presentation, and there is a need for close fetal surveillance during labor, or there are maternal disorders and/or uteral/placental dysfunction with potential adverse influence on fetal oxygen and nutritional supply or there is deviation from the normal course of labor, including induction, augmentation of labor.

            "Does the PMA data support this indication for use?  Do you have any suggestions for modifications?"

            I think, Dr. Ramin, you had some comments?

            DR. RAMIN:  Yes, I have a comment, and, again, it is looking at the differences in how we monitor patients in labor in the United States as opposed to Sweden.  Probably about 90 percent of our patients, at least at our institution in Houston, get monitored, whereas I think the number was significantly less in Sweden.  I guess it would be looking at the indications of kind of narrowing down specifically which fetuses you're going to use this device in.

            CHAIRMAN BLANCO:  Well, would you suggest that the device be utilized in patients who have a high-risk condition, so that they are at higher risk for metabolic acidosis?

            DR. RAMIN:  I think the device should be indicated and make it a little bit more tighter in using it in the high-risk pregnancy, perhaps IUGR fetuses, for example; patients that have a non-reassuring fetal heart rate, for instance.

            There was a question raised about whether or not it could be utilized ‑‑ I think one of the other panel members brought up about breeches.  So this is for vertex presentation, but I don't know if it is applicable to perform a breech presentation.

            DR. RINGEL:  I was going to comment on that.  If you look at their usual manual, they actually give instructions for using it in breech presentations.  So I didn't understand, if it's not applicable, why it is in the manual.

            CHAIRMAN BLANCO:  Okay.  Any comments from any of the other panel members?  Any agreement or disagreement with, that because of the changes, the differences in populations, that maybe a stronger indication, looking at high-risk conditions, needs to be utilized, rather than a somewhat broader wording, as it is now?

            DR. O'SULLIVAN:  I have two questions.  No. 1, do we know anything about its use in an otherwise ‑‑ in a patient's, in a population who has absolutely no problems or no detectable problems in labor?  That's my first question.

            My second question has to do with the indications here as they are listed.  I mean, in this country, where it says, "There is a need for close fetal surveillance during labor," I think we are in a situation from the medical/legal perspective that every baby needs close fetal surveillance in labor.  So that to me means everybody.

            CHAIRMAN BLANCO:  Well, but I think that's Dr. Ramin's point, that that probably is too broad and that needs to be sharpened up to more high-risk patients.  You would agree with that or not?

            DR. O'SULLIVAN:  I would agree with that, but I would also say, is the information out there, has a study been done to look at women who are otherwise uncomplicated, and what the impact is in that population?  I would expect it to be excellent, but, I mean, what I am saying is, are there things from the perspective of the false positive situation that we have yet to see?

            CHAIRMAN BLANCO:  Okay.  Any other comments?

            (No response.)

            Okay, so I guess two issues that were brought up is that, and I don't know if there is a consensus from the panel, so I would like a couple more people to say something, that the indications need to be tightened to include folks with high-risk conditions as opposed to the more broader wording that is currently here?

            Then the other issue is concern or at least some information from the sponsor about in a low-risk population, if it has been used for that, has there been some evidence of significant number of false positives?

            Am I paraphrasing you correctly, Dr. O'Sullivan?

            DR. O'SULLIVAN:  Yes.

            DR. RAMIN:  It's kind of the question that I had asked, and that is, you have normal tracing; what is the false positive rate in those individuals?

            CHAIRMAN BLANCO:  Okay.

            DR. RAMIN:  I have another comment just regarding the deviation from the normal course of labor, including induction and augmentation of labor.  I can't speak for all institutions in the United States, but I do know that we don't necessarily have to do internal monitoring for an induction or augmentation of labor.  We can do external monitoring.

            I guess in order to use this device, it has to be internal monitoring.  So my recommendation would be to have that as an indication.

            CHAIRMAN BLANCO:  All right.

            Jay, you have a question?

            DR. IAMS:  I just didn't understand the last part of what Sue said there.

            CHAIRMAN BLANCO:  Yes, what did you say?

            DR. RAMIN:  Well, part of the indication would be deviation from a normal course of labor.  So if you have a woman that you're inducing, that would mean that you have to insert ‑‑ this is an internal monitor, the scalp electrode.  I think there are cases where you can induce somebody's labor with an adequate external tracing, not have to do internal monitoring.

            CHAIRMAN BLANCO:  Well, but wait a minute.  But, Sue, the issue is not that you have to, because you don't necessarily have to use this technology.  I mean, the question is ‑‑ I would phrase it differently ‑‑ is induction of labor sufficient indication to use the STAN?  Is that enough of a condition that makes a patient non-routine to use the STAN, to justify the use of a STAN?  Not that, okay, I'm inducing somebody, so I've got to use the STAN.  I mean, I think the question is the other way.

            Does that make sense?

            DR. RAMIN:  That makes sense.

            CHAIRMAN BLANCO:  Yes.  Is induction of labor, augmentation of labor, enough of an ‑‑

            DR. RAMIN:  Indication.

            CHAIRMAN BLANCO:  -- indication, a high-risk condition, it's an indication to insert an internal and use the STAN monitor?  Is that what you meant?  Did that clarify it?

            DR. IAMS:  Yes.

            CHAIRMAN BLANCO:  So what do you think?

            DR. WOLFSON:  Yes.

            DR. IAMS:  Do I think the labeling, if modified as described, would be appropriate?

            CHAIRMAN BLANCO:  Would it be appropriate to use the STAN monitor in patients just simply because they're being induced?

            DR. IAMS:  Oh.  No, I wouldn't think that every patient being induced would need a STAN monitor, no.  Some patients, if ‑‑ we all probably have slightly different indications for use of an internal fetal monitor, but in recent years the indication for an internal fetal scalp electrode is often going to be coincident exactly with the kind of patients that were studied in this trial.  The low-risk patient who is being watched in order to be careful with every baby is watched in our unit with an external monitor unless an adequate tracing can't be obtained.

            So if tracing is then normal, I wouldn't think you would need the STAN, but if you have any abnormality of the tracing or any abnormality of the course of labor, or the antepartum conditions there were mentioned, then I think you would have an indication for a STAN.

            CHAIRMAN BLANCO:  Dr. Wolfson, I heard you say "yes."  So would you like to comment?

            DR. WOLFSON:  Well, I agree with what Jay just said.  I think that most inductions can be done without the need for internal monitoring, if not probably nearly all of them.  I just wouldn't want to ‑‑ let me back up.

            I believe the evidence that was presented suggested that one of the greatest, one of the complications of induction, which is hypertonus, will probably be detected through the STAN.  So if it is missed from clinical assessment, the system might alert the person.  So I a value of utilizing it in patients who are under induction.

            CHAIRMAN BLANCO:  All right, anyone else care to comment?

            (No response.)

            I take it from the silence on the other issues, most panel members are in agreement that tighter indications to make it more clear that this is to be used in high-risk conditions until there is more information at least, until there is more information on its use in low-risk patients, would be the way that the panel would see it, would suggest to FDA that they label the use of the product.

            Anybody want to disagree with that?

            (No response.)

            No?  Okay.  Any other comments on No. 5 then?

            (No response.)

            All right, having no other comments on No. 5, we'll move on to No. 6.  "Are the professional labeling, users' manual, and the training materials, Section 1A of the panel package, provided by the sponsor sufficient to ensure appropriate use of the STAN system?"

            Anybody care to open that up?

            DR. IAMS:  Jorge, I'll take a dive into that.  I think that's a big problem in the United States.  You just saw the panel members up here read your written information, put our disks into our computers, and think we're following things.  I can sense your disappointment that we didn't quite appreciate your product the way it perhaps deserves to be appreciated.  And we've had not just a day or two to do it, but we've had weeks, months, even to do it.

