34
UNITED STATES OF AMERICA
+ + + + +
DEPARTMENT OF HEALTH AND HUMAN SERVICES
+ + + + +
FOOD AND DRUG ADMINISTRATION
+ + + + +
CLINICAL CHEMISTRY AND CLINICAL
TOXICOLOGY DEVICES PANEL
+ + + + +
MEETING
+ + + + +
THURSDAY,
OCTOBER 28, 1999
+ + + + +
The Committee met in Room 020B,
9200 Corporate Boulevard, Rockville, Maryland, at
9:12 a.m., Henry C. Nipper, Ph.D., Panel Chair, presiding.
PRESENT:
HENRY C. NIPPER, Ph.D., Chairman
BEVERLY HARRINGTON FALLS, M.D., Member
SHERWOOD C. LEWIS, Ph.D., Member
BARBARA R. MANNO, Ph.D., Member
NADER RIFAI, Ph.D., Member
ARLAN L. ROSENBLOOM, M.D., Member
DAVIDA F. KRUGER, M.S.N., Consumer
Representative
ALTON D. FLOYD, Ph.D., Industry
Representative
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
35
PRESENT (cont'd):
STEPHEN CLEMENT, M.D., Consultant
BASIL T. DOUMAS, Ph.D., Consultant
JAMES EVERETT, M.D., Ph.D., Consultant
JANINE E. JANOSKY, Ph.D., Consultant
ROBERT REJ, Ph.D., Consultant
MARY M. KIMBERLY, Ph.D., Consultant
VERONICA J. CALVIN, M.A., Executive
Secretary
STEVEN I. GUTMAN, M.D., M.B.A., Division
Director
ALSO PRESENT:
ARLEEN PINKOS
ANN HAWTHORNE
DAVID G. BROWN, Ph.D.
JIM CONNOLLY
SUNIL ANAOKAR, Ph.D.
MARK DEEG, M.D.
JOHN PASQUA, M.D.
MARGO ENRIGHT
TELBA IRONY, Ph.D.
HENRY GINSBERG, M.D.
CAROL BENSON
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
36
A-G-E-N-D-A
I. Closed Committee Discussion .................. 4
II. Call to Order
-- Opening Remarks .......................... 37
-- Introductions ............................ 40
-- Conflict of Interest Statement ........... 42
III. Y2K
-- David G. Brown, Ph.D. .................... 44
IV. Open Public Hearing
V. Sponsor Presentation - Polymer
Technology Systems, Inc.
-- Jim Connolly ............................. 55
-- Sunil Anaokar, Ph.D. ..................... 94
-- Mark Deeg, M.D. ......................... 119
VI. Question and Answer Period ................. 120
VII. FDA Presentation
-- Arleen Pinkos, MT(ASCP) ................. 178
-- Carol Benson, MA MT(ASCP) ............... 193
-- Telba Irony, Ph.D. ...................... 196
VIII. Guest Speaker
-- Henry N. Ginsberg, M.D. ................. 204
IX. Question and Answer Period ................. 225
X. Open Public Hearing ........................ 280
XI. Open Committee Discussion .................. 280
XII. Final Recommendations ...................... 340
XIII. Closing Remarks ............................ 340
XIV. Adjourn
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
37
1 P-R-O-C-E-E-D-I-N-G-S
2 (9:12 a.m.)
3 CHAIRMAN NIPPER: The panel and the audience
4 should come to order.
5 At the beginning of the panel meeting, I must
6 respectfully ask that the table right behind the projector
7 be cleared until we're ready to have you occupy that area.
8 Are you able to do that with minimum disruption to what's
9 going on there? Okay. Thank you.
10 I'd like to call on Veronica Calvin, the
11 Executive Secretary of the Clinical Chemistry and
12 Toxicology -- Clinical Toxicology Devices Panel -- for
13 opening remarks, introductions, and a conflict of interest
14 statement.
15 MS. CALVIN: Good morning and welcome to the
16 meeting of the Clinical Chemistry and Clinical Toxicology
17 Devices Panel. Today the committee will discuss and make
18 recommendations on a pre-market notification for an
19 over-the-counter device that measures triglycerides from
20 whole blood finger sticks. Before we move into today's
21 agenda, I will provide brief summary minutes of the last
22 panel meeting.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
38
1 The Clinical Chemistry and Clinical
2 Toxicology Devices Panel last met on February 26, 1999,
3 to discuss a pre-market approval application for the
4 continuous glucose monitoring system presented by
5 Minimed, Incorporated. The panel unanimously
6 recommended approval of the PMA subject to the following
7 conditions: submission of additional data regarding
8 interference, validation of the calibration algorithm,
9 and use in some patient groups not previously selected
10 for the study, and labeling changes. On June 15, 1999,
11 the device was granted full approval to the market.
12 At this time, I would like to acknowledge
13 special guests who will be participating in the meeting
14 today. Dr. Henry Ginsberg, our guest speaker, is the
15 Director of the Irving Center for Clinical Research at
16 the College of Physicians and Surgeons of Columbia
17 University. He is also head of the Division of Preventive
18 Medicine and Nutrition in the Department of Medicine at
19 Columbia, and an expert for the National Cholesterol
20 Education Program, to name a few of his highlights.
21 Dr. Mary Kimberly, from our sister agency,
22 the Centers for Disease Control and Prevention, she
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
39
1 coordinates the Cholesterol Reference Method Laboratory
2 Network, which you will hear more about later.
3 Lastly, Dr. Alton Floyd, industry rep for the
4 Hematology and Pathology Devices Panel, is substituting
5 for Dr. Robert Habig, who could not be here today.
6 Also, I bring you regrets from Dr. Martin
7 Crowell, who could not be here. He called late yesterday
8 to inform me that a matter arose at his hospital requiring
9 his immediate attention and presence today.
10 Now I would like for the panel members to
11 introduce themselves, beginning with Dr. Robert Rej.
12 DR. REJ: I'm Robert Rej, Director of
13 Clinical Chemistry and Hematology at the New York State
14 Department of Health. I'm a former member of this panel,
15 and I'm a temporary voting member to this panel today.
16 DR. EVERETT: I'm James Everett. I'm
17 Medical Director of Madison Memorial Health Care in
18 Madison, Florida.
19 DR. MANNO: I'm Barbara Manno. I'm
20 Professor of Psychiatry at Louisiana State University
21 Health Sciences Center in Shreveport, Louisiana. And I'm
22 a toxicologist and I'm a voting member of the panel.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
40
1 DR. DOUMAS: Basil Doumas, Professor
2 Emeritus, Medical College of Wisconsin, in pathology.
3 DR. JANOSKY: Janine Janosky from the
4 University of Pittsburgh in the School of Medicine. I'm
5 a biostatistician. I'm a voting member of the Dental
6 Products Panel.
7 DR. LEWIS: I'm Sherwood Lewis. I'm the
8 Director of Toxicology in the Office of the Chief Medical
9 Examiner, the State of Connecticut. I'm a voting member
10 of the panel.
11 MS. KRUGER: I'm Davida Kruger. I'm a
12 certified nurse practitioner from Henry Foote Health
13 Systems in Detroit in the area of diabetes, and I am the
14 consumer representative on this panel. Thank you.
15 DR. FLOYD: Alton Floyd. I'm the industry
16 representative for the panel today sitting in. And I have
17 my own consulting company, Trigon Technology.
18 DR. GUTMAN: I'm Steve Gutman, and I'm the
19 Director of the Division of Clinical Laboratory Devices.
20 DR. ROSENBLOOM: I'm Arlan Rosenbloom,
21 Professor Emeritus of pediatrics, pediatric
22 endocrinologist at the University of Florida, and Director
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
41
1 of the Children's Medical Services Center.
2 DR. CLEMENT: Steve Clement here in D.C.,
3 Associate Professor, Georgetown University, and Director
4 of the Georgetown Diabetes Center.
5 DR. KIMBERLY: Mary Kimberly from the Centers
6 for Disease Control and Prevention. I'm Coordinator of
7 the Cholesterol Reference Method Lab Network, and I'm here
8 as a guest today.
9 DR. RIFAI: I'm Nader Rifai. I'm Associate
10 Professor at Harvard Medical School and the Director of
11 Clinical Chemistry at Children's Hospital.
12 CHAIRMAN NIPPER: And I'm Henry Nipper. I'm
13 Dean of Admissions at Crane University School of Medicine,
14 Associate Professor of Pathology and Associate Director
15 of Clinical Chemistry and Toxicology at St. Joseph
16 Hospital in Omaha. And I'm Chair of the panel, except
17 Veronica runs things.
18 (Laughter.)
19 We all know that.
20 MS. CALVIN: Thank you. I will now read the
21 conflict of interest statement.
22 The following announcement addresses
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
42
1 conflict of interest issues associated with this meeting
2 and is made part of the record to preclude even the
3 appearance of an impropriety. To determine if any
4 conflict existed, the agency reviewed the submitted agenda
5 and all financial interests reported by the committee
6 participants.
7 The conflict of interest statutes prohibit
8 special government employees from participating in
9 matters that could affect their or their employees'
10 financial interests. However, the agency has determined
11 that participation of certain members and consultants,
12 the need for whose services outweigh the potential
13 conflict of interest involved, is in the best interest
14 of the government.
15 We would like to note for the record that the
16 agency took into consideration matters concerning Drs.
17 Martin Kroll, Nader Rifai, Arlan Rosenbloom, Basil Doumas,
18 and Ms. Davida Kruger. These panelists reported current
19 and/or past interest in firms at issue, but not in matters
20 related to what is being discussed today.
21 Since these matters are not related to the
22 specific issues of this meeting, the agency has determined
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
43
1 that they may participate fully in today's deliberations.
2
3 The agency would also like to note for the
4 record that Dr. Henry Ginsberg, who is the guest speaker
5 for today, has acknowledged previous interest in firms
6 at issue.
7 In the event that the discussions involve any
8 other products or firms not already on the agenda for which
9 an FDA participant has a financial interest, the
10 participant should excuse him or herself from such
11 involvement, and the exclusion will be noted for the
12 record.
13 With respect to all other participants, we
14 ask, in the interest of fairness, that all persons making
15 statements or presentations disclose any current or
16 previous financial involvement with any firm whose
17 products they may wish to comment upon.
18 Thank you. I'll turn the meeting back over
19 to Dr. Nipper.
20 CHAIRMAN NIPPER: Thank you, Ms. Calvin.
21 Before we begin the open public hearing, we're
22 going to hear about Y2K from Dr. David Brown from the Office
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
44
1 of Science and Technology. And I believe that this handout
2 is from you, and a nice folder.
3 DR. BROWN: Correct. I'm happy to be able
4 to provide that. We have some extra copies over there
5 if you need one.
6 I'd like to give a very brief presentation
7 about our activities in the area of Y2K. I think that
8 you are fortunate in that I guess you haven't had a meeting
9 in some time perhaps because we are almost at the Y2K before
10 you are subjected to the presentation. And I think it's
11 just about too late probably for anything to be done, but
12 at least you can be acquainted with some of the --
13 (Laughter.)
14 -- things which we have been doing.
15 I think we have been concerned about this for
16 some time. We know that there has been a lot of alarm
17 raised. And, in fact, I think one of the major purposes
18 behind our work in this area is that, dating back, say,
19 to 1996, we had quite a few dire predictions made. And
20 so we wanted to see, is there a problem? What do we need
21 to do? And perhaps we are able to, in some way, reassure
22 the public if there are not going to be these dire
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
45
1 consequences.
2 Of course, our efforts will probably be more
3 than overwhelmed by what I am informed is an NBC movie
4 as a Y2K catastrophe movie, which will be shown shortly.
5 However, we will still go ahead.
6 Indeed, we are told that most of our PCs will
7 be unreliable, told that all of our health care systems
8 will be failing to work, our medical devices will be
9 non-functional. I think everyone is aware of the basic
10 problem, of course, that back when some of us were
11 programmers there was a severe shortage of space, really.
12
13 I think that's the basic problem. When you
14 are doing with 80-column cards to input data, you just
15 used two of those columns for the year, and certainly a
16 computer program could be confused between 1900 and 2000,
17 if they just ran across 00.
18 And for those of us who have done more
19 programming than that, we are aware that computer programs
20 can become more confused than just that particular
21 ambiguity, by encountering 00 if their programmer was not
22 allowing for that eventuality.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
46
1 Well, where can this come up? Certainly,
2 there are microprocessors, PC-controlled products, more
3 and more all the time, that we are concerned not only with
4 some of the direct medical devices but also the laboratory
5 equipment, because that certainly applies to the products
6 that are reviewed by this panel in particular.
7 But there are many software applications,
8 device interfaces to databases and recordkeeping systems,
9 and this problem of embedded chips, which even the
10 manufacturers of devices are not thoroughly cognizant of,
11 but are embedded in the devices fairly deeply, and then
12 are used for dates and displays and recording.
13 Well, certainly, there is a reason to consider
14 whether or not some of these things could have really bad
15 consequences when they are misprogrammed. And our
16 definition of year 2000 compliance has tended to be pretty
17 broad.
18 We're just not saying, is there going to be
19 a problem when the clocks ticks over on midnight on December
20 31st; but the question about 9/9/99, was that going to
21 be a major problem; the question about leap year in the
22 year 2000, since there will be a leap year and normally
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
47
1 in those years divisible by four there is not one, has
2 that been taken care of.
3 So we want systems to be year 2000 compliant.
4 On the other hand, we want to see what is the magnitude
5 of any hazards which may arise. Certainly, as far as this
6 panel and all of our other panels, and elsewhere in the
7 medical community, we are always interested in knowing
8 what insights you may have into the problem, as your
9 facilities have checked out your Y2K readiness.
10 Well, I think the primary message of this
11 presentation is the primary thing which we have
12 accomplished -- and I think it's a major
13 accomplishment -- that we do have an FDA product database.
14 This is accessible through the regular FDA site,
15 www.fda.gov. That's the easy thing to remember, and from
16 there you can, of course, link to the year 2000 area of
17 the web site and find out a lot about the status of a lot
18 of medical devices, because I think our primary
19 accomplishment has been to work with manufacturers to find
20 out -- to have them post information about their devices.
21 Well, let me step back a minute. We have
22 concerns of three kinds. First, how about the Center's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
48
1 systems? Certainly, we use a lot of computers inside CDRH,
2 and we had to make sure that our own systems would be Y2K
3 compliant. And we are assured by our computer staff that,
4 indeed, that is the case. We've gone to great lengths
5 to try to ensure that that is the case.
6 The second item is, what about the medical
7 devices which we regulate? Our first activity was to send
8 out notifications to manufacturers so that we could be
9 sure that they really were cognizant of the potential for
10 that problem, and of our interest in their taking steps
11 to address it.
12 The second was to do something very concrete,
13 to put up this database of biomedical devices, where
14 manufacturers have submitted information which we have
15 put up, or they have submitted their own URLs which link
16 to our database, so that people can see what potential
17 problems there are with their medical devices. And I think
18 that has been a worthwhile effort as clinical facilities
19 have used that quite a bit to examine the Y2K compliance
20 of the devices which they use.
21 And the good news in that effort was we have
22 found that although there are a fair number of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
49
1 non-compliant devices, most of the problems are being
2 addressed, and most of the problems are of a fairly minor
3 nature.
4 First, we'll note a couple of pages from the
5 web site, so that you can be familiar with it. We are
6 talking about biomedical equipment, not just medical
7 devices, and also laboratory equipment and other types
8 of equipment. And just note that there are various search
9 capabilities to try and make this database user-friendly.
10 You really have to get up on that database yourself to
11 check that out more thoroughly.
12 What are the kinds of answers which we have
13 discovered? Well, we have about 693 manufacturers who
14 have reported that they do have non-compliant products,
15 345 of these specifically on our database and 348 of these
16 providing their own links to their own web sites.
17 So, indeed, there are problems. We are
18 trying to get manufacturers to address these, and most
19 of these are being addressed. The good news, however,
20 was that most of these have been found to have a very limited
21 impact, mainly the question of -- in fact, mainly the
22 question -- a matter of printing out incorrect dates, for
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
50
1 example.
2 But some of these -- for example, with
3 radiation therapy equipment and the use of
4 radionuclides -- have to do with miscalculations which
5 could have had serious consequences.
6 The kinds of activities which we have been
7 involved in -- letters to manufacturers, guidances to
8 manufacturers, establishment of the database which I have
9 just mentioned on our website -- but we have also gone
10 beyond that to try to monitor and assess the performance
11 of the manufacturers in addressing these problems.
12 In fact, the latest thing we have been
13 involved in is a major contract with Battelle to go out
14 and look at the -- what we refer to as potentially high-risk
15 devices, to really actually go to manufacturers'
16 facilities and check to make sure that their statements
17 about their -- the way in which they have addressed these
18 questions are, indeed, correct. And, indeed, we have
19 found, with an extremely small number of exceptions, that
20 that is the case.
21 In just about conclusion, I would only draw
22 your attention to the fact that further questions about
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
51
1 Y2K could be addressed to the panel Executive Secretary,
2 or to the person who really is in charge within our Center
3 of all Y2K matters, Tom Shope. And here is the contact
4 information for Dr. Shope.
5 The only thing I will append to this is that
6 FDA and, of course, CDRH in feeding into the FDA effort
7 will be operating an emergency operations center as the
8 date rolls over into the next millennium. And we have
9 established a procedure to supplement, which, of course,
10 that procedure which we've had all along to -- when we
11 find out that there are problems with medical devices.
12
13 So that if we do find out that problems are
14 arising, and the year rolls over, we will be able to address
15 those. And this is an addition part of that larger effort
16 being led by John Tuskimen, who is to set up an emergency
17 response center in downtown D.C. for the interest of all
18 of the various sectors of the nation's economy, for
19 obviously no matter how good a job we do with medical
20 devices, if there isn't any electric power or water, then,
21 of course, we'd be in very serious problem. But I think
22 that those things are being addressed.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
52
1 So that is the conclusion of this
2 presentation. I'd be happy to answer any questions. And,
3 again, there are some handouts that are available in the
4 corner of the room.
5 CHAIRMAN NIPPER: Thank you, Dr. Brown.
6 Does anyone have questions?
7 DR. RIFAI: What kind of information are
8 companies providing the FDA for the FDA? Is it just a
9 letter indicating that they believe they are now Y2K
10 compliant, or they are providing evidence indicating that
11 they are compliant?
12 DR. BROWN: The information that they are
13 providing is, one, that their products don't involve a
14 date; two, that they are Y2K compliant; or, three, if they
15 have a problem product, then the nature of that problem,
16 what they are trying to do about it.
17 Now, you are correct -- this is, as with almost
18 everything we do, a self-certification by the
19 manufacturer. That's why we have gone out with these
20 potentially high-risk devices to actually send inspection
21 teams into manufacturers' facilities to do a survey which
22 is a spot check, but a survey to make sure that there is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
53
1 a basis for what they are telling us, and that they have,
2 indeed, carried out the tests which they have said they
3 have done.
4 Okay? Well, thank you.
5 CHAIRMAN NIPPER: Thank you. No other
6 questions?
7 Well, thank you very much, Dr. Brown, for your
8 presentation.
9 At this time, we are a little bit ahead of
10 schedule, but I would like to move ahead so that we have
11 as much time as we need for the panel to deliberate, ask
12 questions, and so forth.
13 So at this point, I would like to open the
14 meeting for an open public hearing. We are not aware of
15 anyone who has requested time to address the panel and
16 present information relevant to the agenda, but interested
17 persons may so state at this point.
18 If there is a speaker, the speaker is asked
19 to state whether or not they have any financial involvement
20 with the manufacturer of the product being discussed or
21 with their competitors. Seeing no one who wishes to
22 address the panel, I think at this time if the sponsor
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
54
1 is ready we should move ahead with the sponsor
2 presentation.
3 Polymer Technology Systems has indicated that
4 three individuals will present to the panel -- Mr. Jim
5 Connolly, the President; Dr. Sunil Anaokar; and Dr. Mark
6 Deeg.
7 And, Mr. Connolly, the floor is yours to
8 address the panel or to ask your associates to participate.
9 We're going to move that overhead projector,
10 if you will just wait for a second. We're just going to
11 set it down so we can see across the room.
12 MR. CONNOLLY: Thank you. The fact that we
13 are a technology-driven company is proof -- now we have
14 another problem with some technology, which we will fix
15 here in just a moment, like the rest of them.
16 The first dry chemistries that I'm aware of
17 were about 300 B.C., done by obviously a Greek person of
18 some kind. So the things that we're talking about today
19 have come a long way since then.
20 I wanted to give you some of our background
21 and some of our thinking as we repair our presentation
22 here. The first dry chemistries I saw were in the clinical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
55
1 laboratory -- not in the laboratory but in an emergency
2 room I think about 1973 or '74, and working in a clinical
3 chemistry department where samples are handled and
4 processes are in place, and people are trained versus going
5 down to an emergency room one day and seeing a slender
6 piece of material with some blood applied to it going into
7 a small instrument and giving an answer that no one
8 believes.
9 Certainly, as a laboratorian, I thought it
10 was interesting, but certainly not the real deal, as we
11 called it. That was a few years ago.
12 I think it's interesting now that we're on
13 the other side of the table trying to sell these pieces
14 of paper as the real deal, and the changes that have
15 occurred over the past 25 years, and particularly the last
16 five years.
17 I'd read from my notes, but if you know John
18 and Margo and I, we all have something in common -- it's
19 our handwriting, which is a struggle. That's why we need
20 these slides as quick as we can.
21 Now, we founded this company in 1992, as a
22 result of a meeting with Charlie Suther, which I believe
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
56
1 one of our panel members knows here, a fellow that has
2 been active in diabetes back in the Ames Company before
3 it was Miles/Bayer, later on Boehringer Mannheim, and then
4 to Medisense, along with a fellow by the name of Dr. Tony
5 Gatto.
6 And I was at dinner one night and someone
7 decided what the world needed was a device that people
8 could use at home to manage the complications of diabetes.
9 And we started talking about diabetes and cholesterol
10 issues, which were the thing of the day in the late '80s,
11 early '80s. And that was truly the genesis of the company.
12 So the company's package is just as it says.
13 It's a multi-test menu. It used to be called an MTM
14 device. It is no longer. It is the BioScanner 2000.
15 But it is a small device that does many tests.
16 And the goal, as I stated earlier, was to bring
17 back the quality or to make sure the quality was in the
18 product that clinical laboratories are used to using but
19 in a package that could go to the consumer and also go
20 through clear waiver things to put it in places where people
21 didn't have access to rapid testing; therefore, quick
22 response and care of patients.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
57
1 So when you see the device -- I'm not sure
2 if there's one here today. We've got some pictures of
3 one. When you see the device, it'll look like a
4 glucometer. I hope I'm not abusing their trademark. I
5 didn't see them as guests here. But it is not a glucose
6 meter. Let me do it like that.
7 This is a true five-wavelength, two-angle
8 spectrophotometer. It just happens to be small. We made
9 it small so it could be used in personal use, carried with
10 you. It's battery-powered. The multiple wavelengths are
11 there for a number of reasons, to make development of the
12 chemistries, the broad spectrum of tests that we wanted,
13 easier for the chemist. In other words, they're not locked
14 into a particular wavelength or measurement algorithm.
15 And the company is focused on diabetes disease
16 management, to provide tools for the diabetic to manage
17 his disease, and share information with his caregiver,
18 and also a tool for the caregiver to use. But the focus
19 is not on the daily glucose measurement because most
20 diabetics don't die from high glucoses at the moment;
21 rather, the complications.
22 So, thus, the emphasis on the lipids and the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
58
1 one we speak about today, triglyceride. But it's using
2 these all together. We think there is controversy about
3 triglycerides, as well as a number of other things. But
4 for sure, when used with other tests, they definitely
5 provide a utility for CHD. When used independently with
6 women at higher risk, and particularly when used with
7 diabetics, they make quite a contribution.
8 So rather than focus on the controversy and
9 go to all of the papers that depended -- as an independent
10 risk factor -- some people much more capable than I will
11 talk about that -- we'd rather focus on the relationship
12 of triglycerides and the other lipids to diabetes for -- so
13 I'd like to call it a half-full glass of water.
14 Just a quick background. There's 200,000
15 that die from diabetes. You can go right through these.
16 Eighty percent of them die from cardiovascular disease,
17 half of Type 2s are discovered after their first heart
18 attack, number one cause of blindness in the U.S., number
19 one cause of non-traumatic amputations, number one cause
20 of kidney failure, third of the dialysis patients in the
21 U.S. are diabetics, 25 percent of the people that die from
22 heart attacks are diabetics. So the relationship is very
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
59
1 clear here.
2 This is the package or the menu that we talked
3 about. We began with a test to make the regulatory process
4 as straightforward as possible, beginning with a test that
5 had a lot of background glucose, getting into a little
6 bit more controversial things, and then as we go down the
7 line even more controversial. But you can see all of the
8 tests definitely have a relationship to diabetes.
9 Obviously, the lipids, ketones, outside the
10 lipid arena, but again used in areas where diabetics are
11 sick, and the gold standard for measuring compliance A1C.
12 We, by luck or by gosh or by circumstance, ended up that
13 the panel with triglycerides. I think it could have just
14 as well been a different test, one of these other tests
15 that ended here, and there are more to come after this.
16 Micro albumin is not on here, but, again, it's more tests
17 in that same arena.
18 This is where they're at approval-wise.
19 There are six approvals. We are waiting for some other
20 approvals, and today, obviously, there are some questions
21 about triglycerides, and these submissions are just around
22 the corner for us.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
60
1 So you can see where the company is going,
2 and I think it's pretty clear why we're headed in that
3 direction -- to provide a utility to the patient and the
4 caregiver that does not exist today, with the same kind
5 of capability information they'd receive from the
6 hospital.
7 There's a lot of crossover between this thing.
8 When we first got started, it was strictly diabetics,
9 and there has been a lot of interest in our products to
10 be used in areas outside diabetes because of the lipid
11 things and cardiovascular disease. And we are hoping to
12 address that market through the consumer approval of our
13 products, which leads to an easier CLIA path, which gets
14 us into the physician's office and into some of the
15 screening areas.
16 So we have 30 percent of the population. Most
17 of the experts here know these facts to be so. A lot of
18 them at risk for CAD, and a bunch of these people are
19 insulin-resistant or early Type 2 diabetics.
20 I'm just going to go over this briefly because
21 there are people here to do a better job of this than I.
22 But it's early onset, and it's around for a long time
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
61
1 before it's detected. The impacts are there for quite
2 a while. I believe about half of this, half of the Type
3 2s, are discovered after their first heart attack.
4 And there is this dyslipidemic thing, which
5 seems to be common, which has been published by guys like
6 Reaven, Grundy, Hafner, and so forth, with the elevated
7 triglycerides and the decreased HDL. Again, used in
8 combination, markers for CAD, as well as some possible
9 indications or -- what is that word? It's not diagnostic.
10 About Type 2 diabetes.
11 Next slide?
12 We took a couple of quotes here, which someone
13 else will explain. The part I like about this, that some
14 of these standards -- this published work has a big thing
15 here and what -- a big impact on what we're trying to do
16 here. And that is that these tools sometimes are not
17 there, or maybe the educational process is not there.
18 But a lot of people that are dyslipidemic in
19 this area were not aware of this, and very few of them
20 are receiving treatment. So this is one of the tools that
21 we hope to provide -- a knowledge device.
22 Same thing. Elevated triglycerides and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
62
1 decreased HDL issues, and coronary heart disease -- well
2 published, well documented. We have a handout here today
3 with several references talking about this and the
4 authors. We'll make those available as soon as I finish
5 here.
6 And the conclusion -- optimal care should be
7 taken -- taking care of these people that are dyslipidemic,
8 particularly those on insulin.
9 And my favorite, especially for today, this
10 study, which we all know very well, most of us know very
11 well, that are in this area anyway, that triglycerides
12 levels were a better predictor of outcome than cholesterol
13 levels. And there is a BioScanner 2000, which used to
14 be the MTM, about the size of a package of cigarettes.
15 There is a bunch of them. We can just go right
16 through these. Half of them, the Type 2 is diagnosed;
17 the other half we believe are undiagnosed -- cost a bunch
18 of money.
19 There's one more.
20 So any impact on this quality of life,
21 reduction of cost, more responsive, more effective, kind
22 of things we think would make a huge impact in diabetes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
63
1 My history with dry chemistry began when I
2 was in school with early test type things from Lilly, later
3 on the Ames Miles/Bayer product referred to. Boehringer
4 got in with a product, B.G., if any of you remember it -- a
5 major milestone in diabetes -- and Lifescan in 1985 with
6 a much easier to use device.
7 And, to me, a company that came out with the
8 proof that you can really do quantitative measurements
9 accurately, not using necessarily paper chemistry but
10 certainly some of the sophisticated membranes that came
11 into being in the mid '80s; and now I think a culmination
12 of that, of using the materials, putting them together,
13 and coming up with a system that does perform to the
14 standards that we'll show you today.
15 Does anyone know who Nicholas Culpeper was?
16 I didn't. I thought he was a guy down the hall that did
17 most of the copy repair. He translated, in his day, a
18 lot of the medical information that was in different
19 languages into English to be used by the English.
20 I think -- I certainly don't draw any
21 parallels between our accomplishments and his in bringing
22 medicine to everyone, but it certainly was a big change
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
64
1 to take some of the mystery out of medicine, or the
2 relationship between physicians and patients, and make
3 it something where the patients are more involved now.
4 And I think patients are getting more involved for a lot
5 of reasons.
6 One, health care and access to it has
7 certainly changed, and the 'net has certainly made a lot
8 of information available to people that wasn't there
9 earlier. So I think this is one of the things we'd like
10 to draw an analogy to.
11 There is a big change in patient access to
12 care. Products like ours are going to provide more
13 information, more utility, for the patient and for the
14 physician. And I've never had hair that long.
15 That's it.
16 We've got a short three-minute presentation
17 to show you what the product is.
18 (Whereupon, a portion of the video was shown.)
19 That micro albumin is the one that's used in
20 England. But I think you get the picture here. It's a
21 single instrument with different test strips that are
22 inserted into it with a small blood sample, a very small
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
65
1 blood sample, some as small as five microliters.
2 Results are usually in less than a minute.
3 Later on, some tests will be combined together in logical
4 panels like cholesterol, HDL, a direct LDL. But, again,
5 it's all focused on the diabetic with a single instrument,
6 with a lot of capabilities.
7 And I think there's one more part to this.
8 (Whereupon, the remainder of the video was
9 shown.)
10 I was just checking on the A/V guy to see if
11 he had it right.
12 The rest of this information is going to be
13 presented -- an overview -- mine was the overview, the
14 introduction of the company. And I'm not sure how the
15 data has been presented. There are four people here to
16 present the data and the nature of diabetes, and I'm not
17 sure who comes up next.
18 Dr. Mark Deeg is an endocrinologist at the
19 University of Indiana, and he is going to help us out here.
20 DR. DEEG: Thank you, Jim.
21 As he said, my name is Mark Deeg. I'm an
22 endocrinologist at Indiana University. I conduct basic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
66
1 research on HDL metabolism, and clinically I'm the
2 Director of the Cardiology Clinic at what I call the Indiana
3 Vascular Disease Center, otherwise known as the Roudebush
4 VA, which has -- veterans have a lot of cardiac disease.
5 I serve as a consultant for TPTS, in terms
6 of the clinical utility of the various tests, and they
7 have asked me today to talk a little bit about what it
8 really means when you measure someone's triglycerides.
9 Personally, I'm actually delighted that you
10 invited Dr. Ginsberg here. He is one of the -- as Ms.
11 Calvin indicated, he's one of the foremost experts in the
12 country with respect to lipids and what they mean.
13 You can leave the lights up.
14 I'm going to throw around a few numbers, and
15 let me just sort of introduce a little bit about
16 triglycerides and what some of these numbers might mean.
17 Let me just give you just a little bit of background.
18 First of all, in the last national population
19 survey, the median triglycerides for women was about 88
20 milligrams per deciliter. This just gives you an idea
21 of where the country stands. For men, it's about 112.