            So I'm not sure that the labeling and training materials are sufficient.  I think this really isn't the province of the FDA, but it's a concern to imagine intrapartum fetal monitoring data entering the current United States atmosphere, with all of its legal concerns and all of our arguments with each other.  I don't mean to suggest that we're resistant to change.  We simply don't agree, and the young colleagues don't agree with each other; they don't agree with the old ones.  We simply don't have an homogeneous approach across the country or even in the same department to how tracings ought to be interpreted.

            So I'm not clearly quite sure what to tell the sponsor of this product how to prepare the marketplace, but there needs to be some preparation of the marketplace for this product.

            CHAIRMAN BLANCO:  Jay, the only thing that I disagree with you is that it's not the province of the FDA.  It is actually the province of the FDA, by controlling how the product is labeled, how the indications for the product are made.  If somebody from the FDA wants to contradict me, feel free, but in my experience it is the province to say, "This is a problem."  Say, "This is a terrific device.  We think it ought to come on the market," whatever, "but it can be misused if people don't understand, or used inappropriately" ‑‑ that may be a better terminology ‑‑ "used inappropriately without the appropriate educational level, labeling, and indications," and then it is up to us, if that is the issue, to suggest not in crossing the "T's" and dotting the "I's" detail, but in some general detail, what types of education, labeling, and indications need to be put on the device in order to market it safely in the United States.

            So I would say that is the only thing I disagree with.  I think we can do that, if that's what you want.

            DR. IAMS:  Well, that's good to hear.  What I meant was I didn't think the FDA could really mandate that stuff.

            I guess the other comment I would make along those lines is that the only time obstetricians in America line up and behave in a fairly uniform fashion is when ACOG says something:  This is how it's done.  So you can't really ‑‑ I don't know if we can mandate that or not, either, but ‑‑

            CHAIRMAN BLANCO:  No, I don't think they can mandate that.

            (Laughter.)

            DR. IAMS:  But getting input from ACOG would be a major step in the direction of preparing the population for following the rules, because as long as they're not inside that process, you'll either get no use or inappropriate use.

            CHAIRMAN BLANCO:  All right, any other comments?

            DR. RINGEL:  Just, again, the issue of breech, I was surprised to see that they tell you how to use it for breech cases.

            CHAIRMAN BLANCO:  Okay.  Obviously, with the current indication of vertex, that shouldn't be in there.

            Dr. Wolfson, I think I saw your hand up?

            DR. WOLFSON:  Yes, I was just going to comment.  I like the Fetal Surveillance Manual.  I read it cover to cover.  I found it ‑‑ I loved the diagrams and I loved the choice in the terminology, and I think this is a wonderful educational tool for anybody in obstetrics and gynecology, whether they're in nursing or whether they are physicians or midwives.

            I, unfortunately, did not have the experience in being able to open their educational tool on my machine.  My machine kept bombing.  It just wouldn't let it open.

            So I think that education is a key point is this particular product, and I don't feel that I am sufficiently expert in their educational process to say that what they have is currently insufficient to train someone who is not knowledgeable in this technology to become knowledgeable in this technology.

            One of my original questions was, what is the endpoint when you know that someone is adequately trained?  Is the software built in such a way where you have to demonstrate a specific number of cases where you have properly interpreted them, and whether those are randomized and blinded to the individual?  If that is not the case, that would be one of my recommendations, because I think we have to have on an individual basis, if you're doing this ‑‑ physicians do what they feel most comfortable with, and the best outcomes occur when there is consistency, which we all know.

            Access will come and utilization will come with that comfort level.  What is paramount here is creating that comfort level, and I think that is going to come through the educational process.

            CHAIRMAN BLANCO:  Well, I think they obviously re-educated their personnel when they were doing the study.  From what we have heard, I think the sponsor certainly has a commitment to education ‑‑

            DR. WOLFSON:  Yes.

            CHAIRMAN BLANCO:  -- and to making sure their product is used appropriately.  So I think the issue is, to again reiterate, and I think we have said it multiple times today, that there needs to be some significant educational component that goes along with the use of the device to ensure that it is used appropriately.

            Yes?

            DR. O'SULLIVAN:  The one thing I would also suggest in this manual, which I am sure both Mike and Gary have pointed out, and I suspect Dr. Rosen knows, is that the teaching tools using the fetal monitor tracings in here are the European as contrasted to our 3 centimeters a minute.  Fortunately, I am familiar with that, so it was not difficult, but for students it will be difficult.

            CHAIRMAN BLANCO:  Well, I think that goes under what we said of a little bit of Americanization for the American market will probably be very helpful to their product.

            All right, any other comments on No. 6?  Anything else that anyone would like to say?

            (No response.)

            All right, let's move on to No. 7.  "If the panel votes to recommend approval of the STAN monitor, is there a need for post-approval studies, and if so, what is the purpose of such studies and what are the key elements of the study design?"

            Who would like to begin that discussion?  Gary, I haven't heard from you in a while.

            DR. EGLINTON:  I think this question, I don't need to get to this question.

            CHAIRMAN BLANCO:  Okay.

            DR. EGLINTON:  I stop short of this question.

            CHAIRMAN BLANCO:  All right.  Well, then, let's move on to someone else and see if they have something to the question.  Dr. O'Sullivan?

            DR. O'SULLIVAN:  I feel the same way Gary does.

            CHAIRMAN BLANCO:  Okay.

            DR. O'SULLIVAN:  I would rather know what we're going to do before I answer this question.

            CHAIRMAN BLANCO:  All right, well, then, unless someone else feels strongly ‑‑ Jay?

            DR. IAMS:  I'll try to bail you out, Jorge.

            CHAIRMAN BLANCO:  Thanks.  I appreciate it.

            DR. IAMS:  I think the gold standard for tests is the Papanicolaou smear.  It ought to be possible to have post-marketing studies for obstetrical interventions or screening tests that do pretty much what the last 50 years have done for the Pap smear.  You ought to be able to see, and the Gothenburg study may be the beginning of such an event, that the introduction of a technology into practice results in appropriate screening and intervention that results in a real clinically-important outcome.

            So I think this is a device which pretty much automatically should qualify, if approved, whenever it is approved, for that sort of post-marketing surveillance.

            CHAIRMAN BLANCO:  Thank you.  Anyone else?  Go ahead.

            DR. WOLFSON:  I wanted to second what Jay said.  I think that, clearly, the experience that occurs in the United States has to be monitored and it has to be reported because, again, of the diversity that we are going to expect to see from country to country.  Without the European Union information, we don't really know even the initial diversity over there.  So I would suggest that there, indeed, be post-marketing studies.

            DR. SHARTS-HOPKO:  I have a question for Jay.  Would the outcome of a post-marketing study still be metabolic acidosis or would you like to carry that forward to be neurological outcomes?

            DR. IAMS:  Well, of course, you would like to see both.  I don't think it is the company's responsibility to follow every patient that uses their device forever and ever, but I think it would be their responsibility to participate actively in that sort of effort over time with whoever wants to follow that sort of thing.  That's a bigger issue than just for one individual company within an industry to do, but it ought to be the kind of thing we look forward, and that the company fosters and doesn't impede.

            We have had examples of that in other technologies where that doesn't happen.  Technologies have been approved, and then when they're not successful in influencing the ultimate outcome, there's a tremendous literature in obstetrics dating back 50-plus years about why it didn't exactly do the way it was supposed to do, but it is because the population got riskier, or all those cocaine babies, or whatever.  We've heard all that stuff before.  I don't think we should accept that anymore.

            CHAIRMAN BLANCO:  Okay.  Any other comments?

            (No response.)

            All right, that ends the section on the questions that the FDA posed to the panel.  What I would like to go through is I would like to go through with the different questions that were asked, and a lot of the questions that were asked of the company, the sponsor, have already been addressed, but maybe, Dr. Rosen, if you could come up and we can go through.  Please forgive me; I have "X'ed" some of these out.

            Yes, sir?  One more question?  You mean a new question of Dr. Rosen?

            DR. RINGEL:  Right.