22 Okay.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
67
1 In 1993, the NCEP classified triglycerides
2 as such. They said less than 200 was considered normal,
3 200 to 400 was borderline, greater than 400 to 1,000 was
4 high, and greater than 1,000 was considered very high.
5 Sort of an interesting scale.
6 Now, triglycerides actually are quite common.
7 So this is 88 for women, 112 for men, less than 200 -- it's
8 200, 400 -- 400 to 1,000, and greater than 1,000. Now,
9 triglycerides actually are a fairly common problem in the
10 United States. On average, about five to 10 percent of
11 the population fit into this category of Americans.
12 Now, if you live in Indiana, like I do, which
13 is one of the heaviest states in the country, there is
14 actually more like 10 to 20 percent of the Indiana
15 population fit into this category.
16 The major focus is triglycerides, and some
17 of the controversy, is their role in cardiovascular
18 disease, in predicting cardiovascular disease. The one
19 complication that I'm not going to talk about, which is
20 actually quite important, that people in this
21 category -- high-risk for what's called pancreatitis.
22 And it's a potentially big problem in our diabetics, but
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
68
1 we're not going to talk about this class of people.
2 Lights, please.
3 So what I want to talk about for the next 20
4 minutes or so is, what does it mean to me as a practicing
5 physician when I measure someone's triglycerides? And
6 what does it mean to the patients, and what do I tell them?
7 Next slide.
8 What I want to talk about -- I've outlined
9 here -- is these five things, and some of these things
10 I'm sure Dr. Ginsberg will touch on. We'll talk a little
11 bit about the ugly details of triglyceride metabolism.
12
13 And the reason I throw that up there is because
14 if you really want to understand what happens when you
15 really measure triglycerides, you need to understand what
16 this is or how -- where triglycerides come from and where
17 they go to, discuss the controversial issue about
18 triglycerides as an independent risk factor for coronary
19 artery disease, triglycerides as a synergistic risk
20 factor, triglycerides as a metabolic marker for other
21 syndromes.
22 Now we're getting a little bit of clinical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
69
1 trials about what happens when we treat people for their
2 triglycerides. Basically, fat flows around in our blood
3 as balls of grease. This is just a cut view of a ball
4 of grease that consists of various fats, which include
5 cholesterols and triglycerides and various proteins.
6 There's a whole bunch of these different balls
7 of grease with different names, the largest being
8 chylomicrons, VLDL, very low dense type of proteins, IDL,
9 intermediate dense, the bad guys -- that is, LDL
10 cholesterol -- and the good guys, HDL cholesterol, which
11 we refer to as happy healthy.
12 Next slide.
13 If you had bacon and eggs for breakfast, this
14 is what's happening. You consume that dietary fat. It's
15 absorbed into the intestine and forms what's called
16 chylomicrons. These chylomicrons are the largest
17 particles that are very triglyceride rich. Okay?
18 These chylomicrons are then broken down
19 through an enzymatic action to remnants, which are then
20 broken up into the liver. The liver can then reprocess
21 this fat and these triglycerides into another particle
22 that I call VLDL.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
70
1 These also are triglyceride rich. These are
2 also broken down further, and by the same mechanisms, into
3 remnants, some of which are called IDL. And there are
4 a number of particles in between this. This is sort of
5 simplified. It's even more complicated than this.
6 Now, one of the important points about
7 triglycerides and what makes it so difficult as
8 triglycerides as risk factors is that it's metabolically
9 linked with all of these other lipoproteins. So, for
10 example, as these VLDL particles are broken down into IDL,
11 some of the constituents end up in HDL. Hence, this
12 interrelationship between triglycerides and HDL. We'll
13 talk about that some more.
14 These remnants can be further broken down into
15 LDL -- again, the bad guy in terms of the, you know, very
16 well-established risk factor for coronary artery disease.
17 The LDL can either go back to the liver or deliver its
18 cholesterol elsewhere. Other remnants can be taken back
19 into the liver.
20 So to borrow a phrase actually from Dr.
21 Ginsberg in one of his editorials a couple of years back,
22 is when you ask the question, what does triglycerides mean,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
71
1 are triglycerides a risk factor, his answer was, it's a
2 simple question but a very complicated answer.
3 But the point I want to make in this
4 presentation is that even though scientifically it's
5 complicated, it's still very clinically useful for the
6 physician as well as for the patient.
7 Whoa.
8 (Laughter.)
9 Let's talk about this issue. For a long time,
10 triglycerides as an independent risk factor of coronary
11 artery disease has been controversial. And let me just
12 visit that issue a little bit. And why has that been
13 difficult?
14 Well, part of it has to do with that the daily
15 variation of measuring someone's fasting triglycerides
16 is actually quite high compared to other particles in the
17 blood. For example, your daily variation in triglycerides
18 can be upward of 20 percent; whereas, for LDL cholesterol,
19 it can anywhere from five to 10 percent.
20 As I alluded to, when you measure
21 triglycerides, there's a lot of different particles in
22 blood that have triglycerides, and you're not really sure
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
72
1 which one you're measuring. And what I didn't emphasize
2 is that those are remnants I talked about -- those broken
3 down particles -- at least scientifically appear to be
4 atherogenic. And that's actually what we may be measuring
5 when we're measuring triglycerides, looking at those
6 things.
7 The issue of an independent risk factor as
8 very -- a synergistic factor we'll talk a little bit about.
9
10 And, finally, sort of the bread and butter
11 of the clinical evidence based medicine just hasn't been
12 there for triglycerides. We're beginning to see that now,
13 and we'll see further trials in the next few years. But
14 it's not here yet.
15 Next?
16 Some more recent data addressing the issue
17 of triglycerides as an independent risk factor are now
18 really coming forth that, indeed, triglycerides are an
19 independent risk factor for men. And that has been the
20 most difficult group to show this.
21 This is a particular study called the
22 Copenhagen Male Study, with about 3,000 men, who measured
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
73
1 their triglycerides and a whole bunch of other things,
2 including body mass. And what was unique about this study,
3 they actually took into account dietary alcohol.
4 And when you adjust the triglyceride levels
5 for risk, for all of the various things that can influence
6 triglycerides, what they found, as you went from the lowest
7 group to the highest group of triglycerides -- and the
8 cut points here was 100 milligrams per deciliter, 140
9 milligrams per deciliter, that as your triglycerides
10 increased, your risk for a cardiac event went up.
11 This is just one example of a study. This
12 is a recently published meta-analysis that looked at a
13 whole bunch of population studies, which took into account
14 some 23,000 men. These are the individual studies. This
15 is the summary data -- relative risk for coronary artery
16 disease. This is the risk for each increase of 90
17 milligrams per deciliter in triglycerides.
18 So you can see some of the earlier studies
19 where you can see that -- why it was so controversial.
20 It wasn't much above one. Being above one, obviously,
21 is an increased risk. Some of the later studies began
22 to prove this, and now with the meta-analysis, which
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
74
1 basically you combine all of this data. And this data
2 is adjusted for HDL. That, indeed, for men it's a
3 small -- it's still a risk factor, albeit small.
4 Keep going.
5 However, and I said, this is for men,
6 middle-aged men. Unfortunately, I'm now in that category,
7 having turned 40 this spring.
8 It's really been controversial for men.
9 However, there are other subgroups where triglycerides,
10 as an independent risk factor, is actually much stronger
11 and much more evident. This is illustrated from the
12 Framingham data. This is, again, looking at the relative
13 risk for having a coronary event in men in the orange here
14 at various triglyceride levels, so going from 50, which
15 is actually close to what I would call normal -- and I'm
16 not going to get into that.
17 But if you had to pick your triglyceride
18 level, you'd want it to be down here -- going up to 400.
19 And you can see there's a small increase, but, again,
20 not very impressive as a risk factor.
21 However, for women, you can see that in this
22 same range that it's a much better predictor for coronary
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
75
1 artery disease. So in post-menopausal women, in this
2 subgroup, triglycerides -- and the studies are all
3 confirming this -- that, indeed, for women, it is a good
4 predictor of coronary artery disease. And, again, this
5 is from the same meta-analysis, and a couple of
6 studies -- this is about 6,500 women -- that for women
7 it was a much better risk factor than it was for men.
8 Next slide.
9 Well, what about triglycerides as a
10 synergistic risk factor? What about people who have
11 combined dyslipidemia, which is a very common problem,
12 combining it with both high triglycerides and a high LDL?
13 What does this mean for their risk? Let's look at a couple
14 of studies.
15 This is a study called the PROCAM
16 study -- about 4- or 5,000 men, middle-aged men, followed
17 for eight years, looking at the event rate for coronary
18 artery disease. When they looked at different levels of
19 LDL, going from less than 130 up to 190, triglycerides
20 are low -- let's say less than 200 -- your risk goes up
21 if you have low triglycerides. And, again, this is not
22 unexpected, that if your LDL is higher you have a higher
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
76
1 event rate.
2 However, if you also have high triglycerides,
3 look at your risk factors or the event rates. At any given
4 LDL level, you have at least a twofold or even higher
5 increased risk for coronary artery disease, if you have
6 both high triglycerides and high LDL.
7 Now, what's interesting -- that this
8 particular group, which had the highest event rate, was
9 only five percent of the study population, yet it accounted
10 for 25 percent of the number of events in the whole study.
11 So this particular combination is particularly bad.
12 Next slide.
13 Again, this is from the past perspective study
14 that Jim alluded to. Again, they're using triglycerides
15 now less than 123 or greater than 123. And, again, these
16 are sort of numbers that would be considered normal, at
17 least by the old criteria.
18 And here they're looking at cholesterol
19 rather than LDL, but the same effect -- that if you have
20 higher triglyceride levels, that for any -- for your given
21 cholesterol level you will have a higher risk, a multiplier
22 effect if you will.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
77
1 Next slide.
2 Now, this is some interesting data that
3 actually was just down the street at Hopkins by a guy named
4 Mike Miller. And they looked at people -- they took about
5 500 men and women who went to the coronary cath lab to
6 have -- so they could look at their coronary arteries and
7 see if they had coronary artery disease, and then followed
8 them for up to 18 years to see who survived and who didn't
9 survive, and based on their triglycerides of being less
10 than 100 or greater than 100.
11 And, again, this cut point was from the data
12 that I had told you, that this is really the median for
13 the U.S. population -- about 100 for men and women. And
14 that what they found is that people who had triglycerides
15 at the beginning of the study had less than -- triglycerides
16 less than 100 lived longer than those who had triglycerides
17 that were higher.
18 So the point being from this study is that
19 what we consider high or low may -- one, may vary in
20 different populations, but these absolute numbers I think
21 we may need to consider ratcheting down in terms of what
22 we consider high.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
78
1 Well, as I alluded to earlier, triglyceride
2 metabolism is quite complicated, and triglycerides is a
3 risk factor, maybe in addition a marker for the various
4 alterations and metabolism of triglycerides and lipids.
5 This is becoming more and more appreciated as we
6 understand the complexities of triglyceride and lipid
7 metabolism.
8 When you have high triglycerides, the
9 interplay with all of these other lipoproteins is quite
10 complicated. And what happens is is that in association
11 with these high triglycerides there is an increase in these
12 remnants, the chylomicron remnants, which are
13 atherogenic, at least certainly in vitro and animal
14 models; IDL, which, again, another remnant; VLDL remnants.
15 All of these are triglyceride rich. This is what you'll
16 be measuring when you measure triglycerides.
17 Also, because of the metabolic link with HDL,
18 you have low HDL levels. And also, you have a form of
19 LDL, the bad cholesterol, called small events, which is
20 felt to be particularly nasty or an atherogenic form of
21 LDL. These are all associated with hypertriglyceridemia.
22 Now, the thing is, what do I mean by
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
79
1 hypertriglyceridemia? Well, it turns out that there have
2 been numerous studies that have shown that certainly when
3 your triglycerides hit 150, or start going higher, you
4 begin to see these metabolic changes.
5 So when I teach my residents about
6 triglycerides and the clinical chemistry thing says less
7 than 200 being normal, I tell them about these studies
8 that, indeed, certainly in this scenario, when you see
9 low HDL and high triglycerides, there is actually abnormal
10 lipid metabolism and an atherogenic lipid metabolism.
11 This has been called the atherogenic
12 lipoprotein profile. These things are metabolically
13 intertwined. An important point is that up to about a
14 third of the U.S. population has this metabolic disorder.
15 This is a very common disorder. This is something we're
16 beginning to appreciate more and more in trying to address
17 how to treat this problem.
18 Now, the other marker for high triglycerides,
19 as Mr. Connolly was alluding to, was that triglycerides
20 in the insulin-resistant syndrome are intertwined. The
21 insulin-resistant syndrome again is very common in the
22 United States, and this syndrome is associated with many
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
80
1 abnormalities that increase the patient for coronary
2 artery disease.
3 These include obesity, hypertension, going
4 on for Type 2 diabetes. There is changes in the
5 coagulation state, that people tend to be hypercoagulable,
6 which is part of the events that occur when you have an
7 acute coronary or heart attack.
8 And there's also endothelial dysfunction
9 associated with that -- endothelial cells being the ones
10 that line the artery wall. And these cells -- it's very
11 important that these cells are happy and functioning
12 properly. In the insulin-resistant state, they do not
13 work properly.
14 Again, also associated with the
15 insulin-resistant syndrome, and very early in this
16 diagnosis, are the high TGs, low HDL syndrome -- again,
17 TGs being over 150 and HDLs being more than 35 for men
18 and 45 for women.
19 Next slide.
20 This is a particularly increasing problem for
21 the United States in terms of coronary artery disease,
22 and the reason being is we know people with Type 2 diabetes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
81
1 have an increased -- two- or three-fold increased risk
2 for coronary artery disease compared to people who have
3 normal glucose tolerance -- diabetics, glucose tolerance.
4 For people with just insulin resistance, not
5 hypoglycemic, not diabetic, they, too, have this same two-
6 to four-fold increase in risk of coronary artery disease.
7
8 This is now finally being labeled a disease,
9 and the reason this is such a big -- and Dr. Clement will
10 confirm this -- that this insulin-resistant diabetes is
11 going to be a huge problem for our country in that currently
12 there are about 16 million Type 2 diabetics, just as many
13 insulin-resistant patients, and this is going to be a whole
14 lot worse over the next 20 or 30 years for a number of
15 reasons.
16 One is, as a population, we're getting
17 heavier, which is associated with the insulin resistance.
18 Another reason is we're getting older. And that
19 conversion -- becoming insulin-resistant and becoming
20 Type 2 diabetes increases as you get older. So now as
21 the boomers are turning 50, there are 10,000 people turning
22 50 a year now, and when I turn 50 in 10 --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
82
1 DR. ROSENBLOOM: 10,000 a year?
2 DR. DEEG: Well, 10,000 a day, and then I
3 represent the peak. In 10 years, it will be closer to
4 50,000 a day will be turning 50. This will become a huge,
5 huge, huge problem.
6 Next slide.
7 As I mentioned, the combined dyslipidemia,
8 which is a particularly nasty lipid profile to have in
9 terms of risk, is very common in diabetics. This is some
10 data from the NHANES population, looking at triglycerides
11 in diabetics that about half -- a little more than half
12 have triglycerides less than 200, a third have 200 to 400,
13 and 10 percent have greater than 400.
14 Now, the thing I want to, again, reemphasize
15 is that diabetics are at increased risk for pancreatitis,
16 particularly if they have diabetes along with a genetic
17 disorder. Again, these are triglycerides in the 1,000
18 range.
19 Next slide.
20 Jim alluded to this -- that diabetes is a very
21 nasty risk factor for coronary artery disease, that if
22 you have diabetes you are more than likely going to die
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
83
1 from an atherosclerotic event -- heart attack or stroke.
2 These account for most complications. And
3 of concern to me as a practitioner is that at the time
4 of diagnosis -- and this is one of the arguments for early
5 diagnosis as well as, you know, looking for people with
6 insulin resistance -- as many as half the patients at the
7 time of diagnosis already have coronary artery disease.
8 We missed the boat. The horses are out of the barn.
9 What are some of the clinical trials that can
10 help us decide if treating triglycerides are important
11 or not? There has not been a trial directly aimed at
12 answering that question. Part of the reason is that it's
13 just difficult to do because of the inner metabolic
14 connections between those lipoproteins.
15 Part of it is that, again, because of the
16 controversy about triglycerides as a risk factor. But
17 there is some data that suggests that treating these people
18 gives them clinical benefit. And let me show you data
19 from a couple of trials.
20 Next slide.
21 This is from the Helsinki heart study. This
22 was a study done in the late '80s using a product,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
84
1 Gemfibrozil, primarily as an agent to lower LDL, which
2 it doesn't do very well. But it does lower triglycerides
3 very well.
4 And what they did is, again, they looked at
5 people who had an LDL-to-HDL ratio of either less than
6 five or greater than five. Okay? So what that means is
7 that these people tend to have either really high LDL levels
8 or low HDL levels. And then when they break it further
9 down where they had low triglycerides or high
10 triglycerides, depending upon -- the cut point was 200.
11 So in the placebo group here in the sort of
12 green, you can see that the event rate in these groups
13 were pretty similar. But, again, this group that had the
14 high triglycerides and had this high ratio vis-a-vis high
15 LDL with the low HDL, they had a much higher event rate.
16 Okay? Three- or four-fold greater than these other
17 people.
18 But when you treated them with Gemfibrozil,
19 you can see there's a tremendous drop. And as a matter
20 of fact, this group accounted for 80 percent of the trial
21 results.
22 And what Gemfibrozil does, which I didn't show
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
85
1 you, is lower triglycerides very well. Okay? And, again,
2 this is a group that has elevated triglycerides, again
3 using a more stringent cutoff if you will.
4 This is a recently published study called the
5 VA HIT trial, which I'm really quite excited about because
6 of the people that they identify. It was about 2,500
7 patients, again treated with or without Gemfibrozil.
8 Again, what we're focusing in on was people with low HDLs.
9 HDL, 32; triglycerides only 161; and, interestingly, an
10 LDL of 111. And this is by far the lowest LDL in any major
11 trial that we've done to date. Okay?
12 This is a very typical lipid profile for
13 people with insulin resistance and Type 2 diabetes. These
14 are people with coronary artery disease. And they treated
15 them with Gemfibrozil in order to raise their HDL and lower
16 their triglycerides.
17 And what they found pharmacologically is what
18 you would expect with this drug -- that actually LDL didn't
19 change and even went up a little bit, HDL went up -- its
20 final numbers were six percent. This is the preliminary
21 data. Triglycerides went down 25 percent; again, a
22 response you would expect, and, again, going 25 percent
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
86
1 from 160.
2 You had the death rate, the stroke rate -- I'll
3 call it mortality -- decreased some 25 percent. So this
4 would suggest that, again, it's a little bit complicated
5 because of the inner mix, but, again this is some data
6 suggesting that treating triglycerides, maybe in
7 conjunction with raising HDL because of the intertwining,
8 was clinically beneficial.
9 Next slide.
10 What are some of the current guidelines for
11 triglycerides? These clearly lag behind the LDL because
12 we lack some of the information. But what is some of the
13 information or guidelines that are out there for
14 physicians?
15 Again, as I mentioned, the NCEP, in 1993,
16 before a lot of data and many of these things were known,
17 classified normal triglycerides as less than 200. And
18 in certain populations, this may be too high.
19 Borderline high of 2- to 400, high 400 to a
20 1,000, and very high being 1,000. The recommendations
21 at that time, in '93, was that people in this group were
22 candidates for therapy, particularly if they had other
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
87
1 risk factors for coronary artery disease; for example,
2 diabetes or a family history or genetic disorder in lipid
3 metabolism. That was '93.
4 Next slide.
5 Let me -- going on, for diabetics, in 1999,
6 the American Diabetes Association has been a little more
7 aggressive based on some -- not really the data that I
8 showed you, but some of the data with respect to, what
9 should the lipid levels be in diabetics?
10 Clearly, a very low LDL level in diabetics,
11 low risk -- HDLs greater than 45, and made it very clear
12 that they wanted to see diabetics with triglycerides less
13 than 200.
14 Next slide.
15 So, in summary, is that there is certainly
16 increasing evidence for triglycerides as an increasing
17 independent risk factor for coronary artery disease. The
18 properties are that it is certainly very important in
19 subgroups -- particularly women in diabetics -- as a risk
20 factor. It's a synergistic factor and metabolic marker
21 as well, and that we are now beginning to see some clinical
22 trial data that it's beneficial in certain groups.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
88
1 And as I said, according to Dr. Ginsberg, it's
2 the complicated answer to a simple question. But it's
3 still a clinical utility.
4 Let me just show you the next two slides,
5 showing you what triglycerides means to me, as a physician,
6 and what these numbers mean to patients.
7 For the physician, triglycerides is a very
8 important component in terms of the global risk
9 assessment. By incorporating triglycerides along with
10 LDL and HDL, you can have a much better idea of what their
11 risk is for coronary artery disease, in terms of predicting
12 disease.
13 For those who have existing disease, or those
14 who are diabetics, it now becomes a target for therapy
15 in terms of a goal. And this is really based on -- now,
16 this -- you know, do we have the absolute number? Do we
17 know what we need to treat to? No. Okay?
18 But certainly based on the epidemiological
19 data, some of which I showed you, and some of these other
20 numbers, it certainly seems prudent that we should be
21 addressing triglycerides in these patients. It certainly
22 can help in situations that are not covered by the NCEP
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
89
1 guidelines in terms of what to do. For example, a
2 middle-aged man like me, who might have an LDL less than
3 130, but my triglycerides might be 300.
4 And, finally, and a very important point for
5 the triglycerides, is that when I see these numbers, you
6 always to think of, is something else causing this? Do
7 they have diabetes? Do they have hypothyroidism? Maybe
8 it's their medication? Or is there a genetic disorder?
9 Those are clues that need to tip off the resident.
10 And, actually, one of my favorite tipoffs for
11 the resident is when the lab results come back, when you're
12 in the hospital and the lab results come printed out from
13 the lab, and it says they had to spin the sample. That's
14 usually a good clue that triglycerides are very high, which
15 usually means they are 2,000 or so.
16 Triglycerides is another number to target.
17 Now, it's not supposed to be another number to target.
18 It's another number that they need to know in terms of
19 their risk, and as well as to be aware of.
20 Measuring triglycerides -- and, again, and
21 all lipids -- is certainly important feedback to them in
22 terms of how they're responding to the lifestyle changes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
90
1 and how they're responding to therapy.
2 And, finally, this feedback is quite
3 important in terms of improving compliance, which is one
4 of the major difficulties in treating people with lipid
5 disorders because they don't feel bad unless they have
6 a heart attack.
7 Next slide.
8 So with that, I have finished. Thanks.
9 CHAIRMAN NIPPER: Thank you, Dr. Deeg. I'm
10 sure that we will have questions for you, so stay close.
11 DR. ANAOKAR: Good morning. I'm Sunil
12 Anaokar with Polymer Technology Systems, and I would like
13 to present to you some specific information on the device
14 that we have, the assay system for measuring triglycerides
15 in whole blood samples.
16 I would like to present to you how the device
17 works, how it is used, and some specific performance data.
18 And if it is okay with you, Mr. Chairperson, and the rest
19 of the panel, I would like to ask two of my colleagues,
20 two scientists from Polymer Technology Systems, to also
21 present some other information, such as total system error
22 of the system, as well as some information pertaining to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
91
1 the labeling.
2 Thank you.
3 First, I would like to present to you the test
4 procedures for the device. You have seen the picture of
5 the device before. To actually use this system is very,
6 very easy, very simple. To use it, all one has to do is
7 first insert the memory chip into the instrument.
8 The memory chip is a micro chip that has
9 information such as the chemistry that is supposed to be
10 run, the assay, the lot number of the reaction strips,
11 the calibration code for that particular lot, and the
12 expiration date for the lot. That way, if a wrong
13 chemistry is run, a wrong strip is used, or an expired
14 lot or strip is used, then the instrument simply doesn't
15 work.
16 The instrument is turned on, and the user will
17 check the messages. If the message on the screen says
18 "insert strip," then all one has to do is apply one drop
19 of blood, by finger stick, to the strip, and then insert
20 the strip into the instrument, wait for generally about
21 60 to 70 seconds, and the results appear on the screen.
22 The next slide shows how the assay works.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
92
1 It's an enzymatic assay. The red blood cells are separated
2 from the plasma by a couple of membranes, and the plasma
3 then comes in contact with the reaction membrane, and the
4 reaction membrane has the lipoprotein lipase
5 glycerol-coronase, glycerol phospheroxidase*, ATP, and
6 the chromagens, and the end result of these reactions is
7 the colored product, the intensity which is measured by
8 the instrument, and the intensity of the color is
9 proportional to the concentration of triglycerides in the
10 sample.
11 I would like to present some performance data
12 now. First, the position -- the position in the hands
13 of the consumer was performed in two different ways.
14 First, we did a study where three lay users
15 who had never worked in the labs, never had the experience
16 of using any lab devices, were given a total of six
17 different samples. Each user got two samples at two
18 different levels of triglycerides, and they were asked
19 to prepare those blood samples and illustrate 20 times.
20 And as the percent CV is shown at the bottom
21 of this table, they varied a little bit, but the highest
22 percent CV that they got was 8.8. And please keep in mind
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
93
1 that these people had never used any lab device before
2 and had never worked in the laboratory.
3 The other study was done at three different
4 sites, where 20 lay users participated in the study at
5 each site, and each user was given a total of three
6 different controls who -- that three controls had
7 concentrations -- three different concentrations of
8 triglycerides, and they used the drop method.
9 They added those controls by putting one drop
10 of the control on the strip, and the -- again, the percent
11 CV is shown here at the bottom of this table.
12 The linearity of the assay was -- in the lab
13 by lab professionals or lab technicians. Five samples
14 were used. These samples were concentrations that fell
15 within the analytical range of this assay system; that
16 is, between 30 and 500 milligrams per deciliter. And as
17 you can see, in the data that's shown on the bottom, the
18 regression data, with a slope of 1.0 and coefficient of
19 0.99, that experiment proved the linearity of the assay
20 quite adequately.
21 The interference study was then performed,
22 by taking a number of substances and their interferences
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
94
1 with this assay system. They are all listed on the left
2 column of this table, and what the right column shows is
3 the highest concentration of these substances where no
4 significant interference was observed. However, this
5 data is included in the package that you received from
6 the FDA.
7 Interference or any possible influence of
8 cholesterol concentration on the triglyceride
9 measurements was tested by assaying a number
10 of -- actually, 65 samples with different cholesterol
11 concentrations. And the triglyceride concentrations, as
12 assayed in our system, was compared to the triglyceride
13 concentrations assayed in the reference assay. And as
14 you can see, there is no significant bias influenced by
15 cholesterol concentration.
16 The accuracy of the assay was further tested
17 by performing a consumer study at nine different sites
18 with about 382 lay users that participated in the study.
19 Every user was given the system, the device, the
20 BioScanner instrument, the strips, and the memory chip
21 and a lancet device, so that they could take their own
22 finger stick.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
95
1 They were only given that written procedure
2 for the -- for performing the finger sticks as well as
3 performing the assay. No other instructions were given,
4 and they performed their own assay.
5 And the readings that we received on the
6 BioScanner system were compared to our in-house reference
7 assay, which is a reagent system from Sigma Diagnostics,
8 which is the reagent for automated lab analyzers.
9 The reference assay was then compared to check
10 the validity of the reference method. It was compared
11 to an assay that had been performed in one of the
12 CDC-recommended labs. In our case, it was Pacific
13 Biometrics in Seattle, Washington. It's one of the
14 network labs, the so-called CRMLN or the Cholesterol
15 Reference Method Laboratory Network.
16 They compared -- they assayed the samples with
17 their method, and when the numbers were compared to the
18 numbers that we got on the Sigma -- with the Sigma reagents,
19 a correlation of 0.98 and a slope of 0.94 was obtained.
20 Then we compared the assay of the -- of the
21 CRMLN lab was as the results we had from the other system
22 that the consumer got directly. And we had a correlation
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
96
1 of 0.93 with a slope of 0.85.
2 Since the finger stick blood drops can vary
3 in volume, we did a volume study where we looked at volumes
4 of the whole blood from a patient that were added to the
5 strip. And as you can see in the table, we don't see any
6 significant difference between 15 microliters of blood
7 and up to 30 microliters of blood.
8 And, finally, hematocrit values varied from
9 patient to patient. We did a study -- a hematocrit study
10 looking at blood samples for different hematocrits, the
11 same blood sample but different hematocrits. The
12 hematocrit was adjusted for the particular sample.
13 We took two samples with two different levels
14 of triglycerides, and they were tested by the reference
15 assay, and then on the BioScanner. And as you can see,
16 the values did not change very much between 30 to about
17 50 percent hematocrit. Above 50 percent, they drop
18 considerably.
19 So, at this point, we will glad to answer any
20 questions, or during the question and answer period. But
21 at this point I would like to introduce to you Dr. John
22 Pasqua, who is the manager and senior scientist in our
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
97
1 R&D group.
2 DR. PASQUA: Thanks, Sunil.
3 Well, it's getting late in the presentation,
4 so there's nothing to energize an audience better than
5 a good discussion of total system error.
6 (Laughter.)
7 The total system error, as I'm defining it,
8 is -- as it is commonly defined -- is bias plus two times
9 the standard deviation. It's a good measure of analytical
10 performance in a diagnostic device. And it becomes
11 especially critical as the panel addresses, on one of the
12 questions, whether NCEP standards for laboratory
13 instruments applied to over-the-counter whole blood
14 tests.
15 The first -- the top table -- well, first,
16 let me say that the random error component of all these,
17 of all the system error calculations, are derived from
18 the lay user study, where three users each had three blood
19 samples, N of 20, and different instruments. And that's
20 how they got the CV percent.
21 The bias I calculated two different
22 ways -- commonly, the first table, the top table, that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
98
1 was calculated from a regression equation versus our
2 reference method. Simply plugged in the reference value
3 into the regression equation, let's look at 100, got 97.
4 The bias was minus 2.8. There are some rounding issues
5 there.
6 The SD from the lay user study was 8.5. I
7 calculated the total system error -- again, bias plus two
8 times SD was 19.9 mils per deciliter. For 100, it's 19.9
9 percent. And that's the way it goes all across.
10 You note that the higher the reference value
11 goes, it seems like if you calculate it from the regression
12 equation, the more negative the bias becomes. None of
13 the values -- 19.9 percent for 100, 23.2 percent for 200,
14 16.5 percent -- you think it's better than this?
15 (Laughter.)
16 19.9 percent, 23.2 percent, 16.5 percent.
17 Neither meet the 15 percent for the NCEP guidelines,
18 not -- all three of them don't.
19 The way -- another way to calculate the bias
20 was from the surrounding data points. I got this
21 recommendation from actually John Dawson, a statistician
22 for the FDA. He thought -- for example, for 200, I took
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
99
1 the point with the reference value from 190 to 290 mils
2 per deciliter, calculated the bias, and took the average
3 of the bias. And the average bias for 200 was minus 10.
4 Again, the random error components were the
5 same. Final calculations come to 23.8, 22.5, 14.1
6 percent. None of these point estimates, again, meet the
7 15 percent requirement that the NCEP has, which is five
8 percent bias and five percent -- the NCEP recommendation
9 is five percent bias and five percent CV.
10 But it's interesting to note that the bias
11 calculated from the surrounding points doesn't go up as
12 steeply as the regression. And I kind of favor this
13 approach. This is real. This is more predicted. I think
14 this is more representative of the true bias in that area,
15 although, I've got to admit, for 400, there were only five
16 points.
17 Here is -- I just want to show you how I
18 calculated the bias by the surrounding points. Between
19 95 and 105, for the 100 sample; 190 and 210 for the 200;
20 381 to 413 for the 400; and 23, 17, and 5. And here are
21 the biases calculated from that.
22 The FDA seemed concerned about the confidence
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
100
1 intervals around our estimates, especially the SDs. So
2 let me just briefly go -- so what I did for these TSE
3 calculations -- all of these TSE calculations here are
4 calculated from the nominal bias. But what I did was,
5 since I calculated the SDs in mils per deciliter, and I
6 gave the corresponding percent, I had the lower confidence
7 interval in nominal and the upper 95 percent confidence
8 interval, and I used these to calculate these numbers here.
9 And, again, against the NCEP standards, even
10 the lowest -- if we use the lowest estimate of that SD,
11 21.3, 19.8, doesn't meet 15 percent. At the high level,
12 at the 400 level, we do approach the 50 percent sample,
13 be it the low estimate, nominal -- if you use the upper
14 estimate, we don't make it.
15 Dr. Anaokar presented a slide the last time
16 showing that we have a negative bias against the CRMLN
17 reference method, and we just want to
18 propose -- we -- that's fairly easy to remedy. We can
19 either switch the CRMLN reference method in our own lab.
20 We can have them target -- we can calibrate the serum,
21 having the CRMLN lab assign values to it by running their
22 method, or we can do both. And that's just something that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
101
1 we're very willing to do if the panel decides it would
2 be in our best interest.
3 And that's all I have to say.
4 Margo?
5 CHAIRMAN NIPPER: Before you leave, I want
6 to make sure I have your name correct. Is it P-A-S-Q-U-A?
7 Is that --
8 DR. PASQUA: Yes.
9 CHAIRMAN NIPPER: -- the spelling? Is it Dr.
10 or Mr.?
11 DR. PASQUA: Dr.
12 CHAIRMAN NIPPER: Dr. Thank you, Dr.
13 Pasqua.
14 MS. ENRIGHT: I'm Margo Enright. I'm the
15 Manager of Clinical Affairs for Polymer Technology
16 Systems, and I'm just briefly going to go over the labeling
17 that we have presented for this product, for the BioScanner
18 triglycerides.
19 Besides a user guide for the BioScanner
20 itself, we have, of course, a package insert that goes
21 through and has all of the key elements that are required
22 for a package insert, giving instructions on how not only
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
102
1 to run the test but also describing our performance. And
2 I'm going to specifically take a look at some of these
3 items.
4 Besides the package insert for the strips
5 themselves and the user guide, we will provide a control
6 material. And this is manufactured for us. We don't
7 manufacture our own controls. We will provide a control
8 material to be used with the BioScanner 2000
9 triglycerides, and this is the package insert for that
10 product.
11 So, briefly reviewing the labeling for this
12 triglycerides product, just to give you a little history,
13 there have been several iterations of this labeling. Our
14 first submission to the FDA for the OTC triglycerides
15 product was December 30, 1998. And we followed all of
16 the FDA recommendations with respect to product labeling.
17 We have reformatted and revised for input from the FDA.
18
19 And if the FDA has additional
20 recommendations, we would like them to make those
21 additional recommendations because we are very willing
22 to make any changes. And if we misinterpreted any of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
103
1 suggestions they've made, we're very willing to make any
2 changes to our labeling. So we wanted to make sure that
3 we went on the record saying that as far as the labeling
4 goes we are very willing to make any requested changes.
5 Very briefly, hit some of the key items in
6 the package insert, the intended use. This is formatted
7 at a seventh grade level for a home user, describes what,
8 why, who, and when, basically tells them that this measures
9 triglycerides and finger stick blood.
10 As you may note, what's in bold we very
11 carefully note to the consumer that they need to consult
12 their health care provider for use of the device. Also,
13 use of this test may give you an early warning that you
14 should see your health care professional. Those are real
15 critical pieces of the labeling.
16 Expected values -- we use the NCEP expected
17 values recommendations, again, with the caveat in bold
18 at the end. If your triglycerides result is above 200
19 milligrams per deciliter, you should contact a physician
20 and follow your physician's advice. We try and make it
21 very clear that this is not to be used as a substitute
22 for seeing your physician.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
104
1 In the performance section, accuracy, there
2 was an issue that as far as how we reported the accuracy,
3 and we look to the FDA to make a recommendation. And maybe
4 their statistician can help us in terms of we added the
5 additional line, "The result of these studies shows that
6 the BioScanner test system compares well to the laboratory
7 instrument and can be run by a consumer or a layperson
8 in their own home with accurate results 95 percent of the
9 time."
10 If we can get some help from the -- suggestion
11 from the FDA as far as how to calculate that. There was
12 an issue regarding that. But we will follow the guidelines
13 that the FDA gives us on calculating that accuracy because
14 there are -- as you know, just like total system error,
15 there are a number of ways to calculate accuracy. So we
16 will report as the FDA would like us to report.
17 And, finally, before I turn this back over
18 to Jim Connolly, just a summary of our performance. As
19 Manager of Clinical Affairs, it is incumbent upon me to
20 just summarize and let you know that we believe that this
21 product, and based on all of the studies that we have done,
22 and the information that we have provided, the performance
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
105
1 studies, show this product to be safe, effective, and
2 substantially equivalent to a predicate device.
3 I'll now give this back to Jim Connolly to
4 summarize.
5 MR. CONNOLLY: Thank you, Margo.
6 Could we turn the lights back on? I have to
7 read my own writing.
8 I had a lot of great ideas. There are some
9 compelling statements here. But I think I'd rather
10 just -- after looking at some of the data, I think part
11 of the things that are missing, especially when we get
12 more significant figures, or words such as "clinical
13 utility" or "clinical significance," and I think it's
14 interesting that the higher the triglycerides went the
15 better the performance seemed to be. I don't know whether
16 it was because of low numbers or the math, but the point
17 was that around some of the critical areas there were some
18 numbers there.
19 I don't think I want to defend the lack of
20 performance, but I think whether someone has got a 50
21 triglyceride or a 75 triglyceride, that's not near as
22 important as if they have one over 400, or certainly over
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
106
1 200.
2 I've got eight minutes to do this, but I don't
3 think it's worth eight minutes to go through this thing.
4 But we think we provide a product for the patient and
5 for the physician to be used in these risk factors. And
6 in Dr. Ginsberg's paper, I also liked the -- or liked or
7 thought about the ending of the paper, about has the
8 question been answered, or is it there yet?
9 And if you combine that with the words that
10 are in the FDA document at the end of the presentation
11 today, as to risk versus benefit, I think the benefits
12 here are enormous, particularly in the diabetics and in
13 women. And I think there is certainly, from Dr. Deeg's
14 presentation, a lot to be said about the general public,
15 the non-diabetic folks.
16 So I think -- I'm not sure what the risks are
17 of having this product on the market. I think the risks
18 of not having it on the market to those that are at
19 risk -- little overusage of that word -- are pretty
20 dramatic.
21 As to the jury's out -- I mean, we hear a lot
22 of that thing about the jury being out on triglycerides.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
107
1 Is it real or is it not? Papers -- some say they are;
2 some say that it isn't. But, again, I think the benefits
3 of the product in giving the patient access to some care
4 and the physician and the patient -- when the tests are
5 performed right there, to do the consultation, and not
6 going back to our old system of drawing blood samples,
7 sending it out to a lab, a couple of days later,
8 questionable thing, the patient was not contacted, or our
9 current health care system really doesn't provide a system
10 to encourage the patient to -- it should be the
11 physician -- to follow up on things like this. It just
12 takes more time.
13 If I fell on that box, it would be a heck of
14 a conclusion.
15 (Laughter.)
16 So let me take to the last part of my notes
17 before I fall down here. You know, these values actually
18 do address the major costs in health care -- diabetes and
19 cardiovascular disease. Any impact we can have on
20 probably the bulk of our dollars -- that is, greater than
21 probably 70 percent of the total dollars spent in the U.S.
22 are on these two diseases.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
108
1 So I think any impact that we can have on these
2 two diseases is certainly worth looking at. And I think
3 it's definitely beneficial, and I'm a little at loss as
4 to what the risk could be.
5 What did I leave out, Sunil?
6 We have a little note to end this with. I
7 got interested in dyslipidemia because, of all people,
8 I became insulin-resistant shortly after this prolonged
9 approval process. I don't know if they're related, but --
10 (Laughter.)
11 -- I now have a genuine interest in
12 insulin-resistance, particularly in high triglycerides
13 and low HDLs.
14 As you know, we also had a hemoglobin A1C
15 product that's about to come to the FDA. I can tell you,
16 nine years ago, when I was working on a somewhat similar
17 project, one day a week we would take a scientist to lunch
18 or to dinner just to kind of keep the company warm and
19 close. We've always worked, since our large company
20 affairs, with small companies.
21 And in small companies you use a lot of your
22 own blood. So not only do people know a lot about you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
109
1 in a small company because they see you every day -- it's
2 pretty hard to hide your warts -- you also find out a lot
3 about your lipids.
4 As a medical technologist, I've seen a couple
5 thousand or so serum samples sitting in racks, and, in
6 1989, I saw some of my own and couldn't believe it. I
7 did some of the original Peculan-David enzymatic
8 triglyceride work back in the '70s, and I know my
9 triglycerides are around 70.
10 But at age -- at an older age --
11 (Laughter.)
12 -- it's about 500. So I just assumed that
13 Chinese restaurant I've taken this scientist to was the
14 issue. Everything else about me was normal. I dieted
15 a little bit, drank three gallons of water, took another
16 measurement, and sure enough I got it to half what it was.
17 So, therefore, it wasn't a problem until about nine years
18 later.
19 I've always had a low HDL, so I'm genetically
20 absolutely fit for this thing. HDL at 35, 34 -- and do
21 I know that's the correct number? Yes, because I've had
22 it run hundreds, if not thousands, of times. Now I've
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
110
1 got an HDL of 25, a triglyceride of -- in the hundreds,
2 certainly above 400 milligrams, and I've become
3 insulin-resistant. Now I'm a proud supporter of
4 Bristol-Meyers-Squibb.
5 But this thing could have been intervened
6 with, as in many people, nine years ago when there was
7 an indication there that something was awry, as Dr. Deeg
8 mentioned. So I think the use of this thing in coronary
9 artery disease is -- if it's a question, I think it's a
10 mute question, about risk and benefit.
11 And the use of diabetes -- I think there's
12 no question about people who are at these high risks that
13 need to monitor their therapy, their diet, their exercise,
14 and their medications. Triglycerides is one of the best
15 ways to do it.
16 And I think there's absolutely no question
17 about triglycerides being used with our other approvals
18 that we're patiently waiting for in HDL that would be even
19 of more impact on this diagnosis of coronary heart disease
20 and some of the assessments used to determine if someone
21 is becoming dyslipidemic, and I won't say the other word.
22 But it certainly has impact outside of cardiovascular
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
111
1 disease.
2 So I think this panel should obviously approve
3 this and make the right recommendation, so we can get this
4 product into the market and make an impact for those people
5 that are in this drug situation right now with
6 cardiovascular disease and those diabetics who are
7 consuming a vast amount of our physician's time and our
8 health care dollars.
9 Thank you.
10 CHAIRMAN NIPPER: Thank you, Mr. Connolly,
11 and I appreciate the presentations made by the other
12 members of your company and other presenters.
13 I'd like to mention to the panel a couple of
14 housekeeping items. One that I continually am reminded
15 of by FDA staff, and I thought of this on my own, so I
16 must be getting the message, is that considerations of
17 cost are beyond the scope of the panel's deliberations.
18
19 We are always interested -- I'm sure that all
20 of us as individuals are interested to hear these from
21 time to time. But in dealing with this device, we confine
22 ourselves to the FDA's mission in this area.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
112
1 The second thing that I wanted to just call attention
2 to the panel is that I was -- I remarked that at the end
3 we will be making some final recommendations. That's
4 different than a vote. And so I think that as we begin
5 to formulate our questions, to direct our questions to
6 the sponsor and to the FDA presenters, we will be thinking
7 in terms of recommendations as opposed to an up and down
8 vote at the end.
9 MR. CONNOLLY: Dr. Nipper, may I have a
10 moment, please?
11 CHAIRMAN NIPPER: Yes.
12 MR. CONNOLLY: If the costs are not to be
13 considered in this product, then I think I need to emphasize
14 quality of life and lack of death as being major points
15 in helping diagnose these diseases.
16 CHAIRMAN NIPPER: Thank you. Safety and
17 effectiveness is a good guideline there. It's not that
18 we're not interested in cost, but I think the FDA is not
19 allowed to bring cost effectiveness or cost of treatment,
20 either high or low, into the decision about safety and
21 effectiveness.
22 At this point, we are a few minutes ahead of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
113
1 schedule, so I'd like to use the time to add to the question
2 and answer period. And I think that if the committee
3 is -- is the panel is willing, what we'll do is try to
4 go around the room and allow individuals to -- are you
5 motioning to me, sir? Okay.
6 We will allow individual members of the panel
7 to ask questions of the presenters. I'm going to ask one
8 question, and then I'll let Dr. Rifai pick up from me.
9
10 The first -- the question that I had for Dr.
11 Deeg is I think a rather simple one. And you can approach
12 the microphone, so you can be on the record. Dr. Deeg,
13 in the clinical studies that you cited in your
14 presentation, I notice the word "non-fasting" appeared
15 many times. Were there any of the clinical studies that
16 you cited in which the triglyceride data were -- I'm sorry,
17 I started to say fasting -- were there any studies in which
18 the triglyceride data obtained was non-fasting or were
19 the subjects fasting for the data?
20 DR. DEEG: The studies I cited were all for
21 fasting. I'll just add in a point about -- an issue about
22 non-fasting triglycerides and post-prandial
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
114
1 dyslipidemia -- is appearing to be a very important
2 predictor for coronary artery disease at this point.
3 CHAIRMAN NIPPER: Okay. But you didn't cite
4 any studies to --
5 DR. DEEG: Correct.
6 CHAIRMAN NIPPER: -- show -- to support that
7 theory today.
8 DR. DEEG: Well, no, I have a slide I could
9 show you quick. Would you like to see it?
10 CHAIRMAN NIPPER: Why don't we go around the
11 room, and then if we have a little time left -- because
12 I know that the panel probably has a lot more questions.
13 We'll have time for deliberation about that.
14 Dr. Rifai?
15 DR. RIFAI: I just have also a few short
16 questions for Dr. Deeg. Forgive me if they are naive.
17 How often do you see your patients?
18 DR. DEEG: How often do I see my patients?
19 DR. RIFAI: Yes.
20 DR. DEEG: Typically, diabetics I see about
21 every four months.
22 DR. RIFAI: What do you order --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
115
1 DR. DEEG: What do I order?
2 DR. RIFAI: -- in terms of lab tests every
3 time you see them?
4 DR. DEEG: For diabetics, I order a
5 hemoglobin A1C for glycemic control; I now order a lipid
6 panel every time.
7 DR. RIFAI: And you think more frequent
8 measurement of lipid -- of triglyceride per se is valuable?
9 DR. DEEG: Certainly, in the initial
10 management of the patients, more frequent -- as you are
11 fine-tuning their therapy and getting to your goals, and
12 things of that nature. Once you have attained your goals,
13 then you can cut back a little bit on the frequency.
14 DR. RIFAI: And at which level do you consider
15 treatment? Let's say just -- you have increased
16 cholesterol. At which level do you intervene?
17 DR. DEEG: Well, for diabetics, because the
18 risk for coronary artery disease is so high, I'm very
19 aggressive, and the ADA is, you know, consistent with this
20 in terms of treating their cholesterol levels and their
21 triglycerides.
22 So the ADA is recommending that LDL
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
116
1 cholesterol be 100 for all diabetics, and that's in part
2 because their risk is so high. The triglycerides, again,
3 along the ADA guidelines, try and keep them certainly less
4 than 200, sometimes even lower than that.
5 DR. RIFAI: Thank you.
6 DR. KIMBERLY: I have some questions for Dr.
7 Anaokar regarding the comparison with CRMLN laboratory.
8 How were the samples collected for the -- that were sent
9 to the network laboratory for the CRMLN laboratory? And
10 also, what type of pre-analytical considerations were
11 taken as far as the patient or the layperson as far as
12 collecting the finger stick and assuming the serum sample?
13 MS. ENRIGHT: Dr. Kimberly, Dr. Anaokar has
14 asked me to answer your question. As far as the samples
15 for the CRMLN laboratory analysis, the samples were
16 collected -- both finger stick samples were performed by
17 a professional on the BioScanner as well as the lay user
18 themselves on the BioScanner. And we drew venous samples
19 and collected serum samples, which were then frozen for
20 the CRMLN labs procedure and shipped to the CRMLN lab on
21 dry ice.
22 DR. KIMBERLY: Okay. What type of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
117
1 pre-analytical considerations did you -- were all of the
2 patients seated? I mean, so that they were all -- I mean,
3 so that the samples were all drawn -- were they all drawn
4 in the same timeframe?
5 MS. ENRIGHT: Yes. The patient finger stick
6 results were run at the same time, within the -- the draw
7 was either prior to the finger stick or after the finger
8 stick, depending on the number of laypersons that we had
9 available. But waiting -- so that was just a logistical
10 thing. But within 10 minutes of each other, so they were
11 drawn at the same time.
12 DR. KIMBERLY: Were they seated during that
13 whole time? Were they in the same posture? I mean, the
14 NCEP makes recommendations regarding how the patient
15 should be treated beforehand.
16 MS. ENRIGHT: Right. Right. We did -- we
17 did follow -- and I think there are some guidelines in
18 the -- there are some NCEP guidelines for finger stick
19 technique, which is that -- those guidelines were provided
20 to all of the persons involved in doing the finger sticks
21 in our clinical trials of professionals.
22 The lay users themselves were given written
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
118
1 instructions on how to perform the testing. And if there
2 was any variation, it was due to the lay user's use. The
3 professional -- in a professional result on the
4 BioScanner, the finger stick was compared to the lay users.
5 And differences between the lay user and the professional
6 results may have had a little bit to do with the fact that
7 many of the lay users that we used had absolutely no
8 experience doing a finger stick. But they were not in
9 any way coached. They were just given written
10 instructions.
11 DR. KIMBERLY: Okay. Thank you.
12 DR. CLEMENT: Steve Clement. I'm not quite
13 sure who to address this question to. The sponsors
14 maybe -- select the person.
15 The way I look at this data is the sponsor
16 is asking the FDA to set up a different standard for
17 accuracy, total system error if you will, with the
18 increased benefit that this is something that is going
19 to have better access for the patient, which I think is,
20 you know, a very good cause.
21 From the sponsor's point of view, what would
22 you consider a standard error that's valuable to the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
119
1 patient? Or total system error. Excuse me.
2 MR. CONNOLLY: I think there's a couple of
3 answers, and I grabbed it quick before -- someone else
4 may want to get in line.
5 (Laughter.)
6 One which is very true is that probably the
7 majority of the people using this product are going to
8 be diabetics who currently know how to do finger sticks.
9 They're going to get better results. It's my feeling
10 that those results would meet the guidelines.
11 I only had 15 minutes to put this together.
12 I think that when you go to the general public and patients
13 where -- we go to several different places to get people
14 involved in this thing, and it's difficult for people who
15 have never done finger sticks. And, as you know, there
16 has never been a market for a product for a finger stick
17 other than glucose.
18 So it's -- I think it's a monumental fact that
19 we got the kind of answer that we did, assuming that
20 probably 95 percent of these patients had never done a
21 finger stick before. So in that case, I think they are
22 probably not going to be the users of the product, but
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
120
1 to hold us to, you know, totally unknown people, to NCEP
2 guidelines that were developed for large analyzers with
3 incredible performance on CVs, using serum. None of them
4 use whole blood, so there really are no standards for whole
5 blood.
6 DR. ANAOKAR: That is absolutely right. We
7 did check for any guidelines. We looked for guidelines
8 for whole blood that NCEP may have. We found out that
9 there are none. I even checked with Dr. Naito, and he
10 said there are none.
11 There is a precedent, though. There is a
12 product on the market. It's been on the market for some
13 time for cholesterol, and that does not meet NCEP
14 guidelines for either the total system error or the
15 position. And it is a whole blood product.
16 Actually, the total system error on that
17 product is more than twice what is recommended by NCEP.
18 MR. CONNOLLY: And that product was approved
19 by the FDA.
20 DR. ANAOKAR: It is approved by FDA. It's
21 been on the market for over-the-counter use.
22 MR. CONNOLLY: Without a panel.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
121
1 DR. CLEMENT: I'm still looking for a number,
2 I mean, compared to -- like, for example, if we compare
3 it to another reference method, such as serum, which you've
4 done in your case.
5 MR. CONNOLLY: Do you mean, what do we think
6 the number should be for whole blood?
7 DR. CLEMENT: Right.
8 MR. CONNOLLY: Non-testers?
9 DR. CLEMENT: Right.
10 MR. CONNOLLY: I don't think I know. I think
11 it is a number in the high range. With untrained people,
12 with a 15 or a 14-point something or other CV, which met
13 the goal, and a CV of 5.06 -- .06 outside the goal -- I
14 think on an unmeasured sample, on an untrained person,
15 that's incredible, because there are many products, as
16 the cholesterol we just mentioned, are approved that came
17 nowhere close to that.
18 So what's the number? What was our number,
19 22? 22, I think would be the right number for whole
20 blood --
21 (Laughter.)
22 -- untrained people.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
122
1 DR. GUTMAN: Let me just correct the record.
2 It is true that the cholesterol product did not meet NCEP
3 guidelines, but it was off by about two percent. It was
4 in the ballpark. There was a lot of discussion about that
5 product. And, in fact, in the labeling of that product,
6 the cut points were set so that the signal was actually
7 below the traditional cutoffs.
8 The signals were set at 190 and 230 rather
9 than at 200 and 240, so that the error in the system would
10 be towards false positives and people would be driven in
11 to see their doctors more often than not, the notion being
12 it wouldn't be terribly harmful to see your doctor an extra
13 time. It actually probably isn't terribly harmful on a
14 single occasion to miss your cholesterol either, since
15 there is no immediate adverse negative impact.
16 But there was an error, but it wasn't off by
17 a factor of two. It was about two percent.
18 CHAIRMAN NIPPER: Are you answering a
19 question here, Mr. Connolly?
20 MR. CONNOLLY: Yes.
21 CHAIRMAN NIPPER: Which question are you
22 answering?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
123
1 MR. CONNOLLY: I'm answering the question
2 about the cholesterol product that did not meet the NCEP
3 guidelines.
4 CHAIRMAN NIPPER: I don't believe anybody
5 asked that question. Does someone want to ask that
6 question, so he can present this information? I believe
7 the sponsor brought up that topic. And I don't want to
8 be contentious here, but we have only a limited time for
9 committee deliberation. And if the committee is willing
10 to hear that, if the panel is willing to hear that, then
11 I'm -- this information -- I am willing to present it.
12 But I'd like to get around the room before
13 we break, if we could.
14 DR. ROSENBLOOM: Yes. And then let's see if
15 we've got --
16 CHAIRMAN NIPPER: Yes, let's see if we have
17 time, Mr. Connolly. I apologize for cutting you off.
18 I don't mean to remove that information from the table
19 if you want to present it. We may have time later today.
20 Dr. Rosenbloom?
21 Were you finished?
22 DR. CLEMENT: Yes, I'm done.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
124
1 CHAIRMAN NIPPER: Dr. Rosenbloom?
2 DR. ROSENBLOOM: I had some question about
3 labeling, if there was any operational research or any
4 opinion to support the recommendation in the labeling for
5 over-the-counter use that the testing be done two to three
6 times per year, and it doesn't say in whom. That means
7 in everybody, even us young middle-aged folks, who have
8 no other risk factors. And at least monthly in those with
9 diabetes and in the post-menopausal state.
10 And that seems a rather dramatic
11 recommendation, which is only supported by the statement
12 of Dr. Naito, who is the Chief of Clinical Chemistry at
13 the VA in Cleveland. And I wondered if there was -- and
14 it's not in keeping with the American Diabetes Association
15 recommendations which are that adult patients with
16 diabetes should be tested annually for lipid disorders,
17 with fasting serum cholesterol, triglyceride, HDL
18 cholesterol, and calculated LDL cholesterol.
19 And then, if the values fall in the lower risk
20 levels, assessment may be repeated every two years. Tests
21 resulting in borderline or abnormal values should be
22 repeated for confirmation. Tests resulting in abnormal
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
125
1 values requiring institutional therapy should be repeated
2 following the NCEP recommendations, and lipid values
3 should be reevaluated following a macrovascular event.
4 And then it goes on to talk about following
5 the National Cholesterol Education Program
6 recommendations for children and adolescents.
7 So my question is: what is the clinical
8 practice, background, clinical experience, or operational
9 research on which these recommendations which would be
10 promulgated to the general public -- that virtually anyone
11 should be tested two to three times per year, and that
12 people with diabetes and in the post-menopausal state
13 should be tested monthly without any qualifications?
14 MS. ENRIGHT: In our initial labeling, we did
15 not address the issue of how often these tests should be
16 used. But one of the FDA requests was to add that
17 information to the package insert. And what we did is
18 we consulted. We were told that we had to consult with
19 an expert in the area in this. And as you alluded to,
20 this suggestion came from Dr. Herb Naito from the VA Medical
21 Center in Cleveland. And we are certainly open to making
22 changes in the labeling that address this.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
126
1 Initially, we did not make a recommendation
2 on labeling. So we are open to your suggestions on what
3 we should recommend in the labeling.
4 MR. CONNOLLY: I think there's an issue with
5 compliance that we've tried to address. We understand
6 from several of the people that supply lipid-lowering
7 drugs that approximately 75 percent of those people are
8 off those drugs in nine months. We think monitoring more
9 often would keep the patient more compliant and on a drug.
10
11 And as we all know, when you take this drug,
12 there is no -- you don't feel any better, so there is really
13 no reason to comply. It's a very expensive drug. A lot
14 of diabetics don't have health insurance, and to buy
15 another $2,000 a year worth of statins, or whatever, is
16 a burden there.
17 But the compliance issue I think should be
18 the driving fact in how often they should test. Whatever
19 it takes to keep them on therapy.
20 DR. ROSENBLOOM: That's a very
21 qualified -- that's far more qualified than is in the
22 labeling. What you are addressing is far more qualified.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
127
1 In the labeling, it just says if you have diabetes you
2 should test every month, or if you're post-menopausal.
3 MR. CONNOLLY: We are -- as Margo suggested
4 earlier, we are happy to -- we would change the label
5 appropriately.
6 DR. ROSENBLOOM: And --
7 MR. CONNOLLY: We're willing to make changes.
8 DR. ROSENBLOOM: And, of course, the ADA
9 guidelines are very specific and based on expert opinion.
10 That was my question.
11 CHAIRMAN NIPPER: Thank you, Dr. Rosenbloom.
12 Dr. Floyd?
13 DR. FLOYD: I don't have too many specific
14 questions, but one I will address -- there has been a lot
15 of data presented to show how variable this test is in
16 the hands of consumers. There are two issues that pop
17 into my mind immediately when I see that kind of data.
18 One is, how much variability is in the device the consumers
19 handle? I didn't see any of that presented.
20 And the second is, what accounts for the
21 variability in the hands of consumers? There was an
22 allusion here earlier, although it wasn't addressed
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
128
1 specifically, that the differences in hematocrit make a
2 difference in the values that are obtained. I'd really
3 like to see some data on how that's addressed. Does that
4 relate to the volume of plasma that's actually available
5 to the chemistry to react? And is that accounting for
6 the variability?
7 So there are a number of those specific issues
8 I'd like to see a little bit more data on, just so I could
9 understand where the variability in the test is coming
10 from.
11 CHAIRMAN NIPPER: Thank you, Dr. Floyd.
12 Ms. Kruger?
13 MS. KRUGER: I have some specific questions
14 about the users. I heard studies about pipatsin
15 controlled, and then I heard about a 382-patient study,
16 and I made the assumption that is the one that was both
17 lay people as well as health care providers doing that
18 test. Is that correct?
19 MS. ENRIGHT: The data that we presented
20 was --
21 CHAIRMAN NIPPER: Would you speak into the
22 microphone, please?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
129
1 MS. ENRIGHT: Sorry.
2 CHAIRMAN NIPPER: Thank you.
3 MS. ENRIGHT: The data that we presented was
4 only the consumer data. We also did, in conjunction with
5 that, do a professional study.
6 MS. KRUGER: And how many consumers were
7 there that used the device? How many consumers? Was it
8 382?
9 MS. ENRIGHT: 382.
10 MS. KRUGER: And these consumers actually
11 used their own finger sticks? They poked their own fingers
12 and put it on the drop?
13 MS. ENRIGHT: Yes.
14 MS. KRUGER: Okay. They weren't
15 using -- they weren't pipetting it?
16 MS. ENRIGHT: No, they were not pipetting.
17 The only time when the consumers used pipettes were when
18 the consumers pipetted whole blood that was other than
19 their own.
20 MS. KRUGER: Okay.
21 MS. ENRIGHT: Or if they were doing control
22 studies, because there is no other good way to do that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
130
1 MS. KRUGER: Okay. I didn't see anywhere in
2 the literature -- and I may have missed it -- or in your
3 presentation, answering questions of size of drop. Can
4 the consumer touch the strip? If the consumer smears the
5 blood, are there any error codes if the patient does
6 anything wrong, so they know that there isn't enough blood?
7 How long after they put the blood on the strip do they
8 have until they can put it in? Is there anything on the
9 strip that tells them they haven't gotten enough blood?
10 Can they add more blood? And those are questions that
11 I think are really important in terms of the consumer being
12 able to use it.
13 The other thing is I think that Mr. Connolly
14 is making some assumptions about our patients. One of
15 the biggest problems we have is technique, and that I
16 see -- that's all I've done for the last 18 years is work
17 with people with diabetes. And the biggest problem we
18 have is their ability to get all of these things
19 accomplished.
20 Some of the newer meters, they can accomplish
21 that. But I'll tell you that the techniques we see and
22 the variabilities and the accuracy in some of the older
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
131
1 models of machines are not good. And if you don't have
2 the ability to do some of these things, even though those
3 patients have diabetes and they are doing monitoring, it's
4 not going to help the accuracy.
5 Yes, in some of our newer machines that's
6 true, but if you don't have it built in, just because they
7 have diabetes, we're not going to see better accuracy than
8 what you are presenting today.
9 And then another comment. I would agree as
10 just supporting Dr. Clement and Dr. Rosenbloom in terms
11 of the ADA standards of care for measurements of lipids.
12 The frequently is certainly -- we need to talk about how
13 you label that in terms of your frequency.
14 MS. ENRIGHT: May I address one of her
15 comments?
16 MS. KRUGER: Please.
17 MS. ENRIGHT: The one comment regarding the
18 consumers. One of the things that we did in selecting,
19 we were very careful that we selected consumers and that
20 we, in our demographic data that we gathered on the
21 consumers, we made sure that we had relatively naive
22 consumers; that is, that they had no previous or almost
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
132
1 no previous experience with finger sticks.
2 The other piece of that is most of these folks
3 were not diabetics, which we asked. And you make a good
4 point that diabetics sometimes do have great difficulty
5 with a finger stick. But the difference, even in a
6 diabetic versus the consumers, some of the places where
7 we performed these studies were at shopping malls, and
8 we really had a -- we were really in a worst-case scenario.
9
10 And I think the diabetic, whether or not they
11 have good technique in performing a finger stick, they
12 at least have the benefit of a diabetes educator or the
13 physician. They're allowed to be given some sort of
14 instruction. And we were instructed that these folks were
15 to be given absolutely no instruction and no coaching.
16 MS. KRUGER: You're making assumptions
17 again. Eighty percent of our patients are never seen by
18 a provider as a diabetes educator or by a person who is
19 in diabetes. That's all I'm saying to you is you need
20 to be careful about the assumptions you're making. You're
21 making too many assumptions.
22 Okay. One other comment. You made a comment
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
133
1 also, Mr. Connolly, that if a patient is on medication,
2 and they can look at their triglycerides, that potentially
3 that will keep them on drug. That's a two-edged sword
4 as well because very often a patient who is doing very
5 well on blood pressure medication, or very well on
6 triglyceride medication, or lipid-lowering medication,
7 when they see good numbers, that's when they stop the
8 medication because, after all, they're doing better. So
9 it's also a two-edged sword.