            CHAIRMAN BLANCO:  Okay.  Why don't we go through what I have here.  Yes, sir, you can certainly ask a new one.

            Dr. Ramin, did you feel like most of your questions were addressed or did you have any questions that you wanted to bring forth at this point, from what we asked earlier on?

            DR. RAMIN:  I think they were addressed.

            CHAIRMAN BLANCO:  Okay, so you're satisfied that they have addressed the questions that you had.  All right.

            Dr. Ringel, you had asked about the decrease in metabolic acidosis that looked more in the control.  I think we had kind of addressed that already.  Were there any other questions?

            DR. RINGEL:  In the Gothenburg trial.

            CHAIRMAN BLANCO:  Yes, in the Gothenburg trial.  So you've got that done.

            Any of the others?  I thought the others were probably addressed.  All right.

            Let's see, Dr. Wolfson, you questioned, and Dr. Ramin also brought up, about, did they have any data on IUGR fetuses that were included in the study.

            DR. ROSEN:  Well, unfortunately, we don't have that many cases in the Swedish randomized trial.  As far as I remember, there were actually two or three cases that showed ST events as well.

            It is the issue of a fetus that is exposed to the stress of labor, regardless of race, would respond, according to what we know, in a similar fashion.

            CHAIRMAN BLANCO:  Okay.  So for IUGR you don't have that ‑‑

            DR. ROSEN:  We don't have the numbers required.

            CHAIRMAN BLANCO:  All right, the other one was, how as the management schema arrived at?

            DR. ROSEN:  This is very much part of the whole development process of the STAN device, starting with the experimental work, technological development, and then followed by observational studies in the clinical sense.

            Obviously, you're observing these cases of metabolic acidosis as they emerge and recognizing events that we previously saw in the animal setup, where we had structured settings with well-defined hypoxia.

            Then obviously there is the issue of the technology, to what extent we can detect significant events, given the source of information, the electrocardiogram.  In the Plymouth randomized trial we used sort of an old technology, analog signal processing.  There was a scatter of data presented, but we also looked for trends.

            Today the database in the ECG quality is so much different and so much improved that we can look for more detailed ST events emerging, thereby ‑‑

            CHAIRMAN BLANCO:  Actually, I think the question was really more addressing the issue that we may have brought up already in terms of, when you put down the way to use this device, you know, are you going to say, you look at the fetal heart rate monitoring and you only use it when the fetal heart rate monitoring is not reassuring.  Then you look at whether you've got ST events, and you add that component to make your decision.  That's what they ‑‑

            DR. ROSEN:  That is basically the way it should be used, according to the instructions.

            CHAIRMAN BLANCO:  Okay.  All right.  Dr. Neuman asked about lead configuration, T wave configuration, from spiral electrode, lead vector changes as labor progresses and fetus moves.  Have you looked at that at all?

            DR. ROSEN:  Oh, yes, very much so.  That's the reason why we have the unipolar electrode configuration, thereby creating an opportunity to have a consistent ST wave form displayed throughout labor, regardless of how the fetus rotates, but also identifying what's called the "Y" lead in the longitudinal plane of the fetus; that is where the T vector is represented.  So that is the way to secure that we will pick up T wave ST events throughout labor.  There is literature on this and papers published on this issue of lead configuration.

            So, as I said, to record the ST, it's completely different from just one heart rate being displayed.

            CHAIRMAN BLANCO:  All right.  Another question was brought up about, what were the complaints of the purchased units?

            DR. ROSEN:  These were related to failures of the printer, technical failures that were easily sorted.

            CHAIRMAN BLANCO:  Technical failures of the printer or of the machine?

            DR. ROSEN:  There was nothing on the safety aspect.

            CHAIRMAN BLANCO:  The concordance of the STAN results with pre-terminal tracings?

            DR. ROSEN:  Well, we have pre-terminal tracings as part of the guidelines.  As you noticed, there was one case that has been displayed here where they didn't follow the protocol.

            The physiology behind this is that, when a fetus is exposed, if you go beyond the point where the fetus is responding, and the heart rate performance or cardiovascular performance starts to decrease, you will not see a further rise in the ST.

            On the issue of when we use the technology and where there is a risk, could we identify cases prior to the onset of labor, if we look at the data presented to you, there are very few cases of an adverse  outcome, where they actually annotate complications, and, in particular, active pushing in second-stage labor, that's where things start to go wrong, when they do go wrong.

            The reason why we put forward the indications being a bit wide is to not miss out.  Clearly, there was the requirement for a scalp electrode, and that is not standard procedure.  With us, roughly 30, 40 percent of all term pregnancies in active labor obtain a scalp electrode.

            CHAIRMAN BLANCO:  Well, I think the panel has sort of addressed that, that they think that rather than reaching out, we need to tighten up.

            DR. IAMS:  Can I follow up on that?

            CHAIRMAN BLANCO:  Yes, sure.

            DR. IAMS:  I'm a little confused by that answer about the pre-terminal tracings.  I thought I was on the same page as you, in that by the time you get to a pre-terminal tracing that should have been recognized before, the introduction of the STAN technology at that point may not recognize it as abnormal because there is no further change?

            DR. ROSEN:  There will be no further changes.

            DR. IAMS:  Would that translate, if you back that process up a little bit, to whatever is right between pretty bad and pre-terminal?  Is there an opportunity for the delayed application of the STAN technology to fail to recognize, so if you wait too long it doesn't help you?  Is that the message that we're to take from this, is that we should ‑‑

            DR. ROSEN:  Well, if you look at the indications that we have been using in the Swedish randomized trial, then that's now used in labor ward settings where the STAN is utilized, that is that you don't start a STAN recording in active pushing second-stage labor.  That is not the case where you start a recording.

            You have to start prior, at least 30 minutes; you attempt to do that.  That is to recognize the fact that we need a period where we have a baseline.

            DR. IAMS:  Right.  Well, refresh me then in that case because I am not quite sure I got that out of the instructions.  Is that pretty clear in the instructions, that you do not ‑‑

            DR. ROSEN:  Yes.

            DR. IAMS:  I noticed that there were different criteria or that you had to be careful during pushing, but I didn't get the idea that you really shouldn't add this to someone's labor at the point where she's already pushing.

            DR. ROSEN:  That is what is told and what is very much part of the education process.

            CHAIRMAN BLANCO:  But is it just when you're pushing or is it when you already have a fetus that has had some myocardial decompensation?

            DR. ROSEN:  The reason we're doing this is that what we found in the study is that this period when you go from early stage to active pushing in second-stage, that is where we find that the fetus is exposed to hypoxia.

            CHAIRMAN BLANCO:  Okay.  I know, but there are going to be some patients, if this gets used in the United States, that will already have some level of hypoxia at the time that patients come in.  A lot of our patients come in late in labor.  They've got severe complications that create a problem, and the fetus may already be hypoxic.

            I think Dr. Iams' question is, if you already have hypoxia, myocardial deterioration, and you put the STAN monitor on, does it show you that that is occurring, or because there are already ST changes, the program will not recognize it?

            DR. ROSEN:  Well, this is obviously a somewhat difficult question to answer.  Depending on where you are, if you are in the pre-terminal phase, you have to rely on the heart rate trace, showing no variability, no reactivity, but if you've got reactivity, if you've got a fetal heart capable of responding, then you will also see ST events, regardless of what degree of metabolic acidosis you're at.

            CHAIRMAN BLANCO:  So the answer is really yes?

            DR. ROSEN:  Yes.

            CHAIRMAN BLANCO:  If you've already got significant myocardial effects, the STAN monitor will not show you anything?

            DR. ROSEN:  Well, it depends on where you are.  You may still have a heart which is exposed to hypoxia where you will find marked ST events associated with that.  But if you take it through the next step, where the cardiovascular system is failing, where the fetus has decreased blood pressure, then you will have the pre-terminal heart rate trace, and don't expect ST events to go with that.

            CHAIRMAN BLANCO:  Okay.  Jay, are you satisfied?