10 MR. CONNOLLY: Two things. One, with the
11 test strip, one of the things we started off with in the
12 very early days of the company were sample size. There
13 is a lot of people in the industry that believe that sample
14 sizes are much larger than they really are. The definition
15 of hanging drops all over the place.
16 But if one were to consider 10-year old
17 standards at hanging drops are somewhere between 20 and
18 40 microliters, with the larger lancets and with
19 professional people doing it, maybe that's true. We
20 believe sample sizes are much, much smaller than that in
21 reality, for diabetics testing each day. And they
22 probably should be.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
134
1 If you want them to test more often, you can
2 get them to use smaller lancets. There is less pain
3 involved. They don't mind testing more often. We've
4 gotten a lot of people who weren't testing involved in
5 testing by using some of the new devices right now that
6 require less sample.
7 We have made the materials in the sample
8 deposition area outside that to be very, very hydrophobic.
9 So it creates -- it loses some of the problem of the sample
10 going in other places.
11 In regards to the people using these things,
12 I've seen people walk -- well, I've seen one, but it was
13 enough to convince me there is no way. There was an
14 instrument that had just been run sitting on the table,
15 and it said -- there was a result displayed, whatever it
16 was. The patient pricked their finger and deposited the
17 drop of blood on the display. Didn't get it result, but
18 it was obvious that this is a huge problem.
19 One of the things we do by dosing outside the
20 meter -- if you -- if a lot of these diabetics we're talking
21 about, or people in the general public -- the over 40 crowd,
22 the bifocal crowd, or, with me, the trifocal crowd -- I
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
135
1 mean, if you've ever tried to look through a finger, or,
2 worse yet, put a strip in a meter and put your head down
3 at the edge of the table, and turn your head sideways,
4 your trifocal/bifocal thing turns into, where is this
5 thing?
6 And by using a strip that you could hold up
7 to eye level, deposit the blood on the strip, I think we
8 minimize the errors that a lot of folks have in sample
9 deposition on those things. And I think, again, the
10 smaller sample size is a big help.
11 And the other assumptions that if your
12 cholesterol is good, and you stop taking the
13 statins -- we've seen that with some of our employees,
14 which I thought was interesting, because we're all on the
15 compliance side. And then the same employee who stopped
16 taking the statin, who does a lot of roadshows for us in
17 demonstrating the product, his cholesterol kept going up
18 and up. And he started taking the drug again.
19 So I guess it is a two-edged sworn. One, will
20 they stop? And, secondly, will they start again? It
21 would seem to me the only thing -- it's like a bathroom
22 scale, I guess. I mean, I've got one I look at regularly,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
136
1 but I have a hard time impacting that unless I take it
2 seriously.
3 CHAIRMAN NIPPER: It sounds like the elements
4 for a clinical study somebody needs to do.
5 Dr. Lewis?
6 I'm sorry. Were you finished, Dr. Kruger?
7 MS. KRUGER: Yes, I am. Thank you.
8 CHAIRMAN NIPPER: Thanks.
9 Dr. Lewis?
10 DR. LEWIS: Yes. My question would go to the
11 labeling again. And I wonder if there's anything that
12 indicates the most suitable time for sampling, so that
13 the user has some well-understood idea if fasting is
14 required and what that means, which I know from experience
15 can mean anything, depending on, you know, how the
16 individual takes that.
17 MS. ENRIGHT: That's an excellent question,
18 because, as you know, triglycerides varies significantly
19 depending on the time of the day and when the last meal
20 was taken. Based on input from the FDA, we used the -- in
21 the labeling, the NCEP recommendation that the sample come
22 from a person that's been fasting for 12 hours prior to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
137
1 the test.
2 DR. LEWIS: Is the term "fasting" spelled out
3 in any way?
4 MS. ENRIGHT: I'd have to take another look
5 at the labeling, but I think -- I think -- and somebody
6 can correct me if I'm wrong because I think we have the
7 labeling -- I think it says not having eaten for the last
8 12 hours. I think that's how we define it to the consumer.
9 DR. LEWIS: As far as drinking and what
10 liquids?
11 MS. ENRIGHT: For best results, you should
12 test after fasting, no food, only water, for 12 hours --
13 DR. LEWIS: Oh, okay.
14 MS. ENRIGHT: -- prior to testing.
15 DR. LEWIS: Thank you.
16 CHAIRMAN NIPPER: I'd like to know if your
17 copy of the labeling is different than ours, because in
18 the packet my copy of the labeling -- and I was -- and
19 I'm Dr. Nipper -- was -- I saw the note for best results
20 test when you have not eaten for 12 hours, a parenthetical
21 comment under "Quick Guide."
22 Did I miss something somewhere? Oh, yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
138
1 Here we go, another one under "Specimen Collection and
2 Preparation." That's where you were. I apologize.
3 MS. ENRIGHT: Okay.
4 CHAIRMAN NIPPER: So it's in two places,
5 Woody.
6 MS. ENRIGHT: Yes. And we may check because
7 there have been -- based on what I presented, we have made
8 several iterations. And I know what you have in the packet
9 you may have -- you'll want to look at the very latest
10 labeling that you have because we did revise the labeling
11 based on the letters from the FDA.
12 CHAIRMAN NIPPER: Good.
13 MS. ENRIGHT: So there were a number of
14 requests, and I believe we -- I believe we -- to the best
15 of our ability, we honored all of those requests. And
16 if we didn't and misunderstood those, we will make those
17 changes.
18 CHAIRMAN NIPPER: Thank you very much.
19 Did your question get answered, Dr. Lewis?
20 DR. LEWIS: Yes, thank you.
21 CHAIRMAN NIPPER: Okay.
22 Dr. Deeg, we'll call on you in the afternoon
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
139
1 to give this, after we -- I'm sorry to be time constrained,
2 but I don't want to hold up -- we need to take a break
3 in a minute or two, and then I don't want to step on the
4 FDA's time.
5 Dr. Janosky?
6 DR. JANOSKY: Yes. I have a question about
7 the consumer clinical study, in particular, and then
8 available data in general, if there are additional
9 available data. It's my understanding of the consumer
10 clinical study -- I would call it clinical, but I think
11 you're referring to it as a consumer study -- DNO380
12 approximately. Were there any values obtained outside
13 the 400 range?
14 And my comment is especially -- is a follow
15 up to some presentation that Dr. Deeg I guess had given
16 this morning about values outside the 400 range, what they
17 might mean, etcetera. If I read that consumer study, I
18 don't see any values outside of that range. And I see,
19 in fact, very few values outside of the 200 range.
20 Am I correct, or do more data exist that we
21 were not presented with?
22 MS. ENRIGHT: You are correct that more data
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
140
1 exist that you were not presented. There has been -- I
2 have in front of me all of the submissions we have made
3 to the FDA, and I know you were given -- you were provided
4 some selected pieces that should give you a flavor for
5 this.
6 But one of the submissions is one that we made
7 June 4, 1999, and that was -- we refer to two different
8 studies in there, and there are -- there were two different
9 studies. One, I believe, was at four sites, and a
10 five-site study, and we summarized those in the September
11 20th submission, which I believe you have those. But the
12 June 4th submission does have the data of the previous
13 study.
14 And it will take me a little bit of time to
15 pull this data out. I have all of the data on disk, and
16 I can tell you -- I can pull it up, and I can tell you
17 exactly how many we have. It is somewhat difficult to
18 get results. We did not screen our consumers. We did
19 not pre-screen the consumers, so they were based on the
20 general population.
21 And, of course, any result above 500 was
22 outside the linearity of the device. We did have results
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
141
1 above 500, but those were excluded from the analysis
2 because they read high on the BioScanner.
3 DR. JANOSKY: Right. I had seen that
4 earlier. I am not necessarily looking for exact point
5 estimates, but one of the issues that I did see by examining
6 your data is that there seems to be bias in terms of
7 underestimation, and actually quite severe in certain
8 cases.
9 So I'm trying to get a feel for what about
10 categorization, and what about these upper levels. How
11 bad is that bias in terms of this underestimation? Because
12 you are talking about making decisions typically on these
13 higher points, a point of 400, a point of 120. And that's
14 where the bias is quite large.
15 MS. ENRIGHT: Yes. I think Dr. Pasqua spoke
16 to that and the total system error. Actually, the total
17 system error, as you noticed, percentage-wise goes down
18 at the higher levels right around 400. It does go down.
19 But the absolute result -- one other thing we did take
20 a look at, because in the analysis we did find that there
21 was somewhat of a negative bias, and that's why Dr. Pasqua
22 presented the recommendation.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
142
1 We did find that there was a bit of a bias,
2 not only between -- there was a bit of a bias that came
3 somewhat from our reference method, and our reference
4 method is not the CRMLN method. Even though it correlates
5 well to the CRMLN method, there was a bias on the reference
6 method.
7 And our suggestion was that perhaps we need
8 to revisit and use -- even though the method correlates
9 well to the CRMLN -- because of the bias what we need to
10 perhaps do, and we're willing to do, is revisit and in
11 our own laboratory for setting the calibration curve on
12 the device, perhaps use the CRMLN method in our own
13 laboratory.
14 DR. JANOSKY: Okay.
15 MS. ENRIGHT: To eliminate that bias.
16 DR. JANOSKY: Yes. I'd probably like to talk
17 more about the issue, but perhaps after the FDA presents.
18 CHAIRMAN NIPPER: Okay. Thank you.
19 Dr. Doumas?
20 DR. DOUMAS: About when to run this test?
21 CHAIRMAN NIPPER: Basil, pull that
22 microphone over a little nearer to you.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
143
1 DR. DOUMAS: I'm sorry.
2 CHAIRMAN NIPPER: Thank you.
3 DR. DOUMAS: About the frequency of running
4 the test, the recommendation is attributed to Dr. Naito.
5 Do you know if Dr. Naito sees or treats patients?
6 MR. CONNOLLY: Yes, he does. He sees them;
7 he doesn't treat them. He's not a physician.
8 DR. DOUMAS: Exactly. So why in the world
9 a recommendation was based on somebody who is not a
10 physician, not a clinician or even a physician? I mean,
11 this is really surprising to me.
12 MR. CONNOLLY: That's a question for the FDA.
13 DR. DOUMAS: The FDA recommended Dr. Naito
14 to you?
15 MR. CONNOLLY: They recommended that we
16 follow the NCEP guidelines, which were I think in much
17 authored by Dr. Naito, is that not so?
18 DR. DOUMAS: I don't think so. I don't think
19 the guidelines -- the NCEP guidelines really agree with
20 the statement that you have when to do the test.
21 The next question that I have is: expected
22 values -- a physician or health care professional will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
144
1 discuss target values specifically appropriate for each
2 patient. It reads, "If the manufacturer has available
3 a physician who will discuss the values with the patient,"
4 is this -- my assumption is that -- is my assumption
5 correct?
6 MS. ENRIGHT: That the manufacturer has a
7 health care professional?
8 DR. DOUMAS: Says a physician or health care
9 will discuss. I mean, I read it two or three times, and
10 my impression is that you have a physician available that
11 will discuss with the patient the values.
12 MS. ENRIGHT: No, that was not the intent of
13 that.
14 DR. DOUMAS: Then, I think this needs to be
15 changed, that the patient should consult with a physician.
16 MS. ENRIGHT: Okay. That was the previous
17 verbiage. That was used in a previous insert and very
18 similar to some that other manufacturers used. But we
19 will change that based on that and based on your
20 recommendations. We'd be happy to make those changes.
21 DR. DOUMAS: Thank you.
22 MS. ENRIGHT: Thank you.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
145
1 DR. DOUMAS: Dr. Manno?
2 DR. MANNO: I'm going to, first of all, say
3 the product insert -- correct me if I'm wrong, but you're
4 going to include two sets of directions, both the OTC set
5 of directions and the professional set for -- because the
6 strips are basically the same.
7 MS. ENRIGHT: That's correct.
8 DR. MANNO: Okay. Well, in that case, then,
9 you have a definition of fasting in the professional kit
10 as nine hours, and you have 12 hours in the OTC kit. If
11 people have a choice, they're going to take the path of
12 least resistance. That's one point.
13 MS. ENRIGHT: Yes, a very good point. Let
14 me address that and lay that to rest. You have -- that
15 is a draft of an older insert. This is only the OTC.
16 The professional was included for informational purposes
17 only. They will be consistent, but these have been on
18 different paths. Our professional triglycerides also has
19 been submitted for FDA clearance, and that insert needs
20 to be, once again, updated to be consistent.
21 DR. MANNO: If you have -- I'm just curious
22 why you just didn't make one insert specific for the OTC,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
146
1 so that people don't get confused in one through the
2 professional.
3 MS. ENRIGHT: Well, actually, we do have one
4 for the OTC and one for the professional. Because the
5 package includes both inserts, because it's the same
6 package that's sold to both markets, we do make, for
7 instance, the references. We don't feel like the consumer
8 is necessarily going to be interested in the references
9 that are included in the professional insert.
10 And if they are interested, they can look at
11 the professional insert because it's included in the
12 package. But --
13 DR. MANNO: Okay.
14 MS. ENRIGHT: -- if you have other
15 suggestions -- this was a difficulty for our glucose, which
16 is both over the counter as well as professional. We used
17 one insert. That was my preference, to use one insert,
18 but there were so many issues surrounding an OTC product
19 that we needed to go to a second insert. We also did that
20 for our cholesterol product for that reason.
21 DR. MANNO: Okay. Again, on instructions to
22 the user on this, there is just kind of a general statement,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
147
1 number 5, is, "Wait. Read results." Most of those
2 things -- most of the hand-held units that I have had
3 experience with make a general statement to the user that
4 this is the time when the reaction is good. Don't put
5 it on there and wait two days to read it.
6 It would seem to me that there should be a
7 reiteration of a time window that that reaction is
8 good -- to give a good result. And in that same vein,
9 when you get down to the end -- again, it's step 5 in the
10 thing -- they say to wait, read the results.
11 In the fine print it tells them to check the
12 strip after they have done the test to be sure that they've
13 gotten the correct drop size on, but yet it's not in the
14 five-step narration here with the pictures of what to do.
15
16 I think that it's a lot to ask of the consumer
17 to remember from the beginning what they should be doing
18 at the end in order to assure that the right drop of blood
19 is there, etcetera, and that they have a good test, in
20 spite of the slide that you have flashing up there.
21 MS. ENRIGHT: So you feel that's better
22 placed at the beginning?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
148
1 DR. MANNO: I think it should just be a short
2 statement and in the last steps here. Add a step. In
3 any laboratory procedure, any place, you've got QA issues.
4 And one of the things you look at is -- at the very end
5 of anything is how -- is everything correct?
6 MS. ENRIGHT: I'll just ask the FDA, perhaps
7 at the end when we have recommendations, I think the FDA
8 will probably summarize -- have the FDA summarize this
9 and come up with -- comes up with a consensus on
10 recommendations for changes in the labeling. And we will
11 gladly make appropriate changes, based on the FDA's input.
12 DR. MANNO: Okay. I would like to ask a
13 question about the statement of compliance. Seventy-five
14 percent of the people nine months later are not on there
15 anti-lipid drugs. And was this a study that was cited,
16 or was this anecdotal from the company reps for the meds?
17 And did they give you any reasons that they thought this
18 may -- why people were non-compliant? Did they bring up
19 things like side effects?
20 MR. CONNOLLY: No. I've not seen any
21 studies. I got this from two sources. One from two of
22 the manufacturers, one of the earlier statin providers
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
149
1 and one of the current ones. And also
2 from -- apparently -- I'm not aware of the mechanism, but
3 apparently there is a way that the pharmaceutical people
4 monitor prescription refillment and usage through the
5 pharmacies and the sales things with patients.
6 DR. MANNO: Well, that is -- yes, that is
7 biased, too, to the extent that people who are not on the
8 drug may be changed to another drug. But I'm concerned,
9 as a pharmacologist/toxicologist, that there may be
10 reasons why people are not taking it.
11 Again, as Ms. Kruger said, you have to be kind
12 of careful with that compliance claim.
13 DR. DEEG: Dr. Manno, just a quick comment.
14 There is actually literature supporting the compliance
15 rates on various lipidemic agents.
16 DR. MANNO: Okay.
17 DR. DEEG: So those numbers are recent.
18 DR. MANNO: And one --
19 CHAIRMAN NIPPER: Dr. Manno, I'm sorry to
20 interrupt. Pull to microphone a little closer to you.
21 DR. MANNO: Oh, I'm sorry.
22 CHAIRMAN NIPPER: There are some people out
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
150
1 in the hall that are watching you on a monitor and they
2 can't hear you.
3 DR. MANNO: Oh, okay. Didn't realize that.
4 Thank you.
5 I'd like to ask you if you've characterized
6 the reaction curve. As I understand it, this is a
7 multi-enzyme type coupled reactions. And you have
8 problems with total air. Have you really characterized
9 that so that you're in the right range, especially when
10 you've got such large errors, so that you know you are
11 in the right range on your measures? Did you just do that
12 zero to 400 or 500?
13 DR. PASQUA: We calibrate our curve
14 with -- against our reference method. We take blood
15 samples against our reference method in venous blood, and
16 then pipetted the strips to capture our raw data, which
17 is percent R. It's an end point, percent R.
18 And then from that, we fitted -- basically
19 fit a curve to a hyperbolic function, and then create a
20 table of percent R versus mils per deciliter, a 12-point
21 table, which the meter, once it gets a percent R reading,
22 will interpolate -- it will linear interpolate between
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
151
1 the 12 points.
2 So we --
3 DR. MANNO: So you've never actually seen how
4 far you go with this reaction, and get the shape of the
5 curve where the cutoffs -- where you're beginning to curve
6 and drop off.
7 DR. PASQUA: We have difficulty obtaining
8 high samples. But we do have -- we have gotten samples
9 above 400, so we know that at least it's 400 for sure.
10 DR. MANNO: Okay. You've answered my
11 question. Thank you.
12 And I want you to just clarify one point for
13 me, and I don't know if it will be you or one of the other
14 people, but you state that it's a whole blood procedure,
15 or test. And, of course, you've checked it against a serum
16 method.
17 Also, it's stated in here that there's a
18 difference in the method of red blood cell separation,
19 and I'm wondering that -- the statement is that you
20 centrifuge the blood through the predicate method to get
21 serum. Then you say that the test strip separates the
22 red blood cells, allowing developed color to read on the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
152
1 reactionary of the membrane.
2 So are you really reading a whole blood sample
3 reaction, or are you reading a serum?
4 MR. CONNOLLY: John said I could give my
5 version.
6 DR. MANNO: Which is the right version?
7 (Laughter.)
8 MR. CONNOLLY: I think it's difficult to
9 tell. It's mostly serum, predominantly serum, that's down
10 there, yes. Obviously, red blood cells don't contain the
11 things that we're measuring here, except in some of our
12 other tests. But for things like glucose, cholesterol,
13 ketones, triglycerides, it's mostly serum. And by "mostly
14 serum," if I had to guess, I would guess it's 95 percent
15 serum.
16 There is probably some lysis that occurs, and
17 there are probably -- red blood cells can do wonderful
18 things in tiny holes, so I'm sure there are some red blood
19 cells getting to --
20 DR. MANNO: Do you know the pore size of the
21 strip?
22 MR. CONNOLLY: Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
153
1 DR. MANNO: Are you at liberty just to make
2 a --
3 MR. CONNOLLY: I think they are -- it's .45
4 micron.
5 DR. MANNO: Okay. Thank you.
6 CHAIRMAN NIPPER: Thank you, Dr. Manno.
7 Dr. Everett?
8 DR. EVERETT: My questions deal largely with
9 whether or not tests, again, measure what you say it does.
10 And I'd like to start by clarifying again, what is the
11 working range of the test strips? Any representative can
12 answer.
13 DR. PASQUA: Right now, 30 to 500 mils per
14 deciliter.
15 DR. EVERETT: Do you know if the precision
16 and accuracy changes over that range? Or is it constant?
17 DR. PASQUA: Well, I'd say from the low end
18 to about 200 mils per deciliter we see a proportional error
19 in the random error -- in the precision. And then about
20 from 200 to 400, it seems to stay relatively constant -- the
21 precision.
22 CHAIRMAN NIPPER: Speak into the microphone.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
154
1 There you go.
2 DR. PASQUA: Okay. Do you want me to repeat
3 everything or --
4 CHAIRMAN NIPPER: No.
5 DR. PASQUA: Okay.
6 CHAIRMAN NIPPER: Just keep going.
7 DR. PASQUA: So, in other words, if you look
8 at the -- that's why the precision at the high end at 400
9 is pretty good. It's less than five percent, and it seems
10 that the SD stayed relatively constant at 200 to 400.
11 The bias -- if you use the linear regression
12 versus reference -- the predicate method, it seems to
13 be -- it gets much, much worse as you go higher. It gets
14 much more negative. But if you take the actual
15 points -- the bias -- around like 400 and 200 that I did
16 and average that bias, it's not as bad. I think what's
17 happening is if I -- we need to have more points at the
18 high end to really get a better idea what the curve is.
19
20 I think that high end isn't being weighted
21 property, and we're underestimating. I mean, we're -- and
22 that's why we might have the low bias, if I were to take
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
155
1 a guess.
2 DR. EVERETT: Okay. And the other thing is:
3 are there any contraindications to the use of the
4 instrument?
5 MR. CONNOLLY: I wouldn't swallow it. Not
6 that I know of.
7 DR. EVERETT: Okay. And I guess my last
8 question is -- this was one of the ones that some of the
9 members have already asked. And that is, in a sense, where
10 does the use of this instrument fit into looking at
11 triglyceride levels in patients as opposed to ordering
12 the traditional wet bench lab test?
13 If we still have the same recommendation of
14 nine- to 12-hour fasting before we do it -- I'm still not
15 sure, where does it fit into managing patients?
16 MR. CONNOLLY: I'd like to answer the first
17 part of that, and Dr. Deeg the second part of that. I
18 have to remember the first part now that I got that
19 statement out.
20 (Laughter.)
21 Diabetes as a disease, compared to other
22 diseases, I think is probably -- categorically, you could
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
156
1 say it's the only disease that's truly managed and treated
2 at home. Other diseases -- cardiovascular disease,
3 cancer, whatever -- where you're in some kind of a clinic
4 or hospital for therapies, or radiation, or chemo, or
5 whatever, but diabetes is something that is truly managed
6 at home. Number one.
7 And I think most of this information that can
8 be used at home by the patient -- be it triglycerides,
9 cholesterol, any of the lipids -- that he can work with
10 himself, and his professional is key.
11 Hang on while I get the rest of this now.
12 How long do I have to answer this, Dr. Nipper?
13 CHAIRMAN NIPPER: Be as brief and concise as
14 possible, sir.
15 MR. CONNOLLY: You can follow up on the
16 compliance and --
17 DR. DEEG: I think from my perspective, in
18 terms of treating these patients, one, for me, it provides
19 additional data points that -- to see how they're doing
20 over the interim between when I'm seeing them in the office
21 sort of thing, and then the issues that have been raised
22 before in terms of their providing feedback, positive or
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
157
1 negative, to the patient in terms of how they're doing
2 in complying.
3 MR. CONNOLLY: Once the patient has gone on
4 a blood draw -- I mean, if the patient comes in, sees the
5 doc, they draw a blood sample, send it to the whatever
6 reference lab, which is common, and the results are back
7 in two or three days, or whatever it is. You know, the
8 patient is gone, and the opportunity of the doctor with
9 the patient is gone. And I think that's a real missed
10 opportunity in working with a patient for the physician.
11 DR. EVERETT: Well, in a real sense, that's
12 not quite accurate either. Many times we've planned to
13 do the test before the patient arrives. The patient comes
14 in a day or two early, got the test done, so when the patient
15 comes in for the office visit the lab results are there
16 as well.
17 MR. CONNOLLY: Right.
18 DR. EVERETT: And that's my concern about how
19 you describe how this test is going to fit into the other
20 things that doctors do and how we counsel the patients
21 and how we deal with them. This has just gone awry with
22 these misinterpretations, or bold gestures of what we do.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
158
1 And that's not quite accurate.
2 But, again, that's a subject for a different
3 time. My other question is this test and its
4 performance -- has it been measured against many of the
5 drugs, looking for things that might interfere with the
6 test as it relates to drugs and medications that many of
7 the hyperlipidemic patients are on?
8 MR. CONNOLLY: Yes. But back to your other
9 question here, to ask the patient to come in the office
10 twice for a visit so it's convenient for the doctor, that's
11 not convenient for the patient. And the patient may not
12 wish to do either one of those or find another physician
13 that goes the other route where he only has to make one
14 visit.
15 Someone else will have to answer interference
16 issues.
17 Are you asking for more information that
18 what's given in the table by Dr. Anaokar for the
19 interferences we look at?
20 DR. EVERETT: Yes.
21 MR. CONNOLLY: No, that's -- we showed you
22 everything we have.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
159
1 DR. EVERETT: But, so if there are various
2 drugs that will interfere with the ability of the
3 instrument to measure triglyceride levels, that has not
4 been done, is that correct?
5 MR. CONNOLLY: I don't know which drugs
6 you're referring to. Can you help me here?
7 DR. EVERETT: Assume the patients, as you
8 already mentioned, are on many of the anti-lipid-lowering
9 medications. Some are on hypertension medications.
10 Diabetics tend to be on lots of drugs. They are to lower
11 their lipid levels, their triglyceride levels, their
12 glucose levels.
13 Have we looked at whether or not any of those
14 will interfere with the test?
15 DR. PASQUA: No. No, we haven't, but I don't
16 think you can -- we to do that, but I thought you meant
17 something like presumed that it will.
18 DR. EVERETT: I don't know. That's why I'm
19 asking.
20 DR. PASQUA: No, I don't --
21 DR. EVERETT: Because these patients are
22 typically on many medications.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
160
1 MS. ENRIGHT: Maybe I can address that. One
2 thing that we did not do with our data, that we did not
3 show you, is for each and every patient, the 382, we have
4 data that shows what medications they're on. And if you'd
5 like, we can sort that data, and we find those patients
6 that are on lipid-lowering drugs. And we can do those
7 calculations. We do -- we have -- we do have that data
8 available.
9 DR. EVERETT: Okay.
10 MS. ENRIGHT: So we can calculate that for
11 you.
12 DR. EVERETT: Well, that's up to Dr. --
13 MR. CONNOLLY: Dr. Everett, I think this
14 reaction principle that we use is pretty standard. It's
15 been around since probably the late '70s, GPO
16 triglycerides, so I think anything that interferes with
17 anyone else using this reaction system we'd have the same
18 thing. I wouldn't expect any difference, since it's the
19 same reaction principle.
20 DR. EVERETT: Sounds reasonable. But if you
21 never look, how would you know?
22 MR. CONNOLLY: Well, if that's a question,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
161
1 I think we need to go back and look at the predicate devices
2 and see if those people looked at the same interferences.
3 I mean, we'd be happy to do the study to see if there's
4 interference there, if the panel feels that it warrants
5 it. But I'm not sure as to -- is anyone here aware of
6 the predicate device we used, that interferences were done
7 in that -- on those products?
8 DR. DOUMAS: May I answer this?
9 CHAIRMAN NIPPER: Dr. Doumas? And then Dr.
10 Rosenbloom had a follow up.
11 DR. DOUMAS: I don't think you can compare
12 systems where the sample serum is diluted 40 or 50 times
13 with the system that you have undiluted serum. I mean,
14 there is another instrument on the market, a major
15 instrument which sometimes that interference is there
16 because there is no dilution of the sample whatsoever.
17 But in most systems, when the sample is
18 diluted 10, 20, 30, 40 times, before the reaction has been
19 run, then you get a different interference. You may have
20 no interference when the sample is diluted, and
21 interference when you have no dilution of the sample.
22 I'm not saying that those things will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
162
1 interfere, because I have no idea if they will. But there
2 is a difference how you do interference studies. In tests
3 that you have, the sample, as such, is being analyzed
4 without dilution, without the serum, and when the sample
5 is diluted quite a bit when the reaction is being performed.
6 MR. CONNOLLY: I'm pretty certain that the
7 amount of sample to substrate ratios in the chemical
8 reaction would be the same in this product. I can't say
9 this categorically, but I'll bet it's within very close
10 to the same thing you see in large analyzers. So there
11 is no dilution effect of sample to substrate. It's just
12 in a dry phase by which there is less water. But as to
13 actual reactants, they would be the same.
14 DR. EVERETT: My last question deals with the
15 target audience for use in your instrument. Is it confined
16 to just diabetics, or any patient that we would like to
17 know what their triglyceride levels are?
18 MR. CONNOLLY: We'd like it to be for all
19 patients. We are focused on diabetics, but there are a
20 lot of patients we think this would be of use to.
21 CHAIRMAN NIPPER: You had a followup
22 question, Dr. Rosenbloom.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
163
1 DR. ROSENBLOOM: Yes, it was a follow up.
2 I share Dr. Everett's concern about the lack of actual
3 testing in patients who are ingesting the materials that
4 might interfere. But also, the package insert, in its
5 list of limitations, does not indicate that certain drugs
6 or vitamins will be -- would interfere, although in other
7 correspondence the sponsors say that it does. And I
8 wondered why this was left out of the package insert, or
9 if this is just an error.
10 DR. MANNO: It's in the -- excuse me --
11 CHAIRMAN NIPPER: Dr. Manno?
12 DR. MANNO: It's in the -- listed on the
13 professional insert that will also be included.
14 DR. ROSENBLOOM: But it's not in the --
15 DR. MANNO: It's not in the --
16 DR. ROSEBLOOM: -- not in the patient insert?
17 DR. MANNO: -- in the OTC insert.
18 MS. ENRIGHT: Yes, that's correct, because
19 that was deemed more important for the professional. It
20 wasn't included in the OTC. And, again, if that's a
21 recommendation, we include it in the OTC.
22 One of the requirements for the OTC package
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
164
1 insert is that it be written at a seventh grade level,
2 and some of that information, you know, based on the
3 guidelines that we were given does exceed what was
4 recommended for the OTC insert.
5 But, you know, again, based on the consensus
6 and what the FDA recommends, we'll make whatever changes
7 you find appropriate.
8 CHAIRMAN NIPPER: Can you make it brief?
9 DR. MANNO: Yes. The three that you chose
10 to list on the professional as not interfering are
11 acetaminophen, ibuprofen, and solicolites. And many of
12 the people -- your target population will be on those drugs.
13 So that would be a natural question for a consumer who
14 is on that drug, one of those drugs, to ask.
15 I think it would be nice if they had that
16 information easily available to them.
17 MS. ENRIGHT: I think that's a reasonable
18 suggestion, and we'll look to the FDA reviewer if they
19 feel that's appropriate in OTC. But we have no issues
20 with including that.
21 CHAIRMAN NIPPER: Thank you.
22 Dr. Everett, did you have other questions?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
165
1 DR. EVERETT: No, I'm done.
2 CHAIRMAN NIPPER: Dr. Rej?
3 DR. REJ: Thank you. I feel a little bit like
4 my favorite Dilbert cartoon, where the hapless engineer
5 who is last around the table --
6 (Laughter.)
7 -- says, "All of the good comments have been
8 taken already."
9 (Laughter.)
10 But I think I have a few questions I think
11 will be useful in keeping in mind before we hear the FDA
12 presentation. It has to do with comparison of the
13 BioScanner results with the reference method. I don't
14 know who is best to answer that question.
15 But in the design of the study, you mention
16 that consumers measured their own triglyceride. There
17 was a professional measurement of triglyceride by the
18 proposed device. And then a venous sample was taken and
19 assayed by the reference method.
20 MS. ENRIGHT: That's correct.
21 DR. REJ: Okay. And the data that's in the
22 package that we received from the FDA, which results from
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
166
1 your device are shown?
2 MS. ENRIGHT: Absolutely the consumer. This
3 was a consumer over-the-counter submission.
4 DR. REJ: Okay.
5 MS. ENRIGHT: The professional is included
6 in a separate submission, and we were asked to submit that
7 separately. So it will be addressed as a separate issue.