            DR. IAMS:  Yes, I think so.  Just let me say it back to you again.  If I have a patient with the appropriate risk factors, my good care of this patient would be maximized by having a STAN electrode, rather than a regular electrode, inserted by the time the time the patient reaches 6 or 7 centimeters, because if I don't do it then, I might end up failing to use the device in its most appropriate fashion?

            DR. ROSEN:  That would be according to the recommendations of use of the device.

            DR. IAMS:  Right.

            CHAIRMAN BLANCO:  Thank you.

            Did I get everybody's questions that were not answered?  Yes?

            DR. WOLFSON:  I just wanted to just clarify what I thought I heard you say, Dr. Rosen.  On the STAN criteria or the algorithm for intervention, the different levels of the .1, .15, and .05 are based on clinical review and interpretation of the large database that you have, and are not necessarily based specifically, for example, on receiver operating characteristic analyses per se?

            DR. ROSEN:  You are correct.

            DR. WOLFSON:  Is that correct?

            CHAIRMAN BLANCO:  All right, thank you.

            Anything else?  Go ahead, Gary.

            DR. EGLINTON:  Do you practice intrapartum fetal stimulation in Sweden, vibrio acoustic stimulation or LS clamp provocation of the fetus, looking for accelerations?

            DR. ROSEN:  Varying between different centers.  We do have the fetal blood sampling in Gothenburg, for instance, where in Lund there would be more vibrio acoustic stimulation, and so on.  So this varies depending on where you would look.

            DR. O'SULLIVAN:  Let me ask you a question regarding pushing, because this was something that I was impressed about.  In terms of at least in those tracings that you showed us where the baby developed or showed significant metabolic acidosis, while it wasn't present in all of them, it seemed that the tracings, at least as you looked at them, showed some degree of abnormality, which got worse with pushing.

            Was there any attempt to stop the pushing, to allow the baby to recover?  Or even to stop contractions to allow the baby to recover, before the baby developed some other abnormality?

            DR. ROSEN:  Well, once there's been a significant event, according to the clinical management protocol, obviously, we recommend immediate delivery under those circumstances.  According to the data we have, we want that baby out as soon as possible.

            To try to stop pushing, to try to give drugs, I mean if you obviously stop pushing and if you give tributalin, for instance, stopping contractions, okay, if contractions are the cause of hypoxia and active pushing, you would expect there to be less hypoxia, clearly.  But how to institute that, I'm not sure it would be practical and what we are recommending is simply take the baby out, please.

            CHAIRMAN BLANCO:  Okay.  Any other questions from the panel members?  Anything else they would like to bring up?

            (No response.)

            Okay, let's move on.  Open public hearing again.  At this point, what we would like to have is not questions or interaction, but just basically if anyone has any short comment from the audience, from the FDA, or the sponsor.  We will take them in that order.  Is there anyone in the audience not affiliated with the sponsor or with FDA that would like to make some final comments before the panel begins its deliberation and votes?

            (No response.)

            Okay, next would be from anyone from the FDA that would like to make any additional comments at this point.

            (No response.)

            Nope?  All right, then we go to the sponsor.  Anyone from the sponsor that would like to make any additional comments at this point?

            Okay, please identify yourself and your relationship.

            DR. NOREN:  My name is Hakan Noren.  I am a consultant obstetrician in the University Hospital, Gothenburg.  I have no economic contract with Neoventa, but they have paid my trip here and my stay here.

            CHAIRMAN BLANCO:  Thank you.

            DR. NOREN:  I would comment on the Gothenburg study.  You must realize this is a clinical situation.  This is not a randomized trial.  So don't look at them in a STAN group or a CTG group.  Look at them together to see what is the number of metabolic acidosis is.  We are now down to .48 percent from October to January.

            So it is two quite different groups:  the high-risk group, the STAN group, and the low-risk group.  This is dependent on the midwives an obstetricians getting better all the time to recognize that there is something wrong with the CTG, and they are immediately going over to the STAN group.  So you can't compare them.  Look at them as one unit.

            DR. RINGEL:  May I ask a question?

            CHAIRMAN BLANCO:  Please, go ahead.

            DR. RINGEL:  You presented, you, the sponsors, presented the data for us to interpret.  I understand what you're saying, but without knowing the historical controls for your bad outcomes, without knowing any comparison, without knowing what just education of your midwives and your obstetricians would produce, how am I to interpret the data?

            For all I know, it could just be the education process that has created such a tremendous statistical improvement in your outcomes.  For all I know, you may have a very low rate of bad outcomes in Gothenburg in general.

            So how am I to interpret the data regarding this machine?

            DR. NOREN:  You must realize that the knowledge in the CTG interpretation before the randomized trial wasn't that good, and a lot of midwives, and even doctors, we looked at CTG in different ways.  Doctor, this study, when we started with the STAN machine, we were forced to speak the same language.  This was the first thing.

            Secondly, the midwives were forced to identify what kind of CTG pattern it was in a much more detailed way than they had done before.  It has been an ongoing education all the time, especially after the relearning period, because we actually took those cases where something has happened and discussed them, and also pinpointed, say, to the doctor who had made violations against a protocol and tell them, "This is why there is a protocol."  That's why, obviously, this fetus has metabolic acidosis.

            So by doing this all week, all months, let's say, we have forced cooperation, to speak the same language, and also to be aware of the situation.  Of course, many times the clinical situation it sounds scientific here today, but many times you have a non-reassuring CTG, and we are more or less waiting for that ST event.  When that happens, everyone is ready to do the C-section.  Of course, the time between event and actually the C‑section or vacuum extraction, what we do is decreasing.  That is one of the reasons why we are better all the time, but it has continued to be better.

            CHAIRMAN BLANCO:  Okay, unless you have something else different, let's go ahead and bring it to a close, that you would like to say, okay?  All right, thank you.

            Okay, anything else the panel wants to bring up?  Questions or comments before we start?

            (No response.)

            All right, if not, Dr. Whang is going to address the issue of the recommendations, options, that we have available to us.  I will go over a little bit about how we're going to do the voting.

            DR. WHANG:  The panel will now be voting to determination its recommendation to the FDA.  Its recommendation can be one of the following:   approval, approvable with conditions, or not approvable.

            Approval means there are no attached conditions.

            Approvable with conditions means there are specific conditions which are discussed by the panel and listed by the panel Chair.  Conditions could be labeling changes such as a revised indication, contraindication, warnings or precautions, or a condition could be a post-approval study or a reanalysis of the data that have been presented.

            Not approvable can be recommended for one or more of the following reasons:  (A) There is a lack of showing of reasonable assurance that the device is safe under the conditions of use prescribed, recommended, or suggested in the proposed labeling.  (B) There is a lack of showing of reasonable assurance that the device is effective under the conditions of use prescribed, recommended, or suggested in the proposed labeling.  Or (C), based on a fair evaluation of all material facts, the proposed labeling is false or misleading.

            If the panel votes not approvable, the panel members will be asked to indicate why they think it is not approvable and what is needed to make it approvable; for example, additional clinical studies.

            CHAIRMAN BLANCO:  Thank you, Dr. Whang.

            So, basically, what we have before us is that we need to take a fairly complicated vote, and I'll try to make it simple.  Essentially, what we are going to do is we are going to see if anybody is interested in putting forth a motion and has a second that would approve the PMA without any conditions, just simply it's approved, done with, that's it.  I guess that will be the start.

            Does anyone have an interest in making a motion to approval without any condition at this point?

            (No response.)

            Okay, not hearing or seeing anyone who is interested in full approval, the next step is to see whether there are members who want to propose and second a motion that the PMA is not approvable.  Is there anyone who would like to make a motion, second, that the PMA is not approvable?

            DR. IAMS:  I'll make that motion.

            CHAIRMAN BLANCO:  Okay, there is a motion on the floor that the PMA is not approvable.  Is there a second?

            DR. EGLINTON:  Second.