8 DR. REJ: Do you have data showing the
9 comparison of your device in the hands of the consumer
10 and a professional that you can share with us today?
11 MS. ENRIGHT: I'll have to see if I can pull
12 that out for you. I brought a large amount of -- a large
13 amount of data, and I have some data. I'll have to see
14 what I can pull out for you today.
15 DR. REJ: Okay.
16 MS. ENRIGHT: We were asked specifically to
17 leave that out of the OTC so -- for your package, and we
18 do have --
19 DR. REJ: Okay. So this is the consumer
20 versus the --
21 MS. ENRIGHT: That is the consumer --
22 DR. REJ: -- the -- and looking at the graph
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
167
1 that you showed of the chemistry used, I assume that your
2 device measures glycerol as well as triglyceride?
3 MS. ENRIGHT: Yes, it's not glycerol blanked.
4 DR. REJ: Okay.
5 MS. ENRIGHT: That's correct.
6 DR. REJ: And your comparative method within
7 your lab, does that include glycerol measurement?
8 MS. ENRIGHT: Yes, it does.
9 DR. REJ: Okay. And so everybody is reading
10 from the same book, does the reference method measure
11 glycerol?
12 MS. ENRIGHT: The CRMLN is a glycerol blanked
13 method, the CRMLN method is.
14 DR. REJ: Okay. So that the bias -- okay.
15 The bias, then, is -- seems to be, then, even greater
16 if you had compared it to a non-glycerol blanked reference
17 method or comparative method, correct?
18 MR. CONNOLLY: I think it's the other way.
19 But no.
20 CHAIRMAN NIPPER: Can we let Dr. Kimberly
21 jump in, if she needs to, here?
22 DR. KIMBERLY: I was looking at the data the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
168
1 other day, and PBRF reported by total and net. And what
2 they used in their consumer study was versus the total
3 and not -- so theirs is a total net -- total triglycerides.
4 DR. REJ: Okay. So the true triglyceride was
5 not used as a comparative?
6 DR. KIMBERLY: Right.
7 DR. REJ: The bias would be even greater had
8 those data been used.
9 DR. KIMBERLY: Yes.
10 CHAIRMAN NIPPER: Thank you.
11 DR. REJ: And one -- I know that you made
12 reference in your presentation that you sought guidance
13 from the FDA regarding comments on the OTC package insert.
14 But I'd like to know -- and you asked the FDA how they
15 would like it calculated, but I'd like to know how you
16 calculated the statement that the consumer layperson, in
17 their own home, with accurate results 95 percent of the
18 time --
19 MS. ENRIGHT: Yes.
20 DR. REJ: -- was calculated.
21 MS. ENRIGHT: Yes, I believe that was based
22 on a 95 percent confidence interval. I'd have to get
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
169
1 those --
2 DR. REJ: No, that says 95 percent of the
3 time.
4 MS. ENRIGHT: Yes. I think I'll have to
5 consult with Dr. Pasqua. And if I can get you that answer
6 later, and let me make sure how we did that calculation,
7 I'll get you that information.
8 CHAIRMAN NIPPER: Thank you, Dr. Rej.
9 In the interest of time, I'd like the panel
10 to take a -- can we take a 10-minute break? And then we'll
11 hear from the FDA when we have less biological urgency.
12 Thank you everyone for bearing with us while
13 we got through this.
14 (Whereupon, the proceedings in the foregoing
15 matter went off the record at 12:06 p.m. and
16 went back on the record at 12:22 p.m.)
17 CHAIRPERSON NIPPER: I would like to
18 introduce a person who needs no introduction, Arleen
19 Pinkos.
20 MS. PINKOS: I don't know about that.
21 CHAIRPERSON NIPPER: You've already been
22 here this morning, remember?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
170
1 MS. PINKOS: As Dr. Nipper so humbly
2 introduced me, I'm Arleen Pinkos. I am one of the
3 reviewers in the Chemistry and Toxicology Branch and the
4 primary reviewer for this product. During the next hour,
5 we are going to have a number of speakers for you.
6 Unfortunately, because we don't share all the slides back
7 and forth before the presentation, there is a little bit
8 of overlap. We'll try not to bore you too much.
9 Before I get into what we are going to be
10 talking about, I was asked to make a clarification on the
11 cholesterol product that was brought up earlier. We did
12 do a little bit of research. Back in 1995, we did have
13 a cholesterol product that came through for review for
14 over-the-counter. At that time, NCEP did not outline
15 goals for total error. They did, however, outline goals
16 for bias. That device did meet it, but apparently since
17 then, the total error that NCEP has identified is beyond
18 what that product was for. It was okay when it prossed,
19 passed by us, is all I can say at this point.
20 In the first 10 minutes, I am going to cover
21 two main topics, a summary of the special considerations
22 given to over-the-counter products, followed by an
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
171
1 overview of some of the performance studies. Next, Carol
2 Benson, also a reviewer in the Chemistry Branch, will
3 present a summary of total error calculations. Total
4 error, as you all know, is identified by NCEP as an
5 important measure of device performance.
6 Then, Dr. Irony, our statistician, will
7 summarize some observations from her review. She will
8 be followed by Dr. Ginsberg, who is a lipid expert. He
9 will be giving us a little bit of perspective about
10 over-the-counter triglyceride products and what's going
11 on there.
12 Then finally, I'll return back to the podium
13 to review some of the questions that we would like you
14 to cover during your deliberations.
15 I would also just like to mention that Dr.
16 Kimberly here has very nicely agreed that if we would like
17 her to, once we move into open discussion, if we would
18 like her to give her overview of CDC's lipid
19 standardization program, she is willing to do that.
20 Okay. When reviewing products, the division
21 considers whether the device will be used professionally
22 or whether it will be used by lay users. The minimum
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
172
1 performance thresholds for granting marketing clearance
2 are different for these two devices because the people
3 running the tests are different, and the environments
4 where they will be used are different.
5 I will first describe the process by which
6 we clear a professional product, then go over some of the
7 special considerations given to over-the-counter
8 products. During the review of a 510 case submission,
9 FDA determines whether or not the device is substantially
10 equivalent to the predicate device. The term
11 "substantially equivalent" is not defined by law, and the
12 Division frequently grapples with determining the
13 appropriate standards by which to judge equivalence. This
14 question is particularly problematic when there is no gold
15 standard by which to evaluate the device.
16 In general, the Division makes its
17 determination through analysis of the fundamental
18 performance parameters of the device, such as accuracy
19 and precision. The device is determined to be equivalent
20 if the performance characteristics are roughly the same
21 as those of the predicate device. When deciding whether
22 close is close enough, the Division frequently allows more
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
173
1 error in a device when it's going to be used by
2 professionals.
3 We do this because professionals are trained
4 to perform instrument procedures, including maintenance
5 and QC. They know how to interpret the results. But
6 perhaps more importantly, because they understand the
7 limitations of the test result and because limitations
8 can be adequately conveyed in professional labeling.
9 Another important consideration is that the
10 test result is interpreted by a clinician, along with other
11 laboratory tests and the patient's clinical signs and
12 symptoms.
13 When the Division reviews an OTC product,
14 performance characteristics are still the primary focus
15 of equivalence of the equivalence determination.
16 However, because a device will be used by lay users,
17 additional considerations must be part of that decision.
18 There are three primary areas of
19 consideration. First, the device should produce
20 clinically meaningful results when operated by the lay
21 user. Second, the results should be interpretable by a
22 lay user, consumers should be able to understand what the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
174
1 results mean, and what to do with them. Third, the
2 potential benefit should outweigh the possible risks.
3 I would like to spend a few minutes on each
4 of these three points. The lay user must be able to
5 generate clinically meaningful results using the device.
6 Those results should be comparable to those obtained by
7 professionals. Results must be reliable enough to provide
8 clinical benefit. As with professional use devices, we
9 are again faced with the question of what is the appropriate
10 minimum performance threshold. How accurate must a test
11 be?
12 In the case of this particular product, we
13 are lucky enough to have a reference point. NCEP has
14 identified performance goals for us, and we'll be talking
15 about those in more detail in a few minutes. To assess
16 whether results are clinically meaningful, FDA suggests
17 that accuracy and precision studies be performed by lay
18 users.
19 Lay users should be representative of the
20 population likely to purchase the device, and have a wide
21 variety of socio-economic and educational backgrounds.
22 FDA prefers that accuracy be measured against an
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
175
1 established reference methodology if one is available.
2 These studies should be designed so as to
3 mimic the setting in which the device will be used. For
4 example, the consumer should only have access to the
5 package insert for instructions. The sample collection
6 procedure should be the same as what will be done by the
7 consumer. A finger sample should be evaluated in the study
8 if that is the sample type that will be used by the consumer.
9 This is important because the Agency has observed that
10 whole blood samples generally don't perform as well as
11 serum or control materials.
12 Why is it so important to have the proper
13 studies performed? In short, because consumers play a
14 significant role in the quality of results that are
15 generated. Those results can vary significantly from the
16 results generated by professionals. There are many
17 factors that influence that end result.
18 Consumers receive no training and have no
19 oversight. They are not as experienced in performing and
20 interpreting tests. They generally do not understand the
21 importance of QC practices, using proper technique or the
22 need to follow directions exactly. In most cases,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
176
1 consumers do not understand the importance of proper
2 reagent storage or performing maintenance.
3 Evaluation of labeling is very important in
4 a consumer study. The consumer must be able to read the
5 package insert and know exactly what functions to perform
6 in order to maintain the system and run the test. Labeling
7 should clearly tell the consumer what to do when they get
8 certain results. A good consumer study will evaluate the
9 effects of all these factors.
10 Test results must be easy for a lay user to
11 interpret. The reliability of an over-the-counter
12 product is very important. Consumers are making a medical
13 decision to act or not act based on a single test result.
14 Unlike their professional counterparts, lay users do not
15 have the benefit of other tests or medical history upon
16 which to base their decision. Interpretation of results
17 can also be complicated by the fact that a test may have
18 limitations that are difficult for a lay user to
19 understand, such as bias or imprecision, or as we discussed
20 earlier, interfering substances.
21 The importance of labeling can not be
22 over-emphasized. The package insert is the only source
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
177
1 of information for the consumer concerning how to perform
2 the test, run and evaluate QC, properly maintain the
3 instrument, and understand what to do with the test result.
4 The Agency recommends a seventh grade reading level for
5 over-the-counter products.
6 Finally, our third consideration is whether
7 the benefits of the device outweigh the risks. This
8 question is a subjective one, but there are a few key
9 elements that are worth reviewing. What are the benefits
10 of having this test at home? Does the device generate
11 clinically-acceptable results? It may not be necessary
12 for a device to be perfect in order to provide benefit
13 to the user.
14 How likely is it for the consumer to get an
15 incorrect result? Are the test procedures simple, without
16 little likelihood of error? Finally, if a consumer does
17 get a wrong result, what will happen? Will they not go
18 to their physician or might they adjust their medication
19 on their own?
20 In conclusion, there are three considerations
21 that must be given to over-the-counter products. Can a
22 lay user reliably and consistently generate a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
178
1 clinically-meaningful result? Did the consumer accuracy
2 and precision studies mimic actual use? Is the result
3 interpretable? Does the labeling capture the meaning of
4 the result, and what to do when certain results are
5 obtained. Finally, do the benefits outweigh the risks?
6 Now I would like to move on to the second half
7 of my presentation, which is a brief overview of the key
8 performance studies that have been done. That's not to
9 say that there aren't other studies in here. I think there
10 were linearity studies that we have already discussed.
11 One of the questions that you asked this
12 morning from the nurse's perspective is one that I might
13 just take a minute on, because there was some discussion
14 about it. That was the short sample problem and the size
15 of the drop. That was certainly something that we're
16 always concerned with, with whole blood assays.
17 The sponsor's address to that is to visually
18 put a drop size or the circle size that's adequate. You
19 can certainly provide some input on that. But they also
20 did a short sample study where they evaluated a sample
21 and titered it down by adding smaller and smaller and
22 smaller drops, to see if the short sample detection system
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
179
1 was adequate.
2 What was found was that the short sample
3 mechanism was not as accurate as you would like it to be.
4 It was actually possible to get a result that was half
5 of what the true value was before the detection system
6 kicked in. That was one of the reasons why we had them
7 add as a final step to turn a strip over and visually look
8 at it to see if it looked like there was a short sample.
9 So whether that's an adequate mechanism for
10 over-the-counter products, you can certainly weigh in on
11 that.
12 Before I move into the performance studies,
13 I just wanted to present the performance goals set by NCEP
14 for triglycerides. Much of our interactions with the
15 sponsor have focused on gathering information to
16 characterize BioScanner's performance in this area. I
17 would like to mention that NCEP recommends that accuracy
18 of triglycerides be characterized against CDC's reference
19 method. Although a direct method comparison to that
20 method is not possible, CDC does have a list of reference
21 laboratories that are standardized against the reference
22 method.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
180
1 Okay. I would like to now present a brief
2 overview of the sponsor's three accuracy studies, along
3 with the resulting regression data. The first study
4 compares BioScanner's results from lay users to results
5 from serum samples run by the predicate device. The
6 predicate device, as we have discussed earlier, was the
7 Sigma reagent system run on an autoanalyzer using serum
8 samples. This study involved 381 samples.
9 The second study compares results from 47
10 serum samples analyzed by the Sigma method to results from
11 a laboratory standardized against the CDC reference
12 method, Pacific Biometrics.
13 Then the third study involved results from
14 fingerstick samples analyzed by lay users on the
15 Bioscanner compared to results from venus samples analyzed
16 by the same CDC standardized lab. That study involved
17 220.
18 I know we have been talking about a lot of
19 different accuracy studies going back and forth, so
20 hopefully that will put a little bit of perspective on
21 that for you.
22 There have been three different types of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
181
1 precision studies performed. The sponsor's precision
2 claim is a C.V. of five percent, or within 20 milligrams
3 per deci-liter, whichever is greater. NCEP's
4 recommendation, you recall, was a precision of less than
5 or equal to five percent.
6 I am not going to dwell on any of the charts
7 that I am going to show you because they are in your packets.
8 Again, I only want to provide them as a reference point
9 for you.
10 The first study involves the only study
11 evaluating whole blood. The analysis was performed by
12 three lay users on two levels of EDTA collected samples.
13 This study was of particular interest because the whole
14 blood sample typically demonstrates poorer precision than
15 serum-based materials. The within run data is here before
16 you, separated into three tables, with each table
17 representing one consumer. The only thing that I wanted
18 to point out to you on this table was that the C.V.s range
19 from 1.7 to 8.8 percent.
20 The second study involves the analysis of
21 serum-based control materials by 20 consumers at three
22 sites. The three levels of control were run on the same
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
182
1 day. The only thing I want to bring out to your attention
2 is that the single point estimates for C.V. range from
3 4.2 to 10 percent.
4 The third precision study involved analysis
5 of control materials by consumers. I have to apologize
6 because there was a little bit of miscommunication about
7 who ran this study. I think the packets that were sent
8 out to you indicated that this study was run by
9 professionals, as my slide does. So can make those
10 corrections. But it is indeed control materials run by
11 consumers.
12 This study is the only study where we had total
13 imprecision results generated. Again, the only thing I
14 wanted to point out were the single points of estimates
15 for C.V.s ranged from 3.2 to 10.6 percent.
16 The third and final area that I wanted to touch
17 on were the interference studies. The sponsor identified
18 two critical substances that interfere with triglyceride
19 measurements, ascorbic acid and elevated hematocrit
20 levels. The sponsor has also taken a rather unique
21 approach to characterizing the interference from
22 cholesterol. We would like you to address that during
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
183
1 your deliberations.
2 Any questions? Okay, then we'll have Carol
3 Benson come up.
4 CHAIRPERSON NIPPER: We'll hold questions
5 for Arleen and Carol and the other presenters until the
6 end. We may lay over until after lunch with a little bit.
7 After we assuage Dr. Rej's feelings of being the last
8 one, we'll start with him.
9 MS. BENSON: I'm Carol Benson. I will be
10 presenting the estimate for total analytical error for
11 the BioScanner triglyceride device. At this time, the
12 cholesterol reference method laboratory network,
13 otherwise known as the CRMLN, has no formal program in
14 place to provide a certification of traceability for
15 manufacturers of triglyceride devices like there is in
16 place for total, HDL, and LDL cholesterol. Therefore,
17 it is necessary to calculate the estimate of total
18 analytical error based upon the information supplied by
19 the manufacturer. The total analytical error takes into
20 account the inaccuracy and the imprecision of the device.
21 The steps used to estimate the total
22 analytical error are as follows. Calculate the percent
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
184
1 bias at the NCEP cutpoints using the percent. The equation
2 percent bias equals the test value minus the reference
3 value divided by the reference times 100.
4 Use the coefficient of variation from the
5 precision study. Calculate the total error using the
6 equation, total error equals 1.96 times the total CV plus
7 the percent bias.
8 First, we calculate the bias at the NCEP
9 cutpoints using the regression equation from the
10 comparison study of the 220 samples. The results are as
11 follow. At 100 milligrams per deciliter, the bias is minus
12 eight percent. At 200 milligrams per deciliter, the bias
13 is -11.5 percent. At 400 milligrams per deciliter, the
14 bias is -13 percent.
15 Next, we calculate the estimate of total error
16 using the C.V. from the different precision studies that
17 were submitted by the manufacturer. The results are
18 presented in different scenarios as follows. Scenario
19 number one is an estimate of total error by using the C.V.
20 from the precision study using the control materials.
21 As Arleen pointed out, that's incorrect about the
22 professionals. It was actually performed by the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
185
1 consumers.
2 The results are as follows: at 100 milligrams
3 per deciliter, the total error is about 28.8 percent.
4 At 200 milligrams per deciliter, the total error is 26.8
5 percent. At 400 milligrams per deciliter, the total error
6 is 22.1 percent.
7 The next two scenarios as estimates of the
8 total error using the within run C.V. from the precision
9 study in which the consumers tested the whole blood at
10 two concentrations of triglyceride, 189 milligrams per
11 deciliter and 423 milligrams per deciliter. These values
12 are near the NCEP cutpoints.
13 The results are presented for the two
14 consumers who obtained the lowest within run CV and the
15 highest within run CV. Scenario number two is the estimate
16 for the total error for the consumer with the lowest within
17 run CV. At approximately 200 milligrams per deciliter,
18 the total error would be estimated to be 14.8 percent.
19 At approximately 400 milligrams per deciliter, the total
20 error would be estimated to be 24.6 percent.
21 Scenario number three is an estimate of the
22 total error for the consumer number one, with the highest
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
186
1 within run CV. At approximately 200 milligrams per
2 deciliter, the total error is estimated to be 26.8 percent.
3 At 400 milligrams per deciliter, the total error is
4 estimated to be 30.5 percent.
5 In summary, we can say that the total
6 analytical error estimates using the total CV from the
7 precision study using control materials ranged from 22.1
8 percent to 28.8 percent. The total error estimates using
9 the lowest and the highest within run CV using whole blood
10 performed by the consumers ranged from 14.8 percent to
11 24.6 percent for consumer number three, and from 26.8
12 percent to 30.5 percent for consumer number one.
13 CHAIRPERSON NIPPER: Thank you.
14 DR. IRONY: Hello. Can you hear me? I need
15 to use the computer, so I will speak from here, if it's
16 okay with everybody.
17 I am a statistician that's in charge of
18 reviewing this device. I will illustrate a little bit
19 what was talked up to here so we can have a deeper
20 understanding of what was going on.
21 For the precision studies, especially in the
22 consumer study, we had only three consumers. Each
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
187
1 consumer used two levels of whole blood. So these were
2 basically the results. As we can see here, there is a
3 huge variation because we are using only three consumers.
4 If we were using 10 consumers, we would also encounter
5 variation, but we'll be representing better the population
6 that will use this device.
7 For instance, if you look at the consumer two
8 and consumer three, they are measuring approximately the
9 same level, level 204 and 189. We can say okay, these
10 levels are close to 200 milligrams per deciliter. But
11 for one of the consumers, the standard deviation was three,
12 for the other one was 13.
13 What this tells us is that there is a high
14 variability and the point estimates are very crude. We
15 shouldn't use point estimates. The best way to do it is
16 to consider interval estimates.
17 So what we have here is the computation of
18 the interval estimates here. This tells us we have more
19 chance to get the right values when you use an interval
20 rather than using a point estimate. But still, if you
21 look at consumer two and consumer three for the levels
22 of triglycerides that are close enough, we have two
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
188
1 confidence intervals that don't even overlap. So again,
2 this reflects the high variability on these measurements
3 on only three consumers.
4 What will be the interpretation of this random
5 error? The interpretation is to follow. If we really knew
6 the standard deviation, if we had true deviation there,
7 95 percent of measurements for the consumer will be within
8 1.96 standard deviations from the mean. That will be true
9 only if we knew the standard deviation. As I told you
10 before, there is a lot of uncertainty about the estimated
11 standard deviations from these particular studies.
12 So let's see what will happen if we will get
13 the upper levels of this confidence intervals. Let's say
14 consumer number two, for instance, mean triglyceride
15 measured was 204. The upper level of the standard
16 deviation was 38. What that means is we will expect that
17 95 percent of the measurements taken by consumer two will
18 be below 240 milligrams per deciliter, when in fact, it's
19 mean was 204. This is for only one consumer. This may
20 vary a lot because the population of consumers will vary.
21 So we have, as I said, a small sample. This
22 was a random error taken in a short period of time. We
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
189
1 had no data to estimate a day-to-day or a long-run
2 variation. That's it for the precision studies.
3 Now I would like to talk a little bit about
4 the comparison of method study. We had 220 patients.
5 We compared the measurements taken by the BioScanner
6 versus the CRMLN lab reference.
7 So it was performed a regression analysis,
8 and again, I like confidence intervals rather than point
9 estimates. So the point estimate for the slope was .85,
10 but at 95 percent confidence interval, was between .81
11 and .90.
12 The intercept, the point estimate for it was
13 in this particular study, 7.2, but the confidence interval
14 was between -2.7 and 17.1. In this case, the confidence
15 interval for the intercept does encompass zero. The slope
16 does not encompass one. Just to remind you, a perfect
17 or ideal linear fit will be when the slope was one and
18 when the intercept was zero. That will be the ideal.
19 It never happens this way.
20 So the linear regression equation is given
21 by the this equation that I am writing on the bottom, but
22 that's again, this is the point estimate. So for us to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
190
1 take a look and see how the regression was behaving in
2 this particular case, we can look at this. This is the
3 regression line, and this, the red line is the regression
4 line, and the points are actually the measurements that
5 were performed.
6 As we can see, there were like a few points
7 that will deserve further investigation. They are far
8 from the line. What we performed is also an analysis of
9 residuals. I will explain what a residual is.
10 A residual in this case is the difference
11 between the actual measurement from the BioScanner device,
12 and the predicted measurement by the regression line.
13 So an ideal residual will be zero. In other words, what
14 you are estimating is actually the right -- what you are
15 measuring. This is zero. So ideally, the residuals will
16 be distributed evenly around this line. As we can see,
17 we have more residuals that are negative, but the positive
18 residuals are larger. You have like three points, the
19 observation number 140 and 77 and 120 there, particularly
20 far from this line. The only thing we can say is that
21 image deserved for the investigation. That is the
22 distribution of the residuals.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
191
1 The median residual was -5. That means 50
2 percent of the residuals were below -5. The maximum
3 residual was that one that's shown, the observational
4 number 140, and was 108. That means for that particular
5 case, the measurement given by the device over estimated
6 the regression line by 108.
7 Now as I told you, I like confidence
8 intervals. I don't like point estimates. So what we did,
9 we estimated the average systematic error or also called
10 bias, for the three decision levels, which are 100
11 milligrams per deciliter, 200 milligrams per deciliter,
12 and 400 milligrams per deciliter. In all cases, what's
13 happening is an under-estimation of the values. For
14 instance, for the case 400 milligrams per deciliter, what
15 was measured by the bias was actually 349, but the
16 confidence interval would be between 339 and 359. That
17 means that the estimated bias in this particular case will
18 be -51. If we build confidence interval for the bias,
19 it's going to be between -51 and -41. That's of particular
20 point of interest because it's a decision point. It's
21 a point when decisions are made.
22 Now it is very important to mention that the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
192
1 regression method assumes no error in the reference
2 method. In other words, we are assuming that the CRMLN
3 lab is a measurement without error. In case there is error
4 in that measurement, the slope could be attenuated, in
5 other words, will be maybe slightly smaller than what
6 should be, using the regression method.
7 As I mentioned before, an ideal linear fit
8 will be intercept zero and slope one. In this case, the
9 residuals are not evenly distributed, about zero, we have
10 more negative residuals than positive ones. Our three
11 positive residuals that deserve further investigation.
12 As was mentioned before in this meeting, we
13 have lack of observations about the level 400 milligrams
14 per deciliter, which is a very important level because
15 it's a decision level.
16 Now I used another method to estimate the
17 bias, which is called the Bland Altman method. What is
18 done, what it does is the following. We plot the
19 differences between the measurements done by the
20 BioScanner and by the reference, which is the CRMLN lab.
21 We plot these differences against the mean, the mean
22 between the BioScanner and CRMLN Lab, because in fact,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
193
1 we don't know what is the actual level of triglyceride.
2 So we use the mean between both.
3 The redlines here, the red bars, they indicate
4 intervals within 15 percent of the mean. So you can see
5 the points here. Again, you can see from this graph that
6 we would need more points around the level 400 milligrams
7 per deciliter to evaluate the device in a better way.
8 Now from these comparison of methods, we can
9 get the mean difference between the reference method and
10 the BioScanner. The mean is -21 milligrams per deciliter.
11 The standard deviation is 33 milligrams per deciliter.
12 The median difference is also negative, it's -25. The
13 maximum positive difference is +95. The maximum negative
14 difference is -92 milligrams per deciliter.
15 The distribution of the differences. This
16 is a histogram. You can tell from the histogram then you
17 have more negative differences than positive differences.
18 In fact, we have 75 percent of the differences are
19 negative, and 25 percent are positive.
20 That concludes my presentation.
21 CHAIRPERSON NIPPER: By that last statement,
22 you don't mean that there's 75 percent of the observations,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
194
1 there's no difference? You mean 75 percent of the
2 differences show a negative bias. Right?
3 DR. IRONY: Right. That's what I meant.
4 CHAIRPERSON NIPPER: Thank you.
5 Holding questions, leaving some time for our
6 next speaker, we are delighted to have Dr. Henry Ginsberg.
7 We appreciate his coming to talk to us.
8 The floor is yours Dr. Ginsberg.
9 DR. GINSBERG: Thank you. I was asked in my
10 view, to give an overview of triglycerides, a little bit
11 about metabolism, and their role as a risk factor. You
12 have heard much of my talk already. Dr. D. gave a very
13 nice presentation. The slides are very similar, so I'll
14 move through my presentation focusing on issues that I
15 think, I believe are relevant to the presentations today.
16 Maybe we'll get more out of the question and answer period
17 than the presentation itself.
18 I thought I would start off, and I didn't know
19 the range of expertise of the panelists, to make a quick
20 point about the fact that we are always interested in
21 cholesterol and triglyceride, and we talk about them
22 together all the time. They are totally different
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
195
1 molecules, other than the fact they are both lipid or
2 organic soluble.
3 This is important for the discussion today
4 in the sense that cholesterol, being a component of all
5 the cell membranes of the body and being used to make
6 hormones and bile, is a very stable measurement in plasma.
7 Total blood cholesterol or serum cholesterol, I think,
8 under ideal conditions of constant diet and activity, the
9 day-to-day variability is about five percent of that
10 measurement, so it's probably somewhat higher than that
11 in real life. Again, the body regulates synthesis versus
12 intake. These membranes are turning over very slowly.
13 In contrast, triglyceride is energy. That
14 is all we use it for. We burn it. We burn the fatty acids
15 to C02 and water, or we store the fatty acids on a glycerol
16 backbone as triglyceride in our fat cells. For that
17 reason, what we move through the blood stream on a
18 day-to-day basis varies much more. Day-to-day
19 variability under fairly controlled situations is about
20 25 percent. In fact, in studies where you are using a
21 drug or a diet or any other intervention and you are looking
22 at changes in triglyceride, you need a minimum of three
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
196
1 measurements over some period of time before or after the
2 pertibation to minimize biologic variability in order to
3 gain the statistical power.
4 The reason that we talk about triglyceride
5 and cholesterol together is because they are both
6 insoluble in plasma, and nature therefore gave us
7 lipoproteins. So we have the triglyceride and the major
8 form of cholesterol, cholesterol esta with a fatty acid
9 attached, in the core of the lipoprotein particle. Those
10 two would not be able to move through the plasma as naked
11 molecules. So they are wrapped up on a phospholipid
12 coating with one or more proteins. The protein is called
13 apoproteins, are key to the function and the sites of
14 delivery of the lipoproteins. They also determine really
15 whether they are atherogenic or anti-atherogenic.
16 The two lipoproteins that we are interested
17 in talking about today are the triglyceride rich
18 lipoproteins, the first being the chylomicron, described
19 as really a fat drop. I call it an oil droplet. The reason
20 that the chylomicron is an oil droplet is that it carries
21 what we have eaten in terms of cholesterol and
22 triglyceride.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
197
1 In an average day, someone eating an average
2 American diet is eating about 40 to 70 grams of fat. They
3 are only eating possibly 300 to 500 milligrams of
4 cholesterol. Even when you add in the cholesterol
5 secreted in the bile, we still -- orders of magnitude,
6 many orders of magnitude differences in amount of
7 triglyceride is being absorbed from the gut versus
8 cholesterol.
9 So we have a lipoprotein that is 95 percent
10 or so triglyceride by weight. Tens of thousands of
11 triglyceride molecules, in fact, and maybe hundreds of
12 cholesterol esta molecules, and then the coating that is
13 phospholipid in protein.
14 So we have the gut making the chylomicrons.
15 They are secreted into the lymph. They enter the
16 bloodstream. They really have two functions. They are
17 to deliver the energy we have eaten to fat cells and to
18 muscle cells. Fat cells for storage, and muscle cells
19 for utilization. They do so by traversing the circulation
20 in the capillary beds of fat and muscle. They meet up
21 with an enzyme, lipoprotein lipase, which can get into
22 the core and lipolyse, hydrolyze the triglyceride. The
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
198
1 fatty acids can be released and they are taken up by the
2 appropriate cells, the atopocytes or the misocytes in the
3 muscle, and utilized.
4 What's left is what we call a chylomicron
5 remnant. It still has the cholesterol it started out with,
6 but has lost, if you're efficient at this lipolysis step,
7 lost about 80 percent of the triglyceride. This particle
8 really heads to the liver. Most of these remnants are
9 taken up by the liver, where they really I think tell the
10 liver what dietary cholesterol intake has been, so the
11 liver then can regulate its own synthesis of cholesterol.
12 That's what happens.
13 Now this process is one in which because
14 there's only milligrams of cholesterol being eaten,
15 there's really very little change in total cholesterol
16 throughout the day. So not only on a day-to-day basis,
17 but even over a 24-hour basis, total cholesterol doesn't
18 really change significantly.
19 On the other hand, triglycerides can go up
20 by folds in different people, depending on how much fat
21 they have eaten, because they are taking in many, many
22 grams of triglyceride in a meal. The level, the rise in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
199
1 triglyceride and the duration of that rise depends on a
2 number of factors. It depends on obviously how much fat
3 you have eaten. Then it depends on the efficiency of the
4 lipoprotein lipase step. In diabetics and in many, many
5 people with hyper triglyceridemia, there are partial
6 defects, sub-optimal levels of protein or activity of the
7 protein, lipoprotein lipase. There are other factors that
8 play into the efficiency of this first step.
9 In general, if someone has a fasting
10 triglyceride of 50 to 100, and they eat a meal with let's
11 say 20 grams of fat, their triglyceride might go up to
12 100 or 150 milligrams per deciliter at the peak. They
13 would be back down to baseline within three to four hours.
14 If someone starts at 250 to 300, and they eat that 20
15 grams of fat, they may go up to 600, 800, even 1,000.