            CHAIRMAN BLANCO:  There is a second on the floor.  Therefore, we are going to discuss this, and then take a vote.

            Jay, why do you think that it is not approvable?

            DR. IAMS:  Well, I would like to congratulate the authors on all their work.  I don't mean for this motion to be taken as a critical comment on all that they have done.

            I need to preface my remarks by ‑‑ I think I can speak for lots of people, myself perhaps specifically, in learning some lessons from the past about more data not necessarily leading to better outcomes.  My presence on this panel may have something to do with my involvement in studies of uterine contraction frequency, where more data did not lead has not led to better outcomes, and yet it was much more data.  That caused me to look back into history and see if other examples might exist, and they do, back to x-ray pelvimetry, for one, once used rather routinely, now used hardly at all.

            Electronic fetal monitoring as standardly practiced is probably the best example.  We have used it progressively and often, and we haven't accomplished, as Dr. Devoe pointed, the goals that we had set for ourselves.

            The second lesson of the past is that what works in one place may not work someplace else.  Baby aspirin in this country in large populations of people at risk didn't seem to do nearly as well as it did in other studies.

            So, with those concerns in mind, I'm concerned about transportability and I'm concerned that more data may not lead to better outcome.  I am hopeful that this will be an exception to both of those historical lessons, because, I confess, I was impressed with the information presented, but I don't think it is yet time to approve.

            I think that the reason for that would be Item B.  I'm not certain that it is going to be effective in the United States population.  I do not have concerns about its safety in general.  I think I'm more concerned about whether it would do here what it so clearly seems to have done in Sweden.

            CHAIRMAN BLANCO:  Let me put you on the spot again.  What would the sponsor have to do to satisfy your concerns and be able to achieve an approvable result?

            DR. IAMS:  Well, I think their EU study will, in part, perhaps address the concern about transportability from one place to another.  An American study would be another avenue to that.

            CHAIRMAN BLANCO:  Okay.

            DR. IAMS:  So that's it.

            CHAIRMAN BLANCO:  Does anyone want to speak to Dr. Iams' concerns or shall we move on to Dr. Eglinton and see his concerns?

            (No response.)

            All right, Gary?

            DR. EGLINTON:  I also want to congratulate Dr. Rosen and all of his co-workers for the years of work.  I learned a lot reading this.  I look forward to reading a lot more and learning much more in the coming years.

            I am very excited about this concept.  I am very excited about this technology, and I hope it works.  I hope that it improves where we are with fetal heart rate monitoring in this country.

            But I am not persuaded by what I have read that this will improve clinical practice in this country.  I hope that there will be a study performed in this country that will have very clear results and will convince me and others that this is a wonderful technology for us to use in this country.

            CHAIRMAN BLANCO:  And the same question to you, and I think I know the answer, but just so that it goes straight on the record:  What would the sponsor have to do to be able to satisfy your concerns, to make the PMA approval?

            DR. EGLINTON:  I think it is going to have to be done in this country.

            CHAIRMAN BLANCO:  All right.  Let me just, before we open the discussion to the other panel members, make note that the PMA review is independent of cost, previous regulatory difficulties, clinical data submitted in other PMAs, and the medical/legal climate and its effect on the standard of care.

            Having said that, does anyone else care to make any comments, either in favor or against the current motion and the comments about what has been said by Dr. Iams and Dr. Eglinton.

            DR. EGLINTON:  Can I just add one more phrase?

            CHAIRMAN BLANCO:  Sure.

            DR. EGLINTON:  I think it needs to be done in this country because of differences in the style of practice.  The way obstetrics is practiced in this country is dramatically different from the way obstetrics is practiced in Scandinavia, and probably many other places in the world.

            CHAIRMAN BLANCO:  Thank you.

            Any comments?  Michael?

            DR. NEUMAN:  Mr. Chairman, I'm just curious what the difference is between not approvable and suggested things to do and approvable with conditions, with the condition being a study in the United States.

            CHAIRMAN BLANCO:  Well, I would appreciate any help from the FDA, if I say it incorrectly, but if you're requesting another total study before it is approved, you're basically not approving it.  You're making the requirement that they have to have a total new study to come forward before the FDA and before the panel.

            As part of a condition, you can get post-market surveillance or another post-market approval study, but it will already have been on the market during that time.  So if it is approved with conditions, and the condition is that you do another study, then it is a post-market approval of study, and it is marketable at that point.

            Am I correct on that?  I would appreciate some help from members of the FDA.

            MS. BROGDON:  I think that's a semantics problem.  If you recommend that it's approvable, but one of the conditions is a whole new study be done, we would essentially take that as a not approvable recommendation because the sponsor would have to essentially start over with their clinical study.  If you mean a post-approval study, you should say that clearly.

            CHAIRMAN BLANCO:  That would be a condition then.  That's what I was trying to explain.

            Mr. Pollard?

            MR. POLLARD:  Right, and I would just like to -‑

            CHAIRMAN BLANCO:  Would you identify yourself, please?

            MR. POLLARD:  Thank you.  Colin Pollard with the Center for Devices and Radiological Health, FDA.

            I just wanted to remind the panel, from some comments that I made at the beginning of the day, just some basic definitions of safety and effectiveness.  With safety, basically, you ‑‑ FDA, actually, ultimately, and you in your recommendation ‑‑ have found that the probable benefits outweigh the risks of the device.  To find it effective, the studies that support that PMA have shown a clinically-significant result.  That is basically the threshold for an approval.

            Now for a post-approval study, for instance, you need to meet that threshold, and then that post-approval study would be looking at, you know, things beyond that threshold, either rare events or long-term results or fairly focused questions or validating a surrogate endpoint, that sort of thing.

            CHAIRMAN BLANCO:  Thank you.

            The floor is open for comments from the panel members.

            DR. O'SULLIVAN:  I really do believe that a study should be conducted in this country.  I think I say this because I think even the Scandinavians have shown that introducing it initially, followed by their interim analysis, re-education, and so on, shows that there is a great deal of learning that needs to go on in understanding how to use this and how to properly interpret it.  That's No. 1.

            No. 2, in contrast to Scandinavia and in contrast to what we would even say as far as product labeling is concerned, if this device is in a hospital, it will be like the pulse oximeter; they will use it routinely, and it won't matter.

            I think that the additional factor regarding its success would be that if we don't do this thing and were to otherwise approve it, that getting it off the ground and getting it working and getting everybody educated would take forever, would not be successful, not as successful, because of the thinking processes of physicians, nurses, and nurse midwives in this country.

            So if this was going to be at all successful in the United States, then I think that there's only one way to go, and that's a randomized controlled trial.

            CHAIRMAN BLANCO:  In the United States?

            DR. O'SULLIVAN:  In the United States.

            CHAIRMAN BLANCO:  Okay.  Let me just point out, in all fairness to the sponsor and the FDA, that the issue of non-approved use, unfortunately, is not the purview of the FDA.  They have to approve for the indications, but they don't and can't control how physicians in the United States then use the devices that are out on the market and available, just for the record, so everyone is aware of that.

            Okay?  Any other comments?  Dr. Wolfson, I think I see your hand.

            DR. WOLFSON:  A question for Dr. Pollard and then some thoughts.

            Colin, can you clarify again that what we are looking at when we are approving the PMA is that the investigations that have been accepted by FDA demonstrate statistical significance, and that the device is considered to not put increased risk of patients at jeopardy?

            CHAIRMAN BLANCO:  Let me just point out, we're not approving.  We're talking a motion for non-approval of the device.  Okay?  Just so we're clear.

            Mr. Pollard?

            MR. POLLARD:  Right, and your question is about to approve the PMA.  Basically, there's three elements that we talked about this morning.  One is that it is based on valid scientific evidence, and we talked about what that means, the different types of studies that any sponsor can submit.

            Based on valid scientific evidence, the PMA would need to show that it is both safe and effective, and sort of the operative words with safety is that the probable benefits outweigh the probable risks when it is used as it is intended to be used in the labeling that is given in the application.  Secondly, for effectiveness, that that sum total of the valid scientific evidence in the application shows when the device is used as it is intended to be used, that it produces a clinically-significant result.

            CHAIRMAN BLANCO:  Now let me just add, if you read it in your package, there should be the definitions, and there are issues of target population.  Maybe it would be worthwhile to read the definition of effectiveness for the panel.  The definition is:  "There is reasonable assurance that a device is effective when it can be determined, based upon valid scientific evidence, that in a significant portion of the target population, the use of the device for its intended uses and conditions of use, when accompanied by adequate direction for use and warnings against unsafe use, will provide clinically-significant results."

            I believe that what Dr. Iams and what Dr. Eglinton and Dr. O'Sullivan have been brought forth is concerns about the significance of it in the target population, and also the uses, the intended uses and conditions of use, by the physicians who would use it, if I could paraphrase that.  Is that all right, Dr. Iams?

            DR. IAMS:  That's fine with me.  I think, to say it differently, if I were a Swedish panel member in Sweden, I would easily vote yes.  I would vote it to be approved with some of the minor considerations we have talked about, but the issue is the target population.

            CHAIRMAN BLANCO:  All right, any other comments here?  Yes?

            MS. MOONEY:  Yes.  In light of the fact that we have two randomized controlled trials, acknowledging the fact that we have had that done outside the U.S., both showing significant improvement and meeting the study objectives, I wonder if perhaps the better course here would be to entertain more narrow indications in approvable status, where we target the highest-risk patients, if that would not tip the risk-benefit in a more favorable light, and then again come back to a post-approval status as far as determining whether any of our concerns with the U.S. population are valid; also, in light of the fact that the sponsor has indicated a willingness and commitment to put together an extensive training program, and at least in the Swedish situation has shown that that training commitment and executing that did, in fact, improve the use of the device.

            CHAIRMAN BLANCO:  Any comments?  Dr. Wolfson?

            DR. WOLFSON:  Jorge, yes, I'm really going to speak against not approving.  I guess that means I'm speaking toward approval, but I'm not sure, based on these definitions.

            I share my colleagues' concerns relative to effectiveness, but I would like to back up for the record and make a couple of points.

            I think it is clear to all of us in the obstetrical realm that one of our principal goals is to improve neurologic outcome.  For decades, we have looked at various tools to achieve that.  This is the first tool that has come along with demonstrable scientific evidence that it has improved neurologic outcome, period.  It's demonstrated in the UK.  It's been demonstrated in Scandinavia.  So I think the validity is there, and I think we don't want to lose track of that.

            Two, I would like to point out that we are looking at a paradigm shift, in my opinion, with respect to the use of fetal monitoring.  For one of the first times, we are now taking a signal, as crude as it is ‑‑ all of us understand that; we always are amazed that we can tell anything about a fetus by heart rate because you can't do that with anybody else, but the idea is we are now analyzing that signal, looking for information that has specific correlates with respect to outcome.  I believe the authors have done decades of work demonstrating the efficacy of this particular technique, and we're not to lose sight of the value of that.

            I think that an offshoot of this process that's been addressed by my colleagues is the fact that there is a great deal of variance or diversity in the interpretation of electronic fetal monitoring in the United States.  This is a concern.  It's an educational issue.  There's no doubt about that.

            I think clearly the same issue was true in the institutions that were part of the Swedish RCT, because we saw that from the beginning to the end, that as they received more training, their outcomes became better.

            I believe that a most important ‑‑ important, important ‑‑ offshoot of this particular tool, this particular product, is that it is going to take the American population of obstetrical providers, whatever be their ilk, and bring them back to restudying what is the electronic surveillance that we utilize for the intrapartum patient.  So that we will become better as a nation and as providers in interpreting the electronic fetal monitor, and with the proper interpretation of the ST segment, produce better outcomes and duplicate, in my opinion, the findings that were seen within the Swedish and U.K. studies.

            I have no evidence that this particular device creates harm when applied in an effective manner.  I would agree with my colleague who said that what would be preferable here, in my mind, is perhaps tightening up the indications of where it is utilized, rather than voting for non-approval, and therefore, obstructing access of obstetrical providers to this device for an undetermined period of time.

            We all know for them to mount a randomized controlled trial, even through the maternal-fetal network of the NIH, and bring it back to this institution is going to take perhaps anywhere from two to even as long as maybe four years.

            So I would share the concern of my colleagues with respect to effectiveness in an American population, and I am still, in my own mind, working through whether that requires non-approval or whether it requires approval with post-market study.

            Thank you.

            CHAIRMAN BLANCO:  Thank you for your very thoughtful statements, Dr. Wolfson.

            Anyone else who would care to address an issue?  Yes, Doctor?

            DR. BROWN:  I just have a point of clarification with something that Dr. Wolfson said.  He made the statement that the studies clearly showed an improvement in neurological outcome.  By that, do you mean they clearly showed a surrogate endpoint of neurological, improvement in the surrogate endpoint, meaning acidosis?

            DR. WOLFSON:  No, I was referring ‑‑

            DR. BROWN:  Because it wasn't my understanding that they were really powered to detect the neurologic outcome differences ‑‑ could you clarify that? ‑‑ on a statistical basis?

            DR. WOLFSON:  I was looking purely at the information that was provided in a Swedish randomized trial on neurologic outcomes.

            DR. BROWN:  But that was not one of the endpoints of the study or the objectives, right?

            DR. WOLFSON:  No, I know it's not one of the endpoints of the study.  I'm simply pointing out to everyone that, in my mind, it may be the most important aspect.  It may not have been specifically what they were after.  They were after reducing the risk of, or the likelihood of, metabolic acidosis, and they were after the idea of demonstrating that you could do it without increasing the Caesarean section rate, because that's been the principal criticism of all these new technologies we bring online, that it becomes nothing more than another indication to deliver somebody abdominally or operatively.

            What I said is that I guess, in my mind, the icing on the cake, the thing I've been grasping for for 25 years of practice is to have a tool that I can tell a patient that, because I am utilizing this, I have the opportunity to perhaps improve the neurologic outcome for your child.  This is part of the data that came out of that.

            So, yes, it is not one of the endpoints, but it is, to me, the most valuable aspect of it.

            DR. BROWN:  Well, let me be specific.  Was that ‑‑ and I'll ask the statistician ‑‑ was that a statistically-significant finding?

            DR. WOLFSON:  Yes, it was.

            MS. TOLEDANO:  The probability of having seven out of seven babies ‑‑ we had seven babies who had moderate or severe encephalopathy.  The probability of having all seven of those babies in one group is less than 2 percent.  So that is "P" less than .02.  That is a two-sided P value.

            DR. WOLFSON:  In both analyses, the intent-to-study as well as the revised, there was a statistically-significant difference.  It is in Volume 1, and God knows what page it is on because I didn't memorize it, but it was one of those things that, when I was going through the data and I was looking at focusing on all this stuff, and all of a sudden it hit me.  It's the detail on the Swedish RCT.  It's those tables that are printed out in landscape in your guide.

            DR. IAMS:  Page 15.

            DR. RINGEL:  Is it in 2, page 16?  Is that what you're referring to?  I'm not sure if that's what you were referring to or not.

            DR. WOLFSON:  I don't now.

            DR. IAMS:  I think it is the page before, page 15, but the first box, "intention-to-treat total."  That's eight versus one.

            DR. WOLFSON:  Yes, neonatal outcome.  The intention-to-treat total, moderate to severe, neonatal encephalopathy, it was 4 in the CTG group, 1 in the STAN group, statistically-significant, a .02 level.  Then when you took those neonatal outcomes and looked at them in ‑‑ what is interesting is ‑‑ where is it?  Well, anyway ‑‑

            DR. IAMS:  Go back a page to 15.  You're on 16.