16 They may not come back down to baseline for 10 to 12 or
17 even 14 hours.
18 So you do have this tremendous variability
19 in the postprandial state. In fact, before coming here
20 today, and not really having much of the information about
21 the product, I thought off hand that this utility of this
22 product would be to have people see what happens to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
200
1 themselves after they eat a meal, and that it might have
2 some influence on compliance with a low-fat diet, because
3 those changes would be much more dramatic than what people
4 might see after a few days on a diet and looking at their
5 fasting levels. I can get back to that.
6 I was a little disappointed with the fact that
7 the method seems to be calibrated in linear up to about
8 500. But clearly, people could see if they were up over
9 that limit, they could see differences. Since we would
10 be dealing with big differences in the patient population
11 of interests, that might be a utility of looking, having
12 people -- and I'm not thinking of people doing this every
13 day, but at least during a diet change and diet instruction
14 period.
15 There is a specific disorder that has two
16 genetic bases. It's called Type 1, hyperlipoproteinemia.
17 Either people are missing the lipoprotein lipase, and
18 so they never break down the triglyceride and make the
19 remnant, or they are missing a protein called apo C2, which
20 is a necessary activator of lipase.
21 In either case, after these people eat a meal,
22 whatever fat they have taken in sits around for very, very
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
201
1 long periods of time. Clinically, these people have
2 platosplenomegoly, because the way they get rid of the
3 chylomicrons is actually to phagatacize them into
4 macrophages in the spleen and the macrophages in the liver,
5 the cutfer cells.
6 They get these eruptors anthromas because
7 they actually get triglyceride levels in the many
8 thousands. Some of them run 10,000 triglycerides. They
9 actually get collections of chylomicrons that block
10 capillaries in the skin and burst those capillaries. You
11 get achnea form eruptions. You can actually see, if you
12 look in their retina, you see tomato soup or yellow looking
13 blood vessels. If you draw their blood, it looks like
14 cream of tomato soup within minutes, very frequently.
15 So these people have this complex of signs,
16 but it's clinically relevant because they get
17 pancreatitis. We don't know exactly why they get
18 pancreatitis because all these lipoproteins are much
19 smaller than red cells. But they probably when they have
20 triglycerides in the thousands and they are mostly
21 lipoprotiens, they act -- chylomicrons is probably an
22 aggregation or a fusion of the chylomicrons. Then they
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
202
1 can be large enough to cause capillary occlusions. You
2 get some microinfarks, and you get pancreatitis.
3 So another potential utility I thought of this
4 over-the-counter home-use triglyceride would be in this
5 very small population of people, maybe one in 100,000,
6 who run triglycerides over 1,000, and who might be able
7 to know when they are getting close, that is about the
8 point where the pancreatitis risk goes up, certainly over
9 2,000. But again, that would be -- that's clearly limited
10 by the methodology. I don't know then if even if this
11 small population would be helped if they knew they were
12 over 500 or not because they are almost always, even on
13 the strictest lowfat diet, they tend to run triglycerides
14 of 500 to 1,000.
15 Now the chylomicrons themselves are not
16 thought to be atherogenic, but the chylomicron remnants
17 are. Chylomicron remnants accumulate in many people for
18 a number of reasons, in part probably because there's less
19 then optimal lipoprotein lipase. Even though I classify
20 the remnant as coming after the lipolysis step, there is
21 a lot of overlap between those two classes.
22 There are also other defects at the site of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
203
1 the liver in terms of remnant removal, that can lead to
2 the accumulation of these remnants. In diabetics and in
3 anyone with hyper triglyceridemia, at baseline and a low
4 HDL, you will get prolonged post perandial lipemia and
5 a lot of those particles will be chylomicrons, but a lot
6 will be the remnants.
7 This is a study done in rabbits, but it has
8 been proven in humans now to be relevant, that if you take
9 rabbits and feed them a high cholesterol diet, they get
10 cholesterols in the thousands. A lot of these particles
11 are the remnants. These remnants get into the vessel wall.
12 This is rabbit aorta stained for lipids, showing
13 atherosclerosis. Then some of these rabbits, blood was
14 taken. The remnants were isolated and radio-ionated,
15 injected into other rabbits. They were sacrificed. The
16 aortas were taken out and stained, but they were also
17 slapped on an x-ray film. An autoradiograph showed that
18 the ionated remnant was getting right into the vessel wall.
19 Now from human studies, we have been able to
20 isolate very low density lipoproteins, which I'll address
21 in a moment, and chylomicron remnant-like particles from
22 the coronary arteries of individuals undergoing bypass
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
204
1 surgery. So it's not just LDL that gets into the vessel
2 wall. It clearly is remnants of chylomicrons, remnants
3 of VLDL, and probably even VLDL. So these triglyceride
4 rich lipoproteins are atherogenic.
5 This is a study showing one of many studies
6 showing that post perandial lipemia, the rise of
7 triglycerides after a meal is higher in people with
8 coronary disease. This was a case control study of men
9 undergoing angiograms, and divided into those with more
10 or less than 50 percent stenosis in one vessel. Those
11 designated as cases had higher triglyceride levels,
12 whether or not they were uncorrected or corrected for
13 starting levels. So again, knowing post parandial
14 triglycerides might be helpful, in other words.
15 Having said that, I would point out that the
16 correlation between the fasting triglyceride and the post
17 parandial peak is about .7 in all the large studies that
18 have been done. You can guess to a fair degree, about
19 a 50 percent accuracy I guess, of who will have a high
20 post parandial triglyceride. But there are people who
21 have low HDL and who maintain a low fasting triglyceride
22 by being active and being thin, but who are unmasked when
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
205
1 they eat a fat meal, and have much higher triglycerides.
2 They probably have lipoprotein lipase problems.
3 So in terms of a screening of the population,
4 just as we like to screen for cholesterol and for fasting
5 triglycerides and HDL, doing a post parandial triglyceride
6 or a non-fasting triglyceride, you will pick up some people
7 who will probably be at risk for coronary disease.
8 Therefore, it should have further screenings on them.
9 Most of them will have, or certainly almost all of them
10 will have a low HDL as well.
11 Now the very low density lipoprotein is sort
12 of the mirror image of the chylomicron except it's not
13 related to the intestine or to dietary energy intake.
14 It is related to endogenous energy balance. So the VLDL
15 is about 70 percent triglyceride, and about 20 percent
16 -- 15 percent cholesterol with the coating phyospholipidem
17 protein.
18 It's coming from the liver, but it has again,
19 a very parallel pathway of metabolism. The first role
20 of the VLDL is the liver energy in the form of triglyceride,
21 the fat and muscle. That energy now is coming from the
22 liver. It does so by interacting with lipoprotein lipase
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
206
1 in those same capillary beds. The importance here is that
2 very low density lipoproteins and chylomicrons are
3 competing for the same lipoprotein lipase pathway. So
4 if you start out with excess fasting triglycerides and
5 you eat fat, you are just going to make it more difficult
6 for your lipase to handle the chylomicron triglyceride
7 load.
8 After this step, you end up with a VLDL
9 remnant. I have already told you that that certainly is
10 an atherogenic particle. A lot of this VLDL remnant goes
11 back to the liver, just like the chylo remnant did, but
12 in addition and unlike the chylomicron remnant, the VLDL
13 remnant can be further modified to become low density
14 lipoprotein.
15 The vast majority of insulin-resistant
16 diabetic patients and the vast majority of people who are
17 just hyper triglyceridemia, hyper triglyceridemic, make
18 too many VLDL in their livers, put out too many VLDL,
19 overcome the capacity of lipase, have higher
20 triglycerides. Then you could also see they are going
21 to make more remnants sooner or later, and they are liable
22 to make more low density lipoproteins as well.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
207
1 The common causes of hypertriglyceridemia
2 Type 2 diabetes, probably the most common cause, plus
3 insulin resistance, and I'll show you a slide a little
4 later to point that out, the mechanisms, very badly
5 controlled Type 1 diabetics can have
6 hypertriglyceridemia. They will both over produce.
7 Lipoprotein lipase is an insulin-sensitive or insulin
8 regulated enzyme. So both the Type 2s and the Type 1s
9 who are not well controlled will have somewhat decreased
10 lipoprotein lipase.
11 Excess energy. Any time we are in excess
12 energy balance, we'll make triglycerides out of anything
13 that's around. We'll do that in our liver, and we'll send
14 out the triglycerides for storage on VLDL. So those are
15 the major. Obviously everything is genetic. The
16 genetics confound the energy excess of obesity, and the
17 various non-insulin resistance genes will confound the
18 effect of insulin resistance.
19 You already saw that hypertriglyceridemia is
20 associated with other risk factors. I have mentioned
21 chylomicron remnants and VLDL remnants. There's also low
22 HDL cholesterol levels, and these "atherogenic" small
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
208
1 dense LDL.
2 I would just put this into a quick scheme for
3 you. In the insulin-resistant or diabetic patient, there
4 are increased levels of free fatty acids in the bloodstream
5 because instead of storing energy efficiently in fat, the
6 fat cells are constantly breaking down their own
7 triglyceride and releasing fatty acids. They get to the
8 liver where they liver doesn't need them for energy, puts
9 them back into a triglyceride molecule, which stimulates
10 the assembly and secretion of the VLDL, so you have
11 hypertriglyceridemia. How high, well that's dependent
12 on a number of other genes, but let's say anywhere from
13 150 to 500 in the average insulin resistant and/or Type
14 2 diabetic patient.
15 Once you have high triglycerides, you will
16 have a low HDL cholesterol because cholesterol from HDL
17 will move into the VLDL and exchange for triglyceride.
18 There is a protein called cholesterol that's the transfer
19 protein, that mediates that step.
20 You will also have a small dense LDL because
21 cholesterol from LDL can move back into its precursor VLDL
22 in exchange for triglyceride. That triglyceride can be
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
209
1 acted upon by lipase's shrinking the LDL. So these three
2 -- and they are all probably atherogenic states. These
3 are atherogenic particles. These are at least, if not
4 more atherogenic than normal LDL, and a low HDL is probably
5 an important -- a loss of anti-atherogenic potential of
6 HDL in terms of either anti-oxidant activity in the vessel
7 wall or clearing cholesterol out of the extra cellular
8 space in the vessel wall and bringing it back to the liver.
9 What's the evidence? You have seen, as I'll
10 move real quickly, certainly in epidemiologic studies like
11 the Paris perspective study, this is a German study in
12 Munster called the pro-cam study. Triglycerides are
13 important, particularly when LDL cholesterol is high or
14 HDL is low.
15 You saw the Copenhagen study. Important data
16 that in a multivariate analysis, there was still at least
17 in a tertile calculation evaluation, there was still a
18 doubling of risks which was significant between the lowest
19 and highest tertiles.
20 This is a biograph showing the results of a
21 multivariate analysis done in many, many studies, with
22 46,000 men and 10,000 women. This is a univariate analysis
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
210
1 showing that a rise in triglyceride of 90 milligrams per
2 deciliter increased risk in men by 30 percent, and in women
3 by 75 percent.
4 When they did a multivariate analysis here,
5 and a lot of us think they are not that useful because
6 triglycerides in HDL and triglycerides in LDL are linked
7 in a variety of complicated ways, this dropped to 15
8 percent, but was still significant. This was significant
9 at 35 percent. So triglyceride, even in this complex
10 mathematical analysis, with enough power stands out as
11 adding prediction for risk.
12 Again, what about intervention trials? Is
13 it important to know your triglyceride and lower it? Well,
14 in the Helsinki heart study, as you heard, the majority
15 of the benefit came from lowering triglyceride in the group
16 that hyper triglyceridemia above 200 and a bad LDL to HDL
17 ratio. This is the placebo group, and this is the
18 treatment group.
19 Treatment in the rest of the groups was not
20 significant in this study. That doesn't mean it's not
21 important, that this people didn't have lipid disorders.
22 All the power was here because this was the worst group
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
211
1 in the placebo group.
2 Finally, just two or three quick slides on
3 a study that was published in July in the New England
4 Journal. Again, you saw one slide, the VA Hit study, which
5 comes as close as we have to a study where not changing
6 LDL and starting with an LDL that was normal to begin with
7 was associated with benefits.
8 So this was a study focused on secondary
9 prevention, men who had had events, who would have low
10 HDL, normal LDLs, and moderate triglyceride levels. They
11 actually had HDLs of 32, triglycerides of 161, which would
12 not usually be considered high. An LDL of 111, which
13 probably wouldn't be treated in most offices, even if a
14 man or woman had an MI already, that's close enough to
15 the goal in most instances that people wouldn't treat that.
16 Yet when they would treat it with gemfibrosel
17 or placebo for five years, HDL went up seven percent.
18 Triglyceride dropped 25 percent. No change in LDL, and
19 a 22 percent reduction in events, MI, and non-fatal and
20 fatal CHD events, at 27 percent reduction in stroke.
21 If you look at these data, getting into the
22 epidemiology, you cannot explain all the benefit from the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
212
1 rise in HDL, not based on the epidemiologic data that's
2 been looked at. Therefore, you would have to assume, I
3 think, that triglycerides falling had some independent
4 effect and/or this change in size in LDL that probably
5 occurred would have some independent effect. There are
6 also data being presented at the Heart Association next
7 month indicating that the level of remnants in these
8 patients was predictive, and it dropped during treatment
9 with gemfibrosel.
10 So those are my slides. I would think that
11 -- I talked a little bit about the utility of an
12 over-the-counter device for post-parandial. For the rare
13 patient with several hypertriglyceridemia, I think it
14 falls a little short in the latter. I think certainly
15 to increase people's self knowledge and self empowerment
16 is always good. I think that allowing people to screen
17 themselves with an over-the-counter device if they haven't
18 been to their doctor and get a triglyceride, particularly
19 a non-fasting one, might send more people to their
20 physicians for appropriate follow-up and fasting
21 screening, et cetera.
22 I think the day-to-day biological variability
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
213
1 is a big problem here. If you add that to test
2 variabilities, I think there is the conundrum to me,
3 because most interventions for triglyceride, diet, weight
4 loss, exercise, even drugs that we use, lower triglyceride
5 in the range of 15 to maybe 40 percent. I'm not sure what
6 people -- will happen if somebody goes on a diet for a
7 couple of weeks, and because of biologic variability and
8 test variability, they look worse or no different.
9 We have had that problem with cholesterol
10 tests with modest interventions. Before the Statons were
11 there, huge ability to lower LDL cholesterol, physicians
12 and the public were not convinced that many of the things
13 we could do for cholesterol were really happening because
14 measurement variability and biological variability
15 clouded that issue.
16 So I think I'll stop. I will be happy to take
17 questions with everyone else.
18 CHAIRPERSON NIPPER: Dr. Ginsberg, we
19 appreciate your presentation. We hope that -- will you
20 be able to stay after lunch to take questions?
21 DR. GINSBERG: For a while.
22 CHAIRPERSON NIPPER: Okay. Because we are
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
214
1 about half an hour beyond the schedule for lunch. I'm
2 sure it's sitting there. I'm sure that we'll all need
3 to replenish our triglycerides, particularly from the
4 potato chips.
5 If there is a member of the panel who has a
6 pressing question for Dr. Ginsberg before he leaves the
7 podium, now may be the time to ask it.
8 Dr. Doumas?
9 DR. DOUMAS: I have two quick questions, Dr.
10 Ginsberg. In view of the very high intra-individual
11 coefficient of variation, which is about 22 to 23 percent,
12 do you feel that accuracy is more important than precision
13 in the measurement of triglyceride?
14 DR. GINSBERG: Gee, that's a -- I mean if I
15 was -- no. I would think -- I think obviously they are
16 very separate issues. I am trying to quickly put them
17 into their appropriate places.
18 If I was testing my triglyceride at home, and
19 I wanted to know if what I was doing was helping, then
20 I would want precision. Right? So that I would know that
21 if I went up or down. Whereas if I wanted to know what
22 I was, certainly the first time, then I would want accuracy.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
215
1 So I think they are obviously both -- if I
2 had, for instance, if I have a triglyceride that's truly
3 200, but it can either be 150 or 250, that doesn't help
4 me. If I do something and I go from 300 to 200, I also
5 need precision there. So I am not sure I can give you
6 an answer which I would pick.
7 DR. DOUMAS: I'll follow up with the second
8 question, the same vein. In view again of the very high
9 coefficient, intra-individual coefficient of variation,
10 if you increase the random error, first of all if you
11 decrease the random error to zero percent, or you from
12 five percent to ten percent, the total variability is
13 hardly affected because we see the total square is equal
14 to the square of the analytical plus the biological CV
15 square.
16 When you have 23 square and five percent
17 square, that five percent doesn't alter the total square
18 root. I mean it alters very little.
19 I have something from a paper published by
20 Cooper, Jerry Cooper. It was published in Clinical
21 Chemistry, showing what the effect is really of the random
22 error on the total variability of triglyceride
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
216
1 measurement.
2 This is the reason I asked the first question,
3 is accuracy more important than precision, than decreasing
4 the analytical error.
5 DR. GINSBERG: I think in general, yes.
6 Again, it's difficult. If we do studies and we have
7 something that we expect to change, we do a multi-center
8 study and we have each of the centers doing the TG at their
9 own place, we want precision. Right? If we're looking
10 for delta, people have a little bias with statistics.
11 But in real life, when a patient is looking at the number,
12 I guess you would want accuracy.
13 CHAIRPERSON NIPPER: Thank you very much.
14 We'll reassemble in about an hour. We appreciate the FDA's
15 presentation and the sponsor's presentation. We will go
16 chew on it a little bit.
17 (Whereupon, the foregoing matter went off the
18 record for lunch at 1:24 p.m. and went back
19 on the record at 2:26 p.m.)
20
21
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
217
1
2
3
4
5
6
7
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
218
1 A-F-T-E-R-N-O-O-N P-R-O-C-E-E-D-I-N-G-S
2 (2:26 p.m.)
3 CHAIRPERSON NIPPER: We know that Dr.
4 Ginsberg has a plane to catch. We appreciate him staying
5 over the lunch hour to catch some questions. So when we
6 start with Dr. Rej and we go around the panel, if we could
7 concentrate our laser beams on Dr. Ginsberg and let him
8 get out of here, I think it would be very helpful to him.
9 MS. PINKOS: Do you want to do that before
10 the questions?
11 CHAIRPERSON NIPPER: That's what we are doing
12 about questions. We have already heard from Dr.
13 Ginsberg's lecture, but we needed to do questions. So
14 I am sure that the panel has some questions for him.
15 Dr. Rej?
16 DR. REJ: It's always so hard to go first.
17 (Laughter.)
18 DR. REJ: Thank you, Dr. Ginsberg, for your
19 presentation. In your presentation, you, if I quote you
20 correctly, you said that it doesn't help you very much
21 -- you can't tell the difference between 150 and 250.
22 Is that correct? Because 200 plus or minus 50 milligrams
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
219
1 per deciliter?
2 DR. GINSBERG: When I was asked the question
3 about accuracy versus precision.
4 DR. REJ: Well, in terms that that 150 could
5 equal 250, would not be helpful to you.
6 DR. GINSBERG: I think that there are two
7 issues. One, if you are working with methods that even
8 if you are working with precise methods, but the
9 intervention is modest, then biologic variability can
10 really be a detriment. We have all had, those of us in
11 the field have had patients who we have put on diets.
12 They can drop their cholesterol 10 percent on a diet.
13 They come back two months later, they have been working
14 really hard, and their cholesterol may be a few points
15 higher than it was previously, because the lab happened
16 to be on the high side and their biologic variability was
17 on that high side that day. That is not helpful.
18 So on the other hand, triglycerides are.
19 Usually you can get bigger changes with things like diet
20 and exercise. So you have a wider window. But obviously
21 you need accuracy still. If someone was 200 and I had
22 sent them home and I told them to really work hard on diet,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
220
1 and they came in and they had lost five pounds and they
2 were truly biologically 150, but because of a bunch of
3 errors in biologic variability, they were still 200, then
4 I have to deal with that as a physician.
5 DR. REJ: Right. Okay, well that goes to the
6 point of the question. That if you follow the NCEP
7 guidelines, then that basically for that same 200
8 milligram deciliter triglyceride, basically 170 equals
9 230. If you allow a total error of 30 percent, which was
10 one of the models that we saw today, then 140 equals 260.
11 You gave the example of 150 equals 250, if you have plus
12 or minus 50 milligrams per deciliter total error. I was
13 wondering if you could comment on the NCEP
14 recommendations, which is something this panel is going
15 to have to get to, and how does that fit into a total error
16 budget for triglyceride measurements. Is that one,
17 realistic, and two, is it clinically useful?
18 DR. GINSBERG: I'm not sure how to answer.
19 I mean I think the NCEP guidelines are outdated for
20 triglycerides. I would agree with what Dr. Deeg said on
21 that. I think that the ATP-3 is meeting, or will meet
22 soon. I think they will be changing those.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
221
1 I think to say that 200 to 400 is borderline
2 is not appropriate any more.
3 DR. REJ: Okay. I am sorry. I was referring
4 particularly to the NCEP guidelines for the measurement
5 of triglycerides, not so much the cut-point
6 recommendations, but the total error of 15 percent and
7 the bias, a bias factor of less than five percent.
8 DR. GINSBERG: I think that you have to be
9 close to that, for any method to be valuable for feedback.
10 Because of the biologic variability. I think you have
11 to be close to that. Wet laboratories have one percent,
12 two percent TG coefficients or variations. We need that
13 because of the biologic variability. So I think NCEP
14 guidelines are appropriate.
15 DR. REJ: Thank you.
16 I have some other questions, but they are not
17 for Dr. Ginsberg.
18 CHAIRPERSON NIPPER: Let's just hold those
19 until we let him catch his shuttle.
20 Dr. Everett, why don't we just proceed on
21 around the room for questions for Dr. Ginsberg, if you
22 don't mind.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
222
1 DR. EVERETT: My questions deal with again,
2 whether this test works. But if you get beyond that, how
3 do you perceive that this test could be used so that
4 physicians will get maximum benefit? That is the part
5 I struggle with at the moment.
6 DR. GINSBERG: Well, I mean I don't see it
7 as a glucose home monitoring in a Type 1, obviously, where
8 people are changing their treatment every day, if we are
9 really going for optimal control. If you look at glucose
10 monitoring in Type 2s, they don't really change their
11 treatment most of the time on a day-to-day or even few
12 day, but they gather data that lets them know where they
13 are. It would fall closer, certainly on that side of the
14 range of utility.
15 But I would think that as I said, it might
16 have utility if I put a patient on a diet and they could
17 actually see what changing their diet did if I explain
18 to them what happens after they eat, and talked about post
19 parandial. But more likely, I would maybe use it if a
20 patient wasn't coming in for four months and maybe in
21 between, they were going to test it and call me. Again,
22 it has to be really accurate to be helpful there.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
223
1 I would see it also as a way for people to
2 screen themselves on a one-time basis, to identify
3 themselves as needing follow-up. But I think, you know,
4 I would see it being used maybe a few times a year.
5 DR. EVERETT: What about the inappropriate
6 use? That is, using it as a diagnostic tool. Do you think
7 it is acceptable for that purpose?
8 DR. GINSBERG: You mean for an individual to
9 test themselves?
10 DR. EVERETT: And determine that they have
11 elevated triglycerides, without seeing a physician.
12 DR. GINSBERG: Well, if they do that and then
13 they see a physician, obviously if they are a false
14 positive, then there is some inefficiency there. But I
15 would think -- I think screening is good. I think the
16 more screening you have, the better off you are. So I
17 would rather get a few false positives and identify more
18 people. That is my view.
19 DR. EVERETT: So you think it is acceptable
20 to use it as a screening tool?
21 DR. GINSBERG: I would be positive about
22 having the availability for someone who hasn't been to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
224
1 a doctor to see this in a store, stick their finger, and
2 find out their triglyceride. Even if its fasting, if they
3 find the triglyceride of 300, that would get them to go
4 to their physician.
5 DR. EVERETT: Okay. Thanks.
6 CHAIRPERSON NIPPER: Dr. Manno, do you have
7 any questions?
8 DR. MANNO: I have no questions.
9 CHAIRPERSON NIPPER: Dr. Doumas?
10 DR. DOUMAS: I had a question, and I am going
11 to ask you, Dr. Ginsberg. When it comes to when in the
12 labels, around this test, two to three times a year or
13 at least monthly if you are diabetic or a woman
14 post-menopause. Do you agree with that statement, that
15 recommendation?
16 DR. GINSBERG: I can't see the reasoning
17 behind the monthly testing for post-menopausal women.
18 The majority of post-menopausal women don't have
19 dyslipodemia. Obviously women aren't as well off
20 lipid-wise or insulin sensitivity wise post-menopausally,
21 but that doesn't mean they become diabetic or dylipidemia.
22 Every woman probably should have her lipids
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
225
1 measured post-menopausally if it hasn't been done before
2 or even if they have been done before, because there are
3 some changes. They do become much more prone to
4 hypertriglyceridemia. But I don't see any reason if the
5 first test is normal, to follow that up on a monthly basis.
6 In terms of the diabetics, most diabetics
7 should be seen every three to four months, depending on
8 their complications. So I would possibly see maybe two
9 to three times a year, you know, somewhere in between those
10 visits as possibly being of value. Or if I change the
11 therapy and they went home and in two weeks instead of
12 coming to see me again, if I knew they could test well
13 and they had in that case maybe precision, to tell them
14 to try that and have them call me rather than having them
15 come into the lab. But I don't think I would see it once
16 a month.
17 DR. DOUMAS: Also for non-diabetics and
18 non-menopause, do you think that -- they are past
19 menopause, do you think that somebody should do it three
20 times a year?
21 DR. GINSBERG: No.
22 DR. DOUMAS: Thank you.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
226
1 DR. GINSBERG: I think if the first one is
2 normal, I would follow NCEP guidelines there. At that
3 age, probably once a year people should get tested. Now
4 whether they should get tested at a lab which is a wet
5 bench or home is a separate issue.
6 DR. DOUMAS: That's another issue. Thank
7 you.
8 CHAIRPERSON NIPPER: Dr. Janosky? Dr.
9 Lewis?
10 DR. LEWIS: I have nothing.
11 CHAIRPERSON NIPPER: Ms. Kruger?
12 MS. KRUGER: I am just struggling with the
13 whole notion of acceptability, when to use an accuracy.
14 Most of the patients we -- we don't see most of the Type
15 2 patients. The majority of us specialists don't see
16 those. Those are seen by the internists. Even with the
17 UK PDS studies, which are hammering away tight control,
18 tight control, control the blood pressure, control the
19 lipids, the fact is, we still haven't been able to get
20 that through for a lot of reasons to the internists. They
21 don't have the time. I mean there's just a host of things.
22 One of the reasons they don't want to do blood
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
227
1 glucose or recommend to their patients to do blood glucose
2 is that it's more data they have to deal with on the
3 telephone, more data that they have to deal with when the
4 patient comes in, and the concern of accuracy when they
5 are manipulating medication.
6 So I guess I am struggling with -- I think
7 blood glucose monitoring today, the meters we see today
8 are accurate. I am not comfortable yet that we are going
9 to see reproducability in this device. I am concerned
10 that while I believe in empowerment of the patients, we
11 still have to have buy-in by the healthcare community.
12 I am not convinced we are going to get the internists
13 buy in who, for the most part are going to be -- in the
14 way, in the global way we are speaking about using this
15 meter, would be using this meter.
16 Those are things that are sort of catching
17 me up, is --
18 DR. GINSBERG: Let me just respond. I think
19 in the lipid field, we are so -- I mean there has been
20 such an abysmal participation by the general health
21 community in terms of treating hyperlipidemia in the face
22 of tremendous data. I mean if you want to practice
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
228
1 evidence-based medicine, there may not be a better model
2 than cholesterol-lowering trials and even the
3 triglyceride lowering trials. That if patients find out
4 they have a bad number, and they go to their physician
5 and ask them why they are not being treated, to me maybe
6 that's what we need to do. It's not optimal.
7 Again, I don't see this as something that
8 people would be doing weekly or monthly to help manage
9 this. I would see it being done at much more infrequent
10 intervals as an adjunct.
11 MS. KRUGER: That would also raise my
12 eyebrows again, is if these people are not monitoring
13 blood, so they don't have the technique for that, and they
14 are doing it so sporadically, are we in fact going to get
15 accuracy and precision? I just throw that out as a
16 concern.
17 The other question I have, are there other
18 labs that need to be done at the same time, say every four
19 to six, eight, twelve weeks, like we do when we do lipidtor,
20 zocor, liver enzymes, or any of those that need to be done,
21 that we'll miss doing for making sure that the drugs are
22 safe if the patients aren't coming in frequently enough
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
229
1 for --
2 DR. GINSBERG: For the Stattons now, I think
3 the FDA actually, I think the package inserts talk about
4 twice a year for LFTs. So hopefully this wouldn't replace
5 the patient's visit to the doctor.
6 CHAIRPERSON NIPPER: Dr. Rosenbloom, did you
7 have anything?
8 DR. ROSENBLOOM: Yes. How many patients
9 have you seen in the last couple of years, the last three
10 or four years since the cholesterol testing has been in
11 existence, who have come to your attention because they
12 have gone out and gotten this test and found it abnormal?
13 DR. GINSBERG: None.
14 DR. ROSENBLOOM: Okay. Secondly, have you
15 ever ordered a triglyceride by itself?
16 DR. GINSBERG: No. Possibly rarely in a
17 patient whose triglycerides in the thousands. But
18 otherwise, no.
19 CHAIRPERSON NIPPER: Dr. Clement?
20 DR. CLEMENT: Interesting. Have you been
21 concerned about the blank, the unblank issue with high
22 triglyceridemia, hypertriglyceridemia in folks with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
230
1 diabetes? Is that a clinical and significant issue that
2 we should be concerned about?
3 DR. GINSBERG: So the issue of glycerol. I
4 think glycerol levels tend to be somewhat greater
5 proportion of triglyceride. Free glycerol tends to go
6 up as triglycerides go up. Post parandially, where I was
7 thinking this might have utility with all the lipolysis
8 going on, glycerol levels could be even a little bit higher.
9 Usually there is still not more than maybe
10 I think 10 percent or 15 percent of value. So there might
11 be an overestimation of the triglycerides by that level.
12 Luckily, they turn over so fast that even during
13 lipolysis, they probably don't go up that much. But it
14 is another built-in error to the system that doesn't have
15 a blank.
16 MS. KIMBERLY: I don't have any questions.
17 DR. RIFAI: I don't either.
18 CHAIRPERSON NIPPER: Well, we thank you very
19 much, Dr. Ginsberg, and hope you make your shuttle.
20 Thanks, Arleen. Were you going to go over
21 the questions?
22 MS. PINKOS: Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
231
1 CHAIRPERSON NIPPER: Okay.
2 DR. REJ: Is there time to ask questions about
3 the FDA presentation?
4 CHAIRPERSON NIPPER: Why don't we just go
5 over the questions. Then we'll just throw it open and
6 ask questions to the FDA persons and others that come up.
7 Is that all right with you? Is the panel okay on that?
8 MS. PINKOS: Okay. I have 10 questions that
9 we would like you to discuss during your deliberations
10 this afternoon. The first one is, is the Agency's
11 requirement for comparison to a CDC reference method
12 reasonable when the device is intended for
13 over-the-counter use? Other options include
14 characterization of only the predicate device by a CDC
15 lab or no requirement at all for any comparisons to a CDC
16 reference lab and simply label the device as not having
17 been evaluated by a method recommended by NCEP.
18 The other thing I want to say here is that
19 we realize that CDC reference laboratory studies are
20 undoubtedly more costly and labor intensive for the
21 sponsor. The Agency is committed to finding the least
22 burdensome path to market for industry. But at the same
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
232
1 time, we don't want to compromise safety or effectiveness
2 in the device. So we're asking here what the minimum study
3 should be for a device.
4 Okay. Should an over-the-counter device be
5 required to meet NCEP performance goals? If a device does
6 not meet NCEP performance goals, is it acceptable to clear
7 it for marketing with cautionary labeling? Or if NCEP
8 is not an appropriate minimum threshold for performance,
9 what criteria should be applied?
10 Because it is difficult to convey the
11 performance limitations of a laboratory test to a lay user,
12 and because results are interpreted as a single entity,
13 FDA generally believes that over-the-counter devices need
14 to be pretty good. So should these NCEP performance goals
15 be the threshold? Or do you think that there is another
16 more appropriate yardstick?
17 I just want to make a couple extra points on
18 this, that the question was raised earlier whether NCEP
19 had specific performance goals for whole blood assays.
20 If you read their information, they don't specifically
21 address whole blood performance. But they do make the
22 statement, I believe it's on the bottom of page 133, that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
233
1 the performance goals are appropriate for all methods,
2 depending on where they are being performed. They go into
3 a little bit of discussion about that, if you are interested
4 in reading it.
5 But NCEP, we have to remember that NCEP is
6 not used to balancing performance requirements with the
7 claims that are associated with the device. We're also
8 considering the benefits that a device being run at home
9 might have. Sometimes there is a little bit of a tradeoff
10 there.
11 In the packet that was sent to you from NCEP,
12 there is also a discussion on total error and why NCEP
13 likes total error so much. That sort of ties into a little
14 bit of the discussion that we were having earlier. The
15 reason that they like total error is it is give and take
16 on accuracy and precision, so that if a device is more
17 accurate and less precise, it's still going to meet a total
18 error target or vice versa. So it is a little bit forgiving
19 in each direction there.
20 Okay. What is the appropriate minimum sample
21 size for evaluating an over-the-counter lipid test, and
22 should there be a minimum requirement for sample
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
234
1 distribution? That was brought up in the review of the
2 product here today, that there was -- a couple people agreed
3 from both sides of the fence, if you will, that they needed
4 to get a few more samples in that upper end. So if you
5 have any recommendations for us in general on lipid tests,
6 we would appreciate that.
7 Has the sponsor done the appropriate
8 precision and interference studies? What is the
9 appropriate claim for this device? Has the claim
10 encaptured in the labeling? Is performance adequate to
11 support the intended use or claims for this device? Again,
12 this just means to make the point that the performance
13 required is very much linked to the intended use of the
14 device or how it is going to be used. That question of
15 course relates to the last question, what should be the
16 intended use of the device.
17 The last four questions that I have all have
18 to do with the labeling. Is it an acceptable approach
19 to refer lay users to a second professional use package
20 insert for additional information such as quality control
21 instructions. The sponsor is taking this rather unique
22 approach to labeling. Do you think this is an acceptable
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
235
1 way of transmitting that information?
2 Does the labeling adequately address quality
3 control instructions, recommendations, and
4 interpretation? Is performance of the device properly
5 conveyed in the labeling? You'll notice that the
6 performance is currently -- and we already discussed this
7 a little bit, the performance is currently presented as
8 being accurate in 95 percent of the time. Precision is
9 described in terms of mean, standard deviation, and
10 coefficient of variation. Sometimes that's not the
11 easiest way to convey information to lay users. So if
12 you have got some guidance for us and the sponsor, we would
13 appreciate that.
14 Finally, do the benefits of this device
15 outweigh the risks? This is almost a rhetorical question
16 by the time you answer the rest of the questions, but it
17 is one that we always ask for over-the-counter products
18 because it is such a fundamental review issue, not just
19 for over-the-counter, but for all products. You would
20 have to consider all parts of the device and what it brings.
21 Considering the performance of the device and
22 its intended use, can it be labeled in such a way so that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
236
1 the information is more beneficial than harmful to the
2 consumer?
3 That's it.
4 CHAIRPERSON NIPPER: Okay. Thanks. Now
5 just to get us back on track here, since Dr. Ginsberg had
6 to kind of scoot, we did the Q&A period for him after the
7 FDA presentation. There may be some questions that the
8 panel has of the three presenters from the FDA, Arleen
9 Pinkos, Carol Benson, or Telba Irony, the statistician.
10 I think we can go ahead and go around the room
11 and try to get FDA questions, specific questions about
12 their presentations, any clarifications that you need to
13 make for that. Then the agenda calls for an open public
14 hearing. We'll do that. Then we will devolve into open
15 committee discussion, where I will try to answer the
16 questions that Ms. Pinkos has discussed with us now.
17 We will try to get as many of these questions
18 answered by panel members as possible, and final
19 recommendations made as soon as possible. But we would
20 like to get a thorough job done here before we lose our
21 panel.
22 So Bob, did you have questions for any of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
237
1 three FDA presenters?
2 DR. REJ: I do have one question that's
3 related to the FDA statistician's presentation, but maybe
4 it can be clarified from the sponsor. It has to do with
5 the study on precision. If you can put those data on the
6 screen, I don't know how hard that is to do. But this
7 was a case where there were three consumers, the precision,
8 inter-consumer precision was estimated. Perhaps the
9 sponsor could explain exactly what was done for that study.
10 DR. IRONY: I'm trying to get the
11 presentation.
12 CHAIRPERSON NIPPER: Bob, that's the one, the
13 slide that says whole blood, three consumers, two levels
14 each?
15 DR. REJ: Yes. Two levels each, and consumer
16 one, consumer two. It has standard deviation CV. So
17 perhaps the sponsor could --
18 CHAIRPERSON NIPPER: I think that's in the
19 packet too.
20 DR. IRONY: Yes. Maybe you can ask the
21 question.
22 DR. REJ: Give me a little information about
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
238
1 the specimen used in those studies, how that study was
2 done.
3 MS. PINKOS: That's the three consumers with
4 the whole blood?
5 DR. REJ: Yes.
6 MS. PINKOS: It was an EDTA whole blood
7 sample.
8 DR. REJ: Right. And across consumers, was
9 it the same sample?
10 DR. IRONY: No. Each consumer had a
11 different sample.
12 DR. REJ: A different sample?
13 DR. IRONY: Actually they had two different
14 samples each.
15 DR. REJ: Each. So that the two samples that
16 were given to consumer one are different samples than those
17 given to consumer two?
18 DR. IRONY: Right.
19 DR. REJ: Thank you.
20 DR. EVERETT: Questions for the
21 statistician. In the slides you showed where the levels
22 of triglyceride was labeled 100 percent, 200, and 400 --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
239
1 DR. IRONY: Let me just try to find it.
2 DR. EVERETT: And you gave the total percent
3 of.
4 DR. IRONY: Which one? The precision study?
5 This one?
6 DR. EVERETT: No. It was a summary slide
7 where you had the levels of triglyceride measured compared
8 to the total percent error.
9 DR. IRONY: I have to go through the whole
10 thing?
11 DR. EVERETT: You may be able to answer my
12 question without the slide.
13 DR. IRONY: Okay.
14 DR. EVERETT: My question is, is the total
15 percent error at each one of those sub-categories, is the
16 total percent error significantly different between any
17 of the two levels?
18 DR. IRONY: Let me just clarify here. Here
19 is not the total percent error. What it shows in this
20 slide is actually the difference, what will be the bias.
21 The difference between the measurement taken by the
22 BioScanner and the measurement taken by the reference lab.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
240
1 What I'm showing here is just like a descriptive analysis.
2 These red bars, they represent the 15 percent. Of course
3 15 percent of 400 is a lot more than 15 percent of 100.
4 So that's why these intervals are larger.
5 The points of the actual measurements, so you
6 can see, for instance, around 100 you have more points
7 outside the bars. Whereas in 400, well we don't have
8 enough, but let's say at 300, we have more points in between
9 the bars because the percentage is higher there.
10 DR. EVERETT: Can I realistically say the
11 total percent area is greater at one level than another?
12 DR. IRONY: No. We can say from here. No,
13 what we can say is that the tolerance limits are different.
14 That's what we can say here. The tolerance limits at
15 the level 400 here will be greater than at a level 100.
16 We can say they are greater -- I mean the actual
17 measurements are greater because we don't have enough
18 points. You can see at 400, we lost some points.
19 DR. EVERETT: Okay. Then my other question
20 is, do you think the target population that they used to
21 generate the data resembles the population at large? When
22 I asked the question, they said there was no limit on who
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
241
1 could use this particular test. I assume it's for the
2 general population. In the study, the population that
3 they use in the study, in my opinion, doesn't represent
4 the general population.
5 DR. IRONY: Yes, I guess the physicians, the
6 clinicians will be able to tell better than myself. You
7 know, this is a small sample of 220. I don't know if it
8 represents the target population or not. Then the sample
9 has to be analyzed and see how many diabetic people are
10 in this population and we don't have this data.
11 DR. EVERETT: So how do you decide then who
12 is recommended to use this particular test?
13 MS. PINKOS: The sponsor hasn't in their
14 claims of their intended use specifically targeted any
15 particular group. I think they have said, and they can
16 chime in if they would like, that probably who was going
17 to end up using it are diabetics, just because of the way
18 their device is packaged with glucose and cholesterol.
19 They are also trying to get some other tests that are
20 probably typically used by diabetics.
21 So whether they shot themselves in the foot
22 or did themselves a favor by having a non-diabetic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
242
1 population doing the study, I don't know. I mean they
2 made the argument, and there is certainly some truth to
3 it, that if you have a diabetic population that are used
4 to doing finger sticks more often, but then we also have
5 to consider that it is being put on the market and anybody
6 can buy it, and anybody can use it. So because they are
7 not limiting their claims for use, the target population
8 that they did use is probably okay.
9 DR. EVERETT: Okay. Then my last question
10 is, are there any statistically strong points about this
11 study that support the intended use of this particular
12 instrument?
13 DR. IRONY: What would be the statistical
14 strong points?
15 DR. EVERETT: Well that's what I'm asking.
16 DR. IRONY: What would you desire from such
17 an instrument?
18 DR. EVERETT: Well, something better
19 precision-wise, as well as accuracy so that when I look
20 at total error, it's better than what's being presented
21 here.
22 DR. IRONY: Well, I guess that is the data
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
243
1 we have got. I can't tell from what I haven't seen, so
2 I guess that's the data we have got and that's what is
3 happening.
4 DR. EVERETT: That, I understand. But my
5 question is what do you think are particular strong points?
6 MR. GUTMAN: Actually, that is why we asked
7 you folks to come together. I think that Dr. Cooper sort
8 of captured I think the issue that we deal with. There
9 is no question that there is a performance deterioration
10 on this assay. There is no question that there is a change
11 in access which might be good or might not be good. As
12 you are deliberating, that should be a light motif behind
13 the thing. Is the compromise from performance outweighed
14 by the access or is it not? We don't know. We're
15 confused. We don't have the expertise that this group
16 has. We are honestly asking you to tell us,does this
17 performance match either this intended use or other
18 intended uses that might make this reasonable to put on
19 the market?
20 DR. EVERETT: Sure. That, I understand. My
21 question is one about statistics, not about all the
22 parameters of the study. She is a statistician. She is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
244
1 an expert in that area. I think she is qualified to answer
2 the question.
3 DR. IRONY: But I still don't understand your
4 question. If you want me to tell what is the good side
5 about the instrument, I don't know.
6 DR. EVERETT: Only about the statistics and
7 as it relates to the data that they presented.
8 DR. IRONY: Right. That is exactly what I
9 gave you, was just analyze the data and interpret the data
10 as it came. I guess so the clinician is to decide whether
11 the error both in precision-wise and bias-wise will be
12 tolerable or not.
13 DR. EVERETT: This could go on forever.
14 CHAIRPERSON NIPPER: It won't, Dr. Everett.
15 That's why they hired me.
16 (Laughter.)
17 CHAIRPERSON NIPPER: Dr. Rej wanted to jump
18 in for a minute.
19 DR. REJ: One follow-up question which I
20 forgot to ask about the Bland Altman plot that's up here.
21 Again, because of the way you have these red bars that
22 show the NCEP guidelines of plus or minus 15 percent, is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
245
1 it safe to say that there are more data points outside
2 of the NCEP guidelines than there are within in this set?
3 DR. IRONY: Well, you are seeing the same data
4 as I'm seeing. Right.
5 DR. REJ: Again, because those bars are not
6 linear, I mean we're not following -- we have these sort
7 of segmented set. Are there more points? I can't count
8 them, especially from this angle. You know more
9 intimately.
10 DR. IRONY: You can see. You can see in the
11 neighborhood of 100, okay, there are lots, lots of points.
12 The further analysis was then when I did the distribution
13 of the frequency. For the 400 level, we can tell not
14 because there are many, but there aren't enough
15 observations to support --
16 DR. REJ: And of course the NCEP guidelines
17 are more forgiving as you get to higher concentrations
18 of triglyceride.
19 DR. IRONY: In this case, it is, yes.
20 However, 400 is a decision point.
21 DR. REJ: Yes, I'm aware of that. But to my
22 eye, it looks to me like from this data set of comparison
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
246
1 with the reference method, more points exceed the NCEP
2 guidelines than are within it.
3 CHAIRPERSON NIPPER: Let's see if Dr. Manno
4 wants to try to catch a hummingbird with a net or wants
5 to try to nail a statistician down to something concrete.
6 DR. MANNO: I'm happy. I have no questions.
7 I'm waiting until later.
8 CHAIRPERSON NIPPER: Dr. Doumas?
9 DR. DOUMAS: I'm referring to a table where
10 you talk about the random error, upper limit. You take
11 your confidence interval. You have, for example, 35 --
12 or 86, 38, 73. What is the probability of having this
13 kind of error statistically?
14 DR. IRONY: Are you talking about this table?
15 DR. DOUMAS: No. I think it's the previous.
16 Yes, those. What is the probability of having random
17 errors of this size? I mean this is the upper limit.
18 The maximum error in other words is 35 milligrams. How
19 many times out of 100, those values will be obtained, these
20 kind of errors?
21 DR. IRONY: Well, it will depend on the
22 viability of the population you are analyzing. You have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
247
1 to make several assumptions.
2 If the instrument will vary precise. Let's
3 say you have it perfect, you never have any variability.
4 Every time you measure it, you get that value. So I can't
5 tell. That's exactly what we are trying to make inferences
6 based on these samples to say what will be the precision
7 of this particular instrument that we are analyzing.
8 The problem I see in this particular study
9 is that we have only three consumers.
10 DR. DOUMAS: I see.
11 DR. IRONY: You see? So as you can see, that
12 for consumer three, for instance, you know there is low
13 variability either because the person was good in
14 measurement or something we can't control. For consumer
15 two, for the same kind of level, which had a lot more
16 variability.
17 What we will need to characterize this
18 instrument, in my opinion, would be much more consumers.
19 DR. DOUMAS: Another question unrelated.
20 Would you define, please, for an ignoramus, what you mean
21 by goodness of fit? How do you evaluate the thing?
22 Somewhere I have seen in the materials here from the FDA
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
248
1 "goodness of fit."
2 DR. IRONY: Well, these basically will tell
3 when you are fitting a line or any curve to points. We
4 have to see how well you fit.
5 When I did the analysis of residuals, I would
6 say the perfect fit will be all residuals will be zero.
7 Let me go back.
8 What is a perfect fit? All points --
9 DR. DOUMAS: Everything on the line.
10 DR. IRONY: On the line. The further the
11 points are from the line, the worse the fit.
12 DR. DOUMAS: Okay.
13 DR. IRONY: There are statistical tests that
14 will measure how good the fit is and how bad the fit is.
15 But intuitively, you know, the closer the points are from
16 the line, the better the fit is.
17 DR. DOUMAS: The better the fit.
18 DR. IRONY: Right.
19 DR. DOUMAS: Thank you.
20 CHAIRPERSON NIPPER: We're going to have
21 dueling statisticians.
22 (Laughter.)
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
249
1 DR. JANOSKY: I have been waiting. I have
2 a series of questions. Two of them are ones that I had
3 asked the sponsor this morning, so I wanted to address
4 them also to FDA and the reviewers. The other one is sort
5 of a take-off on some of the discussion that we have had
6 here. So why don't I start with that take-off one.
7 At the University of Pittsburgh, we call what
8 was presented this morning Syndrome X. Syndrome X is this
9 syndrome that seems to be associated with obesity,
10 diabetes, high blood pressure, et cetera, which the
11 sponsor had presented and they called it something else.
12 Trying to tease apart whether the population
13 that was studied in the consumer study was appropriate
14 or not. So does anybody have an estimate or can someone
15 please tell me what would be the relationship within a
16 group of diabetics for the triglycerides? What would be
17 the expected value? Because it seems to me that you want
18 to, if the intended use is specifically within diabetics,
19 you want to map your sample group to that particular group
20 of patients.
21 I tried to look it up. I couldn't find
22 anything for triglycerides average in that group of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
250
1 population. Is there some data I'm missing?
2 DR. CLEMENT: I can answer.
3 CHAIRPERSON NIPPER: Steve?
4 DR. CLEMENT: As a clinician, one of the
5 several clinicians in the group, if someone has well
6 controlled diabetes, their triglycerides can range from
7 100 to 150, still to as high as 500. Many of these folks
8 that have lipoprotein, lipase hypermenalese as well, even
9 if they are well controlled. So they can meet the whole
10 spectrum even higher.
11 DR. JANOSKY: So within that, you would
12 expect a larger percentage to be within this 300 to 400
13 range than what we would consider an undiabetic
14 population, a normal population without diabetes?
15 DR. CLEMENT: Yes. I mean there would be a
16 higher percentage depending on how well controlled the
17 population is. So I think it is useful to have accurate
18 data.
19 DR. JANOSKY: So now going back to this
20 question as to whether the appropriate population was used
21 or not used, given that piece of information, can you help
22 me sort of get an assessment as to whether the population
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
251
1 that was used for the consumer study would be very similar
2 to the population of the consumers that would be using
3 this instrument or this device?
4 MS. PINKOS: I would ask you, Dr. Clement,
5 since you were the clinician, to look at for instance,
6 the CRMLN plot, and see where most of the points fell.
7 I mean they predominantly fell promptly.
8 DR. CLEMENT: Just a handful, about 400.
9 That is concerning. I think that gets to one of your
10 questions of how many should be the end in different levels,
11 and should there be a certain minimum number within a
12 certain level. I think that is a very valid question.
13 DR. JANOSKY: That is actually the issue that
14 I'm grappling with, which is the issue that I asked the
15 sponsor this morning, as to what do I do with those values
16 of say 350 and above, where there are very few samples
17 taken there, and you might have the expectation of getting
18 a lot of measurements in that area.
19 DR. CLEMENT: In clinical practice you would.
20 DR. JANOSKY: You would, right. That is what
21 I am especially thinking about with an individual who has
22 diabetes, or a group of individuals that have diabetes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
252
1 So going along that vein, in terms of that
2 upper end, it seems to me that it would be very reasonable
3 to -- and again, we could hedge our bet and we can find
4 the type of patient or find the type of population that
5 most likely would have those values before you finger stick
6 and actually take a look at them. That would probably
7 be the way that I would advocate that you look at that
8 upper end.
9 Is there something that you know that I don't
10 know that doesn't make sense in that respect?
11 DR. IRONY: No. I agree.
12 DR. JANOSKY: Okay. Then also talk about
13 that bias or under estimation that you had presented and
14 I had asked the sponsor about this morning. That coupled
15 with what we had just talked about a few seconds ago was
16 also telling me that then you have the probability of a
17 mis-diagnosis, diagnosis being a high triglyceride value.
18 You were talking about an estimate of I think
19 within there about 51 point differences.
20 DR. IRONY: At the level 400?
21 DR. JANOSKY: Right. Exactly.
22 DR. IRONY: Yes, around that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
253
1 DR. JANOSKY: So then you have a large
2 percentage that are anywhere from let's say 380 and above,
3 that might be --
4 DR. IRONY: Different.
5 DR. JANOSKY: That's right. Exactly. So
6 you are talking about missing a percentage of those that
7 we would consider within, refer to a physician.
8 Do you have an estimate as to how many those
9 would actually be in the population? What percentage
10 would be missed?
11 DR. IRONY: I guess we will need more data
12 in that region to estimate that. As I said, we have so
13 few in that region, that it's difficult to estimate from
14 the points we have.
15 DR. JANOSKY: But you didn't do any
16 calculations to see how many would be missed in terms of
17 saying 390, 400, and above, refer on?
18 DR. IRONY: No.
19 DR. JANOSKY: Okay. I didn't have the data,
20 so I would have played around with it to just see. I think
21 for now that probably would be the questions.
22 CHAIRPERSON NIPPER: Thank you.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
254
1 Dr. Lewis, do you have any questions for the
2 FDA?
3 DR. LEWIS: Just one quick one. Not so much
4 one for the statistician. But how typical is it for FDA
5 to have a statement in the users instructions or
6 information that says, and I'm just quoting from this one,
7 "The instrument can be run by a consumer, (lay person),
8 in their own home." I'm not questioning the grammatics
9 there. "With accurate results 95 percent of the time."
10 Is that typical to have that kind of a
11 statement in one of these inserts? If so, how do you
12 anticipate that the lay user interprets that kind of a
13 statement?
14 MS. PINKOS: Characterizing the performance
15 in a lay user package insert is very challenging. There
16 are several documents that have been written describing
17 how to characterize that performance, but I think saying
18 something like 95 percent of the time it's accurate, if
19 that were the case, you know, having something that's
20 general like that might be acceptable for a lay user
21 labeling.
22 We try to, for instance, stay away from
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
255
1 regression analysis or anything like that because a lay
2 user is just not going to understand that. So it's always
3 challenging, based on the performance, to try to capture
4 an easy user-friendly way of telling a consumer what kind
5 of performance they can expect.
6 I mean sometimes people will say 95 percent
7 of the time I agreed within 10 percent of a professional
8 device, or something like that, because people might be
9 able to understand something like that.
10 I don't know whether that answers your
11 question or not.
12 DR. LEWIS: Probably not completely because
13 I am thinking of myself as that lay user and say without
14 having any statistics understanding.
15 MS. PINKOS: Right. To say 95 percent of the
16 time you are going to get an accurate result, you would
17 still have to have some kind of measure of what accurate
18 meant.
19 DR. LEWIS: Or as the lay user, I might say
20 well, if I do this once more and there again have a 95
21 percent accurate chance, I have done it twice. So that
22 brings me pretty darn -- and is that the appropriate
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
256
1 interpretation of that? Obviously not. I don't know what
2 you would better do, but it just seems that that leaves
3 a lot to be desired.
4 MR. GUTMAN: Let me interject, because it's
5 a real challenge. Again, in the course of the discussion
6 of this product, if you have suggestions, we would be
7 grateful.
8 We key off of an NCCLS document for labeling
9 of home use. That document uses the Galin and Gambino
10 context. Actually it's not for continuous stat. It's
11 for nominal data. But it uses essentially their term
12 "efficiency." Then it converts that into accuracy and
13 uses it for lay use so that it's essentially the positive
14 and negative agreement. We have been using that, I mean
15 that NCCLS document has been out for a decade.
16 We actually have it in the context of an
17 existing guidance we were thinking about issuing. We are
18 kind of hung up about it and wondering if that's the best
19 we can do in 1999, because efficiency isn't accuracy.
20 But the trick is, how do you couch in an intelligent way
21 for consumers how reliable a product is. I don't know
22 that we have come up with something better than that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
257
1 Certainly we don't like to get too embroiled in accuracy
2 and precision data because it goes right over most people's
3 heads.
4 So if you don't have suggestions today, but
5 would like to throw something at us post-panel, as you
6 are riding home and thinking about this, we probably could
7 use some help in this area, not just for this submission,
8 but for other akin submissions.
9 DR. LEWIS: Thank you.
10 CHAIRPERSON NIPPER: Ms. Kruger, do you have
11 any questions for the FDA?
12 MS. KRUGER: Yes. I have one for Ms. Pinkos.
13 You had said that there were still some issues with short
14 sampling technique and in terms of accuracy. When I went
15 back and looked at the materials that we were handed, it
16 basically says that if you have a volume of 15 to 30
17 microliters, that it's statistically and clinically
18 indistinguishable, and they used a 200 milligram per
19 deciliter triglyceride, and basically those would be okay.
20 When you look at the data and you go down to 9 or even
21 11, it's 103 to 104, is what you are going to get.
22 My question is, and I know there are some
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
258
1 backups that patients may or may not do, but my question
2 is, at some point the meter will not start if there is
3 not enough volume. I am not sure what that is, and I think
4 that would be important to know, given that at 9
5 microliters, a 200 specimen is only reading 103.
6 MS. PINKOS: I agree. Can you bring some
7 clarification that? I was assuming that -- I can't recall
8 the data set exactly, but at that lowest point that was
9 presented in that study, was that the point at which the
10 instrument read a short sample and wouldn't get a readout?
11 Do you remember that, Margo?
12 CHAIRPERSON NIPPER: Come to the microphone,
13 please, Mr. Connolly.
14 MR. CONNOLLY: The fellow that could answer
15 that question the best is no longer with the company, so
16 the five of us are sitting here struggling, trying to figure
17 out how they measure short samples right now. I don't
18 think we can come up -- we could come up with a guess,
19 but I don't think we have the correct answer.
20 There is software and there is methods in
21 place via one of the other wavelengths to measure and
22 distinguish between whole blood and controls for QC
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
259
1 purposes. Those same algorithms can be used to determine
2 sample volumes, but I can't tell you if they are in use
3 or how they work.
4 MS. PINKOS: We consider the critical element
5 of that study, the fact that the short sample mechanism
6 did not keep the device from giving you an incorrect result
7 if a short sample was added to it. That is somewhat
8 disturbing. That is why they tried to counteract that
9 by having them turn the strip over. But that is a little
10 bit of a concern, is well, because to turn the strip over
11 and say does the color look even, you know, that's somewhat
12 subjective too.
13 MS. KRUGER: When there is already a blood
14 glucose monitor on the market that does the same thing,
15 and patients just don't do it or they reapply samples.
16 It does truly affect the validity of the answer, the
17 accuracy of the answer you are getting. If the meter won't
18 start because the volume is too low or the meter won't
19 start until you have adequate volume, that's okay. So
20 that would be where my concern lies.
21 CHAIRPERSON NIPPER: Dr. Floyd?
22 DR. FLOYD: I have got a couple of questions
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
260
1 about the package inserts, which I went back and took
2 another look at after thinking about this over lunch.
3 One of the things that occurs to me as we're talking about
4 a test here that is intended, recommended in worst case
5 scenario to be used 12 times a year in a post-menopausal
6 female, but in reality as I'm understanding the
7 discussion, we're thinking may be used three to one time
8 a year, depending upon who you are listening to around
9 the discussion.
10 So my question is very simple. What is the
11 number of strips in a package? What is their shelf life?
12 There are instructions in the package insert as to storage
13 conditions, but it doesn't tell me what the shelf life
14 is.
15 So the question is, are they going to still
16 be good if you have got 12 in a package and you don't get
17 to them for four years down the road?
18 MR. CONNOLLY: You have to make money
19 somehow. There are six test strips in a package, and the
20 dating gets longer as the product gets more mature. Right
21 now, we would like to have a goal of 12 to 15 months.
22 We have seen stability in excess of 24 months.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
261
1 DR. PASQUA: Right now, we have stability of
2 six months, but we need to get more data. That's all I
3 am willing to go. These are ongoing. Right now, I can
4 just say six months.
5 CHAIRPERSON NIPPER: Are there any
6 questions? Dr. Rosenbloom, do you have any questions
7 specifically for the FDA about their presentation? We
8 can ask the sponsor questions after we get around the table
9 once, and we can go to open public hearing. Then we can
10 throw it open to anybody.
11 DR. ROSENBLOOM: Yes. The presentation on
12 performance goals.
13 CHAIRPERSON NIPPER: How about getting to the
14 microphone?
15 DR. ROSENBLOOM: Oh, I'm sorry.
16 CHAIRPERSON NIPPER: That's okay.
17 DR. ROSENBLOOM: On the performance goals,
18 the slide on interference, the discussion on interference
19 only talked about ascorbic acid, elevated hematocrit
20 levels and cholesterol. What happened to the drugs?
21 MS. PINKOS: They just didn't make it onto
22 the slide. That wasn't to slight them.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
262
1 DR. ROSENBLOOM: Okay. I was wondering why,
2 if that had been forgotten.
3 CHAIRPERSON NIPPER: Thanks. Any others?
4 Dr. Clement?
5 DR. CLEMENT: I'm confused about the
6 precision testing. I am going over this document over
7 and over. This is the one that was given to us, TZI-2.
8 It talks about precision testing. The precision study
9 says, "precision testing should be done under reasonable
10 conditions of use, i.e. conditions under which the regular
11 patient will use the device."
12 Then we are presented with the data where we
13 get a "consumer" we're basically testing their pipet
14 technique. You don't teach patients to draw their blood
15 at home, and then pipet it in to see how good they are
16 in pipeting to get the same results on the device.
17 So is this a valid test of precision?
18 MS. PINKOS: This is a question that we
19 grapple with for whole blood devices because there is no
20 way, for instance, to characterize day to day usage on
21 a device like this because of the problem with the matrix.
22 When you have a test strip such as this, the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
263
1 reproducability is really -- it's not like you have a
2 reagent system where the reagents are deteriorating or
3 the calibration is drifting out of place. Those are pretty
4 much set.
5 So what we are really looking for is
6 characterization of the technique as you described it,
7 captured it. The technique for the patient getting a
8 finger stick and pipetting it on. Not pipetting it on,
9 but getting a finger stick and analyzing it. But you can't
10 do a finger stick blood and get precision. So the closest
11 thing that we can get is having whole blood samples by
12 a group of consumers repeatedly run.
13 If you have got any suggestions, because we
14 grapple with this with glucose meters as well. It's
15 difficult to characterize precision for whole blood
16 analyzers. You have to get it -- you know, the best thing
17 to do, we think, is perhaps just having a lot of users
18 do the test, because day-to-day has no meaning, and you
19 can't do it.
20 DR. CLEMENT: But even, I'm thinking like for
21 a YSI machine, if you are doing whole blood, there's even
22 techniques to get accurate precision with that. You have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
264
1 to mix it, make sure it's air rated. If it's not air rated,
2 then your numbers can drift considerably. We are having
3 "consumers" who are never trained to do this that may be
4 taking samples over time that this data may be falsely
5 high.
6 For example, compared to if they did a
7 multiple serial fingerstick test very frequently over 10
8 minutes during the fasting state when they are fasting
9 triglycerides. In theory would be the same, or you could
10 check a fasting triglyceride at the beginning of the series
11 of tests at the end of the series of tests and make sure
12 it is steady, and then average those two out and make sure
13 it hasn't drifted upwards and downwards during the start
14 of the beginning of the test.
15 I mean potentially you can measure precision
16 that way directly from a patient's finger to the test.
17 It just seems kind of crazy, actually, because there's
18 techniques on pipetting it out if it's not mixed and all
19 these other things that may actually make these numbers
20 higher than they potentially really are.
21 MS. PINKOS: You're right. It's a tradeoff.
22 CHAIRPERSON NIPPER: Dr. Kimberly?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
265
1 MS. KIMBERLY: I don't have any questions.
2 CHAIRPERSON NIPPER: Dr. Rifai, did you have
3 a question for the FDA?
4 DR. RIFAI: I just want to add to the
5 interference study that you showed. We were told this
6 morning -- when I read the information here, I thought
7 that usually the red cells lyse, and there's a way to
8 capture the hemoglobin prevented from going down to the
9 reaction. But we were told this morning that actually
10 capture the cells to prevent them from going down to the
11 reaction.
12 Is hemoglobin or checking for interference
13 from hemoglobin was on the list?
14 MS. PINKOS: I think that was on the list,
15 wasn't it?
16 DR. RIFAI: Because it was not on the list
17 that we got.
18 MS. PINKOS: I believe it was done.
19 CHAIRPERSON NIPPER: I appreciate the FDA's
20 forbearance while we ask all sorts of questions. Thank
21 you very much. Why don't you all step back and let me
22 open the meeting as an open public hearing this afternoon.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
266
1 We still have some time allotted for interested persons
2 who may address the panel and present information relevant
3 to the agenda. Speakers are asked to state whether or
4 not they have any financial involvement with the
5 manufacturer of the product being discussed or with their
6 competitors.
7 Okay. Hearing none, seeing no one ask for
8 recognition --
9 DR. DOUMAS: May I ask for recognition?
10 CHAIRPERSON NIPPER: I think it's meant for
11 people outside the panel. I'll get back to you, Basil.
12 DR. DOUMAS: I just looked at something which
13 could explain something.
14 CHAIRPERSON NIPPER: Okay. We will get back
15 to that in an open committee discussion in just a minute.
16 Why don't we use the open committee discussion
17 meeting to try to get some consensus on the review
18 questions. Then if comments that are appropriate come
19 up as we go around the room or questions come up that we
20 need to call on the FDA or the sponsor, we can do that,
21 because we don't want time to get away from us and we would
22 like very much to have this panel reach consensus as much
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
267
1 as possible on the review questions.
2 Therefore, if we go around the room and ask
3 for questions, ask for answers to questions, let's be as
4 concise as possible, and stick to the question. Or if
5 you have something that just can't wait, go ahead and ask
6 it when the floor is yours.
7 To take Dr. Rej off the hook, we are going
8 to go around the room this way this time, with question
9 number one. Question number one is, is the Agency's
10 requirement for comparison to a CDC reference laboratory
11 reasonable when the device is for over-the-counter use?
12 We might want to display these questions while we're
13 thinking about it. I apologize for not alerting you to
14 the fact we were going to go to the questions right away.
15 I'll give you a couple of seconds to put them up.
16 What do you think? Is the Agency's
17 requirement for comparison to a CDC reference laboratory
18 reasonable in the devices for OTC use?
19 DR. RIFAI: I think so. I believe so. I
20 don't see why you should have different standards if the
21 test is going to be done OTC versus as a test that is going
22 to be in the laboratory.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
268
1 Unfortunately, the issue for triglyceride is
2 a little bit more complex than let's say if you are
3 measuring cholesterol because triglyceride, as done by
4 the CDC method, as we alluded to earlier this morning,
5 is with correction for endogenous glycerol, and the great
6 majority of labs around the country do not correct for
7 endogenous glycerol.
8 Here, although you are sending the samples
9 to a CDC certified laboratory, from the data that's used
10 in the comparison study, is the one that included the
11 endogenous glycerol. So it's not so-called traceable to
12 the CDC method.
13 One would argue that at least if you go to
14 a CDC-certified lab, they will have tighter measure or
15 small measure of variability than you go to other lab,
16 where you set it up yourself.
17 CHAIRPERSON NIPPER: Thank you.
18 Dr. Kimberly, do you have a -- you are from
19 a CDC laboratory.
20 MS. KIMBERLY: I certainly do have an
21 opinion. I believe absolutely that an OTC device should
22 be compared to the CDC reference method. I would like
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
269
1 to point out though that the method that was used at Pacific
2 Biometrics is an enzymatic method that is a method that
3 is monitored monthly by CDC because we do that for all
4 of the cholesterol method lab network participants.
5 While we are evaluating them on their glycerol
6 blank method, we at CDC have a method in-house, where we
7 can determine the free glycerol in the reference materials
8 that we're using so that we can get an idea of how these
9 CDC standardized labs are doing on their total
10 triglyceride methods.
11 Another thing, concern I have about this
12 particular device is that they are calibrating against
13 their predicate device. I think that it would be important
14 to calibrate with the reference method or by a standardized
15 method. I think that's something that the sponsor should
16 take into consideration as they proceed with this.
17 CHAIRPERSON NIPPER: Dr. Clement, what do you
18 think about question one?
19 DR. CLEMENT: I will defer to my expert
20 analytical colleague. So I would say yes.
21 CHAIRPERSON NIPPER: Dr. Rosenbloom, how
22 about you?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
270
1 DR. ROSENBLOOM: I said yes.
2 CHAIRPERSON NIPPER: Okay. How about you,
3 Dr. Floyd?
4 DR. FLOYD: I would say yes. I bring it up
5 from a different issue. In this case, over-the-counter
6 use, when I see that term, it's for tests that in general
7 we have sort of assumed the consumer interprets for
8 themselves.
9 As I interpret the remarks we have heard
10 today, what we are really talking about in this test is
11 a number that the consumer collects and then reports to
12 their attending physician, we hope. If in point of fact
13 the number doesn't correspond to anything the physician
14 is used to hearing, how can they interpret it? So I would
15 say yes, it has to be pegged to some kind of reference.
16 CHAIRPERSON NIPPER: Ms. Kruger?
17 MS. KIMBERLY: Yes.
18 CHAIRPERSON NIPPER: Dr. Lewis?
19 DR. LEWIS: Yes.
20 CHAIRPERSON NIPPER: Let the record show that
21 Dr. Janosky is nodding yes.
22 DR. JANOSKY: I did say yes also.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
271
1 CHAIRPERSON NIPPER: Dr. Doumas?
2 DR. DOUMAS: Yes.
3 CHAIRPERSON NIPPER: Dr. Manno?
4 DR. MANNO: Yes.
5 CHAIRPERSON NIPPER: Dr. Everett?
6 DR. EVERETT: Yes.
7 CHAIRPERSON NIPPER: Dr. Rej?
8 DR. REJ: Yes. I think there must be
9 traceability to the system somehow. I think direct
10 comparison of the whole blood device with a reference
11 method has some problems because you are actually using
12 a different physical sample even if they are taken at nearly
13 identical times. So the comparison is not quite as easy
14 as if you have a plasma or serum versus serum measurement.
15 So I think that this system that's in place has done a
16 lot of good.
17 I know that most of you know that my laboratory
18 is one of the CDC reference laboratories. I pointed out
19 this to the FDA. We're not accepting any samples from
20 manufacturers. We're not doing certification for
21 triglyceride. So there is no conflict of interest. But
22 I think it is a very important system. Our involvement
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
272
1 is mainly for an accuracy base for our proficiency testing
2 program. So I think the whole system is worthwhile. I
3 think that OTC devices should not be exempt from it.
4 But again, I am not sure that direct
5 comparison of the OTC device with a plasma sample done
6 by the CDC certified laboratories is absolutely necessary.
7 There has to be some traceability.
8 CHAIRPERSON NIPPER: My opinion is that the
9 answer to question one should be yes. The question says
10 requirement for comparison. It does not say requirement
11 for equivalency or it doesn't imply stringency, but it
12 implies traceability. I think it is very important, as
13 Dr. Rej says. I know Dr. Rej and Dr. Doumas and others
14 in the room, including myself, who have served on standards
15 committees for the AECC in times past and we have seen
16 a history of improvement if we can trace to a benchmark.
17 I think that serves us well as a profession.
18 Question number two is should an
19 over-the-counter device be required to meet NCEP
20 performance goals, which is approaching the next issue.
21 The question is should the performance be in line with
22 other analytical systems. So let's go back around the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
273
1 room.
2 Dr. Rej, what do you think about that?
3 DR. REJ: Yes, with the caveat that was
4 mentioned I think by Ms. Pinkos, is that you could trade
5 off accuracy for precision, I think, especially in an
6 over-the-counter device. But meeting the NCEP goal for
7 total error of less than 15 percent is particularly
8 important. But I think one could trade off some precision
9 for accuracy in such a device.
10 CHAIRPERSON NIPPER: Now there is an inherent
11 second question in here, is if it doesn't meet the goals,
12 can you clear it for marketing with cautionary labeling?
13 DR. REJ: I don't believe so. I think the
14 goals -- you have to meet the NCEP goals for an analytical
15 technique.
16 CHAIRPERSON NIPPER: How about you, Dr.
17 Everett?
18 DR. EVERETT: My answer is yes, that it should
19 meet the goals.
20 CHAIRPERSON NIPPER: Okay.
21 DR. MANNO: It should meet the goals.
22 DR. DOUMAS: I tend to differ a little bit.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
274
1 I will go with Dr. Ginsberg. If it doesn't meet exactly,
2 it should be close. So I will leave a little bit of window.
3 Maybe 15 percent over the goals. That doesn't mean
4 another 15 percent, 15 of the 15. Probably you can go
5 to 17.5. That's my opinion.
6 CHAIRPERSON NIPPER: Thank you, Dr. Doumas.
7 Dr. Janosky?
8 DR. JANOSKY: Yes, to meeting the goals.
9 CHAIRPERSON NIPPER: Dr. Lewis?
10 DR. LEWIS: I am inclined to agree with Dr.
11 Doumas. In fact, before he made his comment, I was
12 thinking of something of that sort, whether it's 17.5
13 percent or whatever. But that there might be some
14 opportunity for a slight movement from that 15 percent
15 based on everything I have heard about the system and its
16 performance characteristics in this case.
17 CHAIRPERSON NIPPER: Ms. Kruger?
18 MS. KRUGER: I think it should meet its goals.
19 CHAIRPERSON NIPPER: Okay.
20 DR. FLOYD: Yes.
21 CHAIRPERSON NIPPER: Dr. Rosenbloom?
22 DR. ROSENBLOOM: I have said that I thought
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
275
1 that I was concerned that false security might arise from
2 an inappropriately low reading, and that that would be
3 more harmful than not testing, and that inappropriately
4 high readings would create needless anxiety or misguided
5 treatment changes. So the answer is yes, it should meet
6 the NCEP performance goals or some modest modification
7 of them.
8 CHAIRPERSON NIPPER: Dr. Clement?
9 DR. CLEMENT: Well, I'll be different from
10 the group. I'll say no, because I think this device could
11 be clinically useful if the labeling is appropriately
12 made.
13 Obviously I had issues in terms of the
14 precision testing, but I think -- I mean it sounds like
15 a major, an industry problem, and how do you measure
16 precision on whole blood. Obviously if that can be shown
17 to be reduced just by testing it a little bit differently,
18 then it may help narrow some of the total system error.
19 I mean I think labeling could be modified.
20 I even wrote out some labeling, which we can talk about
21 a little bit later. I think possibly a total system error
22 of plus or minus 20 percent would be clinically useful.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
276
1 CHAIRPERSON NIPPER: Okay..
2 MS. KIMBERLY: I'll say yes, they should meet
3 the goals.
4 CHAIRPERSON NIPPER: Did you need to say
5 something, Mr. Connolly?
6 MR. CONNOLLY: No. I'm just making some
7 notes.
8 CHAIRPERSON NIPPER: Okay. I'm not sure
9 why, but I think that we need to step back from the table.
10 I don't know why we need to do that, but that is kind
11 of a rule at these things. Thank you.
12 Dr. Kimberly, did you have an opinion about
13 question two?
14 MS. KIMBERLY: Yes. I said that they should
15 meet the goals.
16 CHAIRPERSON NIPPER: Okay. I'm sorry. I
17 was preoccupied.
18 Dr. Rifai?
19 DR. RIFAI: Yes. I think they should meet
20 the goal. I have some concern about reaching by consensus
21 how much extra error you will allow a particular device
22 to exceed. I think if a particular device will have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
277
1 different utility than the laboratory test in a different
2 setting, then you have to demonstrate somehow by a study,
3 not -- by consensus how much error you can tolerate.
4 I think for the general public, if you add
5 a label, cautionary label about the error, I am not sure
6 really it's going to be -- the public is going to understand
7 it.
8 CHAIRPERSON NIPPER: Thank you.
9 Dr. Rifai took the words out of my mouth.
10 My problem with asking this group to deviate or to move
11 away from the NCEP goals is that I would like to see some
12 evidence that good medicine will result by moving away
13 from those goals. Not necessarily that the goals are too
14 stringent or that someone else got a product in that somehow
15 didn't meet the goals five years ago, or that -- I would
16 just like to see the fact that the claims for improved
17 public health can be met by this device. I don't see that
18 study.
19 So that is why I think that if you are not
20 going to come in and make a compelling case for marketing
21 the device without meeting the goals, then I think you
22 should meet the goals.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
278
1 Okay. Question three. Do we know what the
2 appropriate minimum sample size for evaluating an OTC
3 lipid test is. Should there be minimum requirements for
4 sample distribution?
5 You took my answer away from me, so you get
6 on the hot seat again.
7 DR. RIFAI: I don't know about the
8 appropriate minimum sample size, but for the distribution,
9 I think it must encompass everything you are going to be
10 seeing clinically, with the special emphasis on the area
11 where the cutpoints for clinical decision making usually
12 are.
13 CHAIRPERSON NIPPER: Thank you.
14 MS. KIMBERLY: I think that statisticians
15 could probably adequately answer the question about a
16 minimum sample size. That would depend on the difference
17 to detect and the precision of the instrument. But
18 ideally, I would think that the difference to detect would
19 be based on the MCEP bias recommendations. So I think
20 that giving a number there would depend on the statistics
21 and the math there.
22 As far as minimum requirements for sample
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
279
1 distribution, the NCCOS has recommended sample
2 distribution guidelines in their EP9, which is a method
3 comparison and bias estimation protocol. I think there
4 are some guidelines out there that could be used. But
5 certainly you definitely want to cover the medical
6 decision points that are relevant to the particular
7 analyte.
8 CHAIRPERSON NIPPER: Dr. Clement?
9 DR. CLEMENT: I agree with the previous.
10 Clearly, it needs more numbers in the 400 to 500. I would
11 say at least 20 within the 400-500 end, if we had to come
12 up with a number for measuring accuracy.
13 CHAIRPERSON NIPPER: Is the 20 an arbitrary
14 guess or did you base that on anything other than seat
15 of the pants?
16 DR. CLEMENT: Pure arbitrary guess, that
17 could be validated by the statisticians.
18 CHAIRPERSON NIPPER: Okay.
19 Dr. Rosenbloom?
20 DR. ROSENBLOOM: Yes. I agree. I would
21 certainly like to see more values in the range of clinical
22 utility that we live in.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
280
1 DR. FLOYD: I agree with the previous
2 comments.
3 CHAIRPERSON NIPPER: Okay. Ms. Kruger?
4 MS. KRUGER: Same.
5 CHAIRPERSON NIPPER: Dr. Lewis?
6 DR. LEWIS: I agree.
7 CHAIRPERSON NIPPER: Dr. Janosky?
8 DR. JANOSKY: I agree, but I would add that
9 it should actually be powered within each of those ranges
10 of clinical decision making. My major concern, as I have
11 voiced today, was the 400 and above. But also, there's
12 other clinical decision points along the way. So I would
13 categorize the range and then power within each of those
14 categorizations.
15 CHAIRPERSON NIPPER: Dr. Doumas?
16 DR. DOUMAS: Oh yes. I think they can
17 consult an NCCLS document. It specifies there.
18 CHAIRPERSON NIPPER: Dr. Manno?
19 DR. MANNO: I agree.
20 CHAIRPERSON NIPPER: Dr. Everett?
21 DR. EVERETT: I agree. There are a number
22 of biostatistical calculations to determine sample size
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
281
1 based on the study and how the study is to be done. I
2 would recommend they select one of those and follow it
3 through as opposed to what appears here, is that the studies
4 are done backwards. The studies are done first, and then
5 the statistical tests are picked at the end. That's
6 inappropriate. I suggest they just follow one of those
7 and continue it all the way through.
8 CHAIRPERSON NIPPER: Thank you.
9 Dr. Rej?
10 DR. REJ: I agree with Drs. Kimberly and
11 Janosky regarding both points, especially at least some
12 guidance can be given by the NCCLS guideline in terms of
13 distribution of samples for this particular case.
14 CHAIRPERSON NIPPER: I like the NCCLS
15 approach too.
16 The follow-up question to this, which many
17 of you have already answered both before this round of
18 questioning as well as during this round of questioning,
19 is has the sponsor done the appropriate precision and
20 interference studies? If anyone would like to elaborate
21 on that, we can go around the room again, starting with
22 Dr. Rej, if you want to nail down any addition information.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
282
1 DR. REJ: I think the points made by Dr.
2 Doumas are relevant, that whole blood devices are not
3 exactly the same as traditional chemistry device. Even
4 the sponsor alluded to that. And that interference
5 studies done in highly diluted samples don't necessarily
6 turn out the same way with neat samples, even though the
7 basic chemistry for measurement is similar.
8 I think that since this is intended for
9 over-the-counter use and uses individual disposable
10 devices for each measurement, three persons is nowhere
11 near adequate to demonstrate precision of the test in the
12 hands of the consumer.
13 CHAIRPERSON NIPPER: How about
14 interferences?
15 DR. REJ: Again, I agree with Dr. Doumas.
16 In terms of specific interferences, again, it depends on
17 the intended use. It was alluded to that this would be
18 used largely by diabetics as an adjunct to their treatment
19 for diabetes. But that's not spelled out in the package
20 insert, that it's really for anybody. So limiting it to
21 the drugs either lipid lowering drugs or other drugs taken
22 by diabetics, I don't think is -- that's certainly the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
283
1 minimal list for drug interference.
2 Again, I think there are some general
3 recommendations on the types of drugs, at least certain
4 classes of drugs that can be used. You don't have to check
5 each and every one, but that certain generic classes
6 certainly. They did some studies with reducing agents
7 like ascorbate, but I think other drugs need to be
8 considered.
9 Just falling back on the literature to more
10 traditional wet chemistry may not apply in this case.
11 CHAIRPERSON NIPPER: Dr. Everett?
12 DR. EVERETT: Yes. I agree. Those were my
13 points exactly.
14 CHAIRPERSON NIPPER: Okay.
15 DR. EVERETT: They just have to do the rest
16 of the work.
17 CHAIRPERSON NIPPER: Dr. Manno?
18 DR. MANNO: I would like to add something to
19 the two previous comments, that if any more work is done,
20 and it should be done on the interference studies with
21 relation to the drugs, but if you are using whole blood
22 from people who are taking drugs, you are not only going
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
284
1 to be finding parent drug, in many instances, you are going
2 to be finding a variety of metabolites. So you are going
3 to have to have some feel for interference there, perhaps.
4 I am still not clear in my own mind. This
5 goes a little bit to what I asked this morning in terms
6 of interference in the response that I got about looking
7 at pore size. When I was told that there are fragments
8 of red cells that get through and the like, I would like
9 to ask the sponsor if they used any other pore size for
10 their strip. Have they looked at that to get the optimum
11 pore size?
12 Because my experience with doing things like
13 this is such that you can go lower and get rid of some
14 of the cellular components by going to a smaller pore size.
15 So if you could address that for me, I would appreciate
16 it.
17 MR. CONNOLLY: We said the right thing this
18 morning, but it was the wrong answer. The pore size of
19 the reaction membrane has nothing to do with exclusion
20 of the red blood cells.
21 DR. MANNO: But that's what is provided to
22 us, because this goes back to my original question this
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
285
1 morning. Are you really testing whole blood or are you
2 testing the aqueous component?
3 MR. CONNOLLY: We are testing the aqueous
4 component, mostly.
5 DR. MANNO: Then we have got some definitions
6 here that need to be clarified in my mind in describing
7 the reactions that are taking place and the intended use.
8 It also may present us with, in truth, we're talking about
9 the same matrix when we're talking about serum as a matrix
10 for the comparison here. So in terms of those studies,
11 I think that that needs to be clarified.
12 MR. CONNOLLY: That's your opinion. I'm just
13 answering the pore size issue. That had nothing to do
14 with exclusion of red blood cells.
15 CHAIRPERSON NIPPER: Okay.
16 DR. MANNO: I have no more questions. Thank
17 you.
18 CHAIRPERSON NIPPER: Dr. Doumas?
19 DR. DOUMAS: I agree with the others about
20 omission of drugs that are taken. However, we missed
21 something. I apologize that I missed this. If you look
22 at interference by ascorbic acid even at the lowest level,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
286
1 that's a 10 percent inaccuracy, negative bias, when only
2 allow five percent.
3 Now .75 milligrams per deciliter of ascorbic
4 acid, you will not find it in people who do not get extra
5 vitamin C. But those who follow the recommendations of
6 the late Dr. Linus Pauling, they may easily get to that
7 level or even higher. So you want 1.5, that's a 20 percent
8 bias already.
9 I want to emphasize again what I said in the
10 morning, that samples here are not diluted like in wet
11 chemistry. So interference, concentration that will
12 interfere is much higher for the same level in this type
13 of chemistry, dry chemistry, as it is in the wet chemistry.
14 So .75 milligrams may not have the same effect in any
15 other instant, I mean non-dry chemistry. Probably this
16 is the reason, you see 11 percent bias with a 0.75.
17 I would suggest or maybe this can explain all
18 the negative bias that we have seen, because a lot of people
19 swallow ascorbic acid, 500 milligrams or a gram daily to
20 protect themselves against some cold.
21 CHAIRPERSON NIPPER: Thank you. Dr.
22 Janosky?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
287
1 DR. JANOSKY: I have nothing to add.
2 CHAIRPERSON NIPPER: Dr. Lewis?
3 DR. LEWIS: Nor I.
4 CHAIRPERSON NIPPER: Have they done
5 appropriate precision in interference studies?
6 DR. JANOSKY: I have nothing to add. I defer
7 to my colleague.
8 CHAIRPERSON NIPPER: Okay.
9 DR. LEWIS: I thought I had nothing to add,
10 but to answer the question as posed, I would have to say
11 no.
12 CHAIRPERSON NIPPER: Okay, thank you.
13 Ms. Kruger?
14 MS. KRUGER: I would agree.
15 CHAIRPERSON NIPPER: Okay. Dr. Floyd?
16 DR. FLOYD: I would agree with the previous
17 comments.
18 CHAIRPERSON NIPPER: Thank you.
19 Dr. Rosenbloom?
20 DR. ROSENBLOOM: Yes. I am also concerned
21 about the real life situation with people with these drugs
22 circulating.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
288
1 CHAIRPERSON NIPPER: Thank you.
2 DR. CLEMENT: I was going to mention on the
3 precision data, I think it's unfair to measure precision
4 and putting so much weight on it, on having untrained
5 "consumers" whoever they are, doing this. If we are going
6 to measure in that way, we should have trained people
7 measuring samples using oxygenated whole blood.
8 CHAIRPERSON NIPPER: Thank you.
9 Dr. Kimberly?
10 MS. KIMBERLY: I defer to my colleagues who
11 have already spoken.
12 CHAIRPERSON NIPPER: Okay. Dr. Rifai?
13 DR. RIFAI: I just want to add one thing about
14 the other interference studies that were suggested by
15 others like checking for interference from fibrate,
16 stattons, even diuretics, whatever type of drugs that
17 patient will be taking.
18 I think Dr. Manno suggested that taken from
19 patients who are taking these medications because you want
20 to not look at the effect of the parent drugs, but you
21 want to look at the effect of the metabolites.
22 Unfortunately, in these particular drugs, the drugs affect
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
289
1 the levels of triglyceride. So it would be very difficult
2 to determine if that's the effect of the drug or that's
3 the effect of the interference. Probably the only way
4 to check for these drugs is just by spiking experiment,
5 and that would be the best way to do it because I can not
6 see any other way.
7 CHAIRPERSON NIPPER: Dr. Gutman?
8 MR. GUTMAN: Yes. I would like to actually
9 raise that because there is a side issue that Ms. Pinkos
10 raised that's relevant, frankly, to the issue of spiking,
11 which is a traditional way of doing things, again, building
12 off of an NCCLS document. That is, the sponsors propose
13 that cholesterol interference be done by looking at
14 regression testing, comparing triglycerides with
15 cholesterol levels. We have not seen that either ever
16 or certainly not seen that very often. I was wondering
17 if anyone on the panel would like to comment as to whether
18 that is an insightful approach or whether in fact we should
19 ask the sponsor to go back and do more standard NCCLS type
20 spiking experiments with lipids.
21 CHAIRPERSON NIPPER: Does anybody on the
22 panel want to tackle that question?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
290
1 DR. REJ: It's not so easy to do spiking
2 experiments with cholesterol. It may be unusual, but it
3 seemed a reasonable way to do it. I didn't find any fault
4 with that. Although I did notice that there was a bias
5 in the actual cholesterol measurement that exceeded NCEP
6 guidelines.
7 CHAIRPERSON NIPPER: Does anyone on the panel
8 have anything to add to what Dr. Rej said?
9 Thank you. I agree with that.
10 I had a question that may not be an
11 interference study, but I had a question for the sponsor.
12 I am curious to know whether the actual signal is taken
13 as a reaction rate or whether it's taken as what we
14 analytical chemists called an endpoint, whether you waited
15 for a maximum color to develop and then you measured say
16 a single measurement, or whether you measured a series
17 of points as the reaction occurred.
18 MR. CONNOLLY: They are all endpoint values.
19 CHAIRPERSON NIPPER: Okay. That reassures
20 me a little bit about maybe a substrate depletion turning
21 out to be a very low triglyceride measurement.
22 Question number five, please. What's the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
291
1 appropriate claim for this device, and has that claim been
2 captured in the label?
3 I can't remember where I started last time.
4 Did I zing you, Dr. Rifai? Why don't we mix it all up,
5 and we'll start with Dr. Floyd.
6 DR. FLOYD: I would have to say that I'm not
7 sure in my own mind, and maybe it's because I missed it
8 as I read the thing. But in my mind, I am not sure if
9 the appropriate claim has been captured in the labeling
10 for this device.
11 CHAIRPERSON NIPPER: We're talking about
12 intended use here. I checked with my co-colleague here.
13 Intended use, according to the document I have is this
14 test measures triglycerides in fingerstick blood. Then
15 it goes on to say why. Then this test can be run by any
16 person -- who -- this test can be run by any person at
17 home as easily as in the doctor's office or a hospital.
18 When, run this test two to three times a year or at least
19 monthly if you are diabetic or a woman past menopause.
20 Did I get it right?
21 DR. FLOYD: I'll stick by my comments. It's
22 not clear to me that the average consumer is going to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
292
1 understand what that means.
2 MS. KRUGER: I think they have got part of
3 that in the claim. I think we have discussed at length
4 the issue of post-menopausal women, that probably that
5 needs to be rethought.
6 As far as the claim, I think that we have made
7 a lot of comments that it appears that the sponsor is
8 comfortable taking back and going back and looking at their
9 labeling. So I guess I would say that not really to the
10 first, and no to the second. But that there has been a
11 lot of discussion today, very positive discussion that
12 could be incorporated, that the labeling could be made
13 appropriate.
14 CHAIRPERSON NIPPER: Dr. Lewis?
15 DR. LEWIS: I'm inclined to agree that with
16 all the discussion that's taken place regarding the
17 intended purpose, if the sponsor were to incorporate that
18 to some good measure, I would say yes. Has it been captured
19 in the labeling? When that's done and the labeling then
20 reflects it, I would say yes.
21 CHAIRPERSON NIPPER: Thank you.
22 DR. JANOSKY: The labeling as it stands, the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
293
1 answer is no. But with the flavor of the intention of
2 what I believe the panel has presented today, then yes.
3 CHAIRPERSON NIPPER: Thank you.
4 Dr. Doumas? How about appropriate claims and
5 has that claim been captured in the labeling?
6 DR. DOUMAS: I would say no in general because
7 I found the statements on procedure and accuracy very vague
8 and maybe not correct. I think the intended use, I
9 strongly disagree with what, as it's written right now.
10 CHAIRPERSON NIPPER: What would you think the
11 appropriate claim should be or do you have an opinion about
12 that?
13 DR. DOUMAS: I think as far as when to use,
14 I think the patient should consult his doctor when to use
15 that test. Second, about procedure and accuracy, I think
16 they should follow recommendation of the statistician,
17 who presented a very good case about what is the total
18 error.
19 CHAIRPERSON NIPPER: So to follow up that,
20 not to nail you to the floor on it, but if you believe
21 the patient should consult the doctor before they use the
22 test, it sounds to me like you are not comfortable with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
294
1 this being over the counter, or am I putting words? You
2 would like it to just be prescription only? Or do you
3 think that over-the-counter is okay on this?
4 DR. DOUMAS: I am not going to express an
5 opinion there. There is a difference when measuring
6 glucose and measuring triglyceride. Glucose, the patient
7 can take action or must take action sometimes if it's too
8 low or it's too high.
9 CHAIRPERSON NIPPER: Yes.
10 DR. DOUMAS: This one, what action the
11 patient is going to take? None. He has to go to the
12 doctor. I think if he needs to be tested, I think his
13 physician should tell him if and when.
14 CHAIRPERSON NIPPER: Okay. Thank you.
15 Dr. Manno?
16 DR. MANNO: I agree with Dr. Doumas.
17 CHAIRPERSON NIPPER: Dr. Everett?
18 DR. EVERETT: Well I think the claim is
19 appropriate, but I don't think it's really captured in
20 the labeling at all. I think it is too ambiguous and
21 misleading for someone with a seventh grade reading level.
22 CHAIRPERSON NIPPER: Okay. Dr. Rej?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
295
1 DR. REJ: In terms of the what, I think not,
2 because it fails by a long shot to meet NCEP guidelines.
3 Therefore, it does not measure triglycerides accurately
4 or precisely. The why, probably correct. The who, that
5 hasn't been demonstrated with only three individuals.
6 The when, there's been a lot of criticism from my clinical
7 colleagues. I'll defer to them.
8 CHAIRPERSON NIPPER: Thank you.
9 Dr. Rifai, are you ready?
10 DR. RIFAI: Yes. I am in total agreement
11 with what Dr. Rej said. Also, I don't know if you want
12 to even clarify under who, run by any person at home.
13 I mean you have five-year-olds at home. We don't want
14 them to run that test. So I wonder if you want to say
15 any person above a certain age or any adult. I don't'
16 know if that needs to be specified or not.
17 CHAIRPERSON NIPPER: Dr. Kimberly?
18 MS. KIMBERLY: I agree with Dr. Rej.
19 CHAIRPERSON NIPPER: Dr. Clement?
20 DR. CLEMENT: I think the labeling could be
21 changed. In fact, I wrote out something. Is it
22 appropriate to just read it on here?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
296
1 CHAIRPERSON NIPPER: I believe so.
2 DR. CLEMENT: In terms of the accuracy issue,
3 this is just a suggestion. It could possibly put the
4 results of the triglycerides measure can be plus or minus
5 X, whatever that is actually found to be, of the reference
6 method. Due to this variability in accuracy, the device
7 does not replace the lipoprotein profile as measured by
8 your healthcare provider. The results should be used as
9 a guide between visits to your healthcare provider for
10 monitoring the management of your condition. This
11 management could include diet, exercise, or the efficacy
12 of medication.
13 CHAIRPERSON NIPPER: Thank you.
14 Dr. Rosenbloom?
15 DR. ROSENBLOOM: I agree with the previous
16 comments from the laboratory experts. I think that that
17 wording is excellent.
18 The who is rather an ambiguous and strange
19 statement. This test can be run by any person at home
20 as easily in the doctor's office or hospital. Well any
21 person at home doesn't run tests in the doctor's office
22 or hospital, which is what this sentence says. So the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
297
1 grammar is bad, as well as being ambiguous. Of course
2 it is a very different as the recommended wording
3 indicates, it is a very different issue than what's done
4 in the doctor's office or the hospital. So I think that
5 whole thing of who is strange.
6 Then the when, of course, we have addressed
7 the inappropriateness. I think the when should be
8 entirely consistent, unless they can come up with
9 expertise, to suggest otherwise that the when should be
10 consistent with ADA guidelines for those people with
11 diabetes, and for the general public, should be consistent
12 with the NCEP guidelines. Of course the ADA guidelines
13 are based on NCEP guidelines. But I see no reason why
14 the availability of this device changes those
15 recommendations.
16 CHAIRPERSON NIPPER: Thank you, Dr.
17 Rosenbloom.
18 I believe that the claim is too broad. I
19 believe that the accuracy and precision of the device needs
20 to be improved. The sponsor made a statement this morning
21 as he began, that his company's goal was hospital quality
22 results in a dry phase. I think I captured that quote.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
298
1 I do not believe these are hospital-quality results.
2 I have no problem with dry phase. I would like to see
3 people in the field be able to get hospital quality results,
4 because I believe those have the chance of misleading
5 results fewer than we have with this device as it currently
6 stands.
7 I see no problem in targeting this device once
8 it's improved to the diabetic population or to people who
9 suspect they may have a lipoprotein abnormality