            DR. WOLFSON:  No, I've got 15.  What I was looking for is that, when you then go into where you take out the adequate recordings, that moderate to severe neonatal encephalopathy becomes seven up against zero in the STAN group, giving a statistically significance of ‑‑

            CHAIRMAN BLANCO:  What page are you on?

            DR. WOLFSON:  Oh, I'm now on page 22.  It is the upper table.

            MS. TOLEDANO:  Yes, that's a one-sided P value.

            DR. WOLFSON:  It's the fifth line down.  Now you are at a P level of .007.

            MS. TOLEDANO:  That's a one-sided P value.  That's just for ‑‑ all right, you guys know what's a one-sided versus a two-sided.

            CHAIRMAN BLANCO:  Any other questions or anything else you want to elaborate on?

            DR. O'SULLIVAN:  I would just like to say I didn't follow him at all because I haven't figured out where he was.

            CHAIRMAN BLANCO:  Okay, well, let's make sure that you do.  Let's start off ‑‑

            DR. WOLFSON:  We're in Volume 2.

            CHAIRMAN BLANCO:  Volume 2.

            DR. WOLFSON:  Page 15.

            CHAIRMAN BLANCO:  Page 15.

            DR. WOLFSON:  Go to page 15, the upper part of the table where it says, "intention-to-treat total," within the box it's the fifth line down.

            DR. O'SULLIVAN:  Yes, I can see where it says, "Moderate to severe encephalopathy," there's a .02 P value.

            DR. WOLFSON:  Right.  Then on page 22, they now have taken out the ‑‑ page 22 includes the adequate recordings.  So they have removed the ‑ what is it? ‑‑ 300-plus cases, or whatever it is, where the recordings weren't adequate.  Then the similar analysis now is seven against zero.  So we have eliminated one.  So now it is at .007.

            DR. O'SULLIVAN:  Yes, but where is the long-term neurologic outcome?

            DR. WOLFSON:  That's it.

            DR. O'SULLIVAN:  It's short-term, though.

            DR. WOLFSON:  Oh, I'm sorry, no, there is no long-term because that wasn't the purpose.  I'm just looking at this from the standpoint that what happened in the nursery and the diagnosis that was made acutely, it disappeared.

            MS. TOLEDANO:  Somewhere there was long-term.

            DR. WOLFSON:  Was there long-term?

            MS. TOLEDANO:  Somewhere there was a table of long-term.

            DR. WOLFSON:  I don't remember the long-term.

            DR. O'SULLIVAN:  There was one thing where they did ‑‑ I think Dr. Wolfson would know this, but they did 18 months of followup or 16 months.

            CHAIRMAN BLANCO:  Please come.

            DR. O'SULLIVAN:  But I don't remember where it is.

            CHAIRMAN BLANCO:  Identify yourself.

            DR. KJELLMAR:  Ingemar Kjellmar, Professor of Pediatrics and Neonatology.

            I had included in the material that was distributed to you in my presentation on perinatal markers of quality in Sweden a few slides dealing with an old material of cases that I have resuscitated in the 1970s, and where we could provide real long-term follow-up studies.  They were about 25 years of age.

            In that material there is a table on the assessment at 18 months of age.  Out of the 74 resuscitated babies, we had 14 cases who became disabled.

            The point with the table that I have included, the points are actually two.  The first one is to demonstrate that there are recognizable patterns of cerebral palsy that dominate the outcome.  These are either dystonic cerebral palsy or tetraplegic or hemiplegic cerebral palsy.

            The second point is that those will be recognized at a very early stage, and already when the babies are 18 months of age, you recognize that they have poor motor performance and that they have a neurologic disability, although you are probably not able at 18 months of age to definitely tell which type of cerebral palsy this patient will have.

            CHAIRMAN BLANCO:  All right, let me interrupt you for a minute.  These 14, those were part of this study?

            DR. KJELLMAR:  No, no.

            CHAIRMAN BLANCO:  I wanted to make sure that was clear.

            DR. KJELLMAR:  Absolutely not.  They were all born in the 1970s.  So that is the reason why I did not ‑‑

            CHAIRMAN BLANCO:  So you don't have 18-month data for these babies.  You just have the acute data?

            MS. TOLEDANO:  That was my question actually.  Are we allowed to ask them directly or do I ask you?

            CHAIRMAN BLANCO:  Either way.

            MS. TOLEDANO:  I can ask him directly?

            Do you have any followup, 18-months' followup, one-year followup of the babies who were in your special care?

            DR. KJELLMAR:  Not yet.

            (Laughter.)

            CHAIRMAN BLANCO:  The answer, for the record, was "not yet" from Dr. Kjellmar.  Thank you.

            All right, any other comments concerning the motion on the flow?

            DR. RINGEL:  I would like to say, not in response to Dr. Wolfson, but after you mentioned these things, I went to look at it again.  I'm a technology junkie basically, but my problem here is that there are basically three studies we have been presented with.

            The Plymouth study does not show any difference in neurologic outcome and just barely shows a difference in the acidosis outcome, and that's why they felt they needed to go further and to use the automatic detection. Okay, so the Plymouth study doesn't really help us.

            The Gothenburg study, I know I keep harping on it, but as best as I can tell, the Gothenburg study does not tell us anything about the device itself, only about the possibilities of training and expertise and learning.  Maybe all this is about learning how to read the standard traces better.

            So we are left with one study where the data is a bit questionable because of the re-training that happened in the middle of the study, and because there are some questions about how significant the changes are.  For that reason, I pretty much agree with my colleagues who say that there needs to be another study done in the United States where there are comparable patients.

            I notice jitteriness is an outcome; irritability it an outcome.  There are some nurseries in the innercities where all the patients come out jittery and irritable.  We don't have to get into why, but I happen to know that is true because I have worked in the innercity my entire professional career.

            So I have to say that while, yes, we would love to believe that we can start preventing or decreasing the amount of cerebral palsy we see, the amount of learning disorders we see that may be related to asphyxiated deliveries, I'm not sure we have the proof yet.

            CHAIRMAN BLANCO:  All right, any other comments?  Dr. Toledano?

            MS. TOLEDANO:  I have to agree with Ms. Mooney and with Dr. Wolfson.  I would much rather see this going to a setting of approval with conditions.  If we are talking about being able to learn how to read tracings, these are the people and they have the materials to teach us how to read the tracings better.  I think we need to have access to that in this country in a more timely manner.

            CHAIRMAN BLANCO:  Thank you.  Any other comments?

            MS. MOONEY:  Just one final remark:  I think that, obviously, if we think we need a U.S. trial before the device can come here, then that's what needs to be done, but we should recognize that there is the potential that the cost associated with this trial would preclude the sponsor going forward with the study.  So this sort of decision could actually be a decision to not make the devise available at all.

            CHAIRMAN BLANCO:  Unfortunately, I must remind the panel that cost is not one of the things that we are able to consider here.

            All right, unless there is something different, we're going to go take a vote.  Do we have something different, Dr. Wolfson?

            DR. WOLFSON:  No, I was just going to address the type of trials.  Thank you.

            CHAIRMAN BLANCO:  Okay.  Well, I think we can do that depending on what the outcome of the vote is.  Unfortunately, there's no way to hide here.  So we're just going to go down the side.  Michael, Dr. Neuman, you've been here a long time, so I'm going to start on your side, so you'll be the first one.

            Now let's remember that the motion is for non-approval.  So that a yea vote on the motion is a motion to not approve the PMA.

            Your vote, Dr. Neuman?

            DR. NEUMAN:  Yes, I vote not to approve the motion.

            DR. RINGEL:  Yea.

            CHAIRMAN BLANCO:  Dr. Ringel?

            DR. RINGEL:  Yes.

            CHAIRMAN BLANCO:  Not to approve?

            DR. RINGEL:  Not to approve.

            CHAIRMAN BLANCO:  Dr. Toledano?

            DR. IAMS:  No, no, no.  He got that wrong.

            CHAIRMAN BLANCO:  He said "yea."

            DR. RINGEL:  I said "yea" three times.

            CHAIRMAN BLANCO:  Yes, "yea," not to approve the motion, correct?

            DR. RINGEL:  Yes.

            CHAIRMAN BLANCO:  Not to approve ‑‑

            DR. RINGEL:  I'm the one that messed it up.  I'm sorry.  Not to approve the PMA.  I apologize.  I'm the one for non-approval.

            CHAIRMAN BLANCO:  Yea or no?

            MS. TOLEDANO:  No, I do not approve the motion.

            CHAIRMAN BLANCO:  Okay, so you vote for some other ‑‑

            MS. TOLEDANO:  That is correct.

            CHAIRMAN BLANCO:  We would like confirmation of Dr. Neuman's [sic] vote.  You voted "yea" for the motion to not approve the PMA?

            DR. RINGEL:  No.  I voted not to approve the motion to not approve ‑‑

            (Laughter.)

            CHAIRMAN BLANCO:  We've got too many noes there.

            DR. BROWN:  It might be better if we said for or against the motion, for or against it.

            CHAIRMAN BLANCO:  Why don't we just say do not approve of the PMA.  Let's just say, not approve of the PMA ‑‑

            DR. SHARTS-HOPKO:  No, no, you can't do that.

            CHAIRMAN BLANCO:  -- to make sure that we don't get into any difference, and let's start all over again.

            DR. O'SULLIVAN:  That violates parliamentary procedure.

            DR. BROWN:  For or against the motion.

            MS. TOLEDANO:  The motion is to not approve.  Are you for the motion or against?

            CHAIRMAN BLANCO:  All right.  Dr. Neuman, are you for the motion or against?

            DR. NEUMAN:  I am against the motion to not approve.

            CHAIRMAN BLANCO:  Thank you.

            DR. RINGEL:  For.

            MS. TOLEDANO:  Against.

            CHAIRMAN BLANCO:  Against.

            DR. WOLFSON:  Against.

            DR. SHARTS-HOPKO:  Against.

            DR. IAMS:  For.

            DR. RAMIN:  For.

            DR. EGLINTON:  For the motion.

            DR. O'SULLIVAN:  For the motion.

            DR. BROWN:  For the motion.

            CHAIRMAN BLANCO:  Wait, wait, not so fast.

            (Laughter.)

            Dr. O'Sullivan, for the motion.

            Dr. Brown?

            DR. BROWN:  For the motion.

            CHAIRMAN BLANCO:  Dr. Seifer?

            DR. SEIFER:  Against.

            CHAIRMAN BLANCO:  Against.

            The motion carries 6 to 5.  The motion is approved.  The PMA is not approved.

            We now need to go around and explain why we voted in a brief way.  I think we have already said most of the things that needed to be said in terms of what the sponsor will need to do to be able to bring forth an approval PMA.

            Dr. Neuman?

            DR. NEUMAN:  While I agree with the concerns about the different patient population in North America compared to Sweden and the U.K., I feel there is compelling evidence that looking at the configuration of the electrocardiogram is providing additional information that we currently are not using, and that the studies that have been done up until now certainly have concerns.  Nevertheless, I think that, provided that additional post-marketing studies are done in the states to make sure that these concerns of the panel are not as significant as some members feel they are, would be sufficient.

            CHAIRMAN BLANCO:  Thank you.

            DR. RINGEL:  I don't think there's reasonable assurance that the device is effective.

            CHAIRMAN BLANCO:  Thank you.

            MS. TOLEDANO:  I agree with Dr. Neuman.

            CHAIRMAN BLANCO:  Thank you.

            DR. WOLFSON:  I think I have expressed my concerns.

            CHAIRMAN BLANCO:  Thank you.

            DR. SHARTS-HOPKO:  I agree with Dr. Neuman, and I didn't want to close the door against continued development of the product.

            CHAIRMAN BLANCO:  Thank you.

            DR. IAMS:  I think I've already stated my rationale.

            CHAIRMAN BLANCO:  Thank you.

            DR. RAMIN:  I, too, have concerns about the effectiveness or efficacy, and I think just a randomized clinical trial in the U.S. population is what is needed.

            CHAIRMAN BLANCO:  Dr. Eglinton?

            DR. EGLINTON:  I have the same concerns.  I think we'll see clinical use in Europe and it will spread, and it will spread around the world, and maybe someday in the future it will spread far enough that we'll be happy with it here.  But I think at this point it is still a research tool, the way we are looking at it.  I think it's an exciting research tool, but I think it's still a research tool at this point.

            CHAIRMAN BLANCO:  Dr. O'Sullivan?

            Thank you, Dr. Eglinton.

            DR. O'SULLIVAN:  There's really nothing I can add to that.

            CHAIRMAN BLANCO:  Thank you.

            DR. BROWN:  I specifically have concerns that I did not believe there was enough proof of clinical significance, although there was statistical significance of the endpoints.

            CHAIRMAN BLANCO:  Thank you.

            DR. SEIFER:  I agreed with Dr. Neuman's points.

            CHAIRMAN BLANCO:  Thank you.

            I would like to thank all the panel members.

            I would like to see if the FDA, Dr. Pollard, do you have anything to say?

            MR. POLLARD:  Yes, I just wanted to get a little bit of clarification.  I mean, I did hear the panel ask for a randomized clinical trial to be conducted in the U.S.  It might be helpful to have a little better sense of just what kind of trial are you talking about.  You know, specifically, what kinds of endpoints would a trial like that require?

            DR. IAMS:  I think a trial very much similar to the one they have proposed, powered to look at metabolic acidosis, powered, if possible ‑‑ you have to look at the numbers to look at markers, clinical markers, of newborn encephalopathy, seizures and some of the markers you've already used in your current trial.  That may not be possible.  Metabolic acidosis is fine.  I have no problem with that.

            CHAIRMAN BLANCO:  Dr. Iams and Dr. Eglinton and Dr. O'Sullivan, and the others, anyone else, can you specify a little further?  I mean, it wasn't the issue of the endpoints.  So that this helps the sponsor and the FDA.  I don't think it was the issue of the endpoints as much as the issue of the population in terms of the patient population and the issue of the management in terms of what occurs in this country as opposed to what might occur in other countries.

            Am I ‑‑

            DR. IAMS:  That's correct.

            CHAIRMAN BLANCO:  -- paraphrasing it for you all appropriately?

            DR. O'SULLIVAN:  The only thing that I think would be of benefit, although, again, I think it would be very expensive, you can't certainly do a two-year followup.  So from that perspective, I think cost-effective-wise that would be ridiculous.  But there are now currently emerging imaging studies that one might think about tagging on here, at least before the baby is discharged or at some point after discharge, but, ideally, during the hospital course of the baby because that's when the baby is a captive.  I think those could be looked into, and certainly I'm not expert in those.

            CHAIRMAN BLANCO:  Any other points?

            DR. WOLFSON:  Can I make one?

            CHAIRMAN BLANCO:  Yes, sir.

            DR. WOLFSON:  I would encourage that the studies be done not solely in academic centers.  Most centers tend to be urban, low-income, minority, and the majority of births that are affected ‑‑ well, I should say they just simply take place within the United States are clearly occurring in what are principally Caucasian, non-urban environments.  So it is important from that standpoint, though, granted, if you demonstrated in an urban environment, then you'll demonstrate it anywhere.

            CHAIRMAN BLANCO:  Thank you.

            If there are no other comments, I would like to thank the sponsors for their presentation, their patience, and their information.

            I would like to thank the FDA for their presentations and their assistance.

            I would like to thank each and all of the panel members for their very erudite comments and their involvement.

            And almost all the members who have been to other hotels, I would like to thank the FDA for having it at this hotel as opposed to the one that we shall not mention, but you're well aware of that none of us like.

            (Laughter.)

            Thank you all very much.  This meeting is adjourned.

            (Whereupon, 3:58 p.m., the foregoing matter was concluded.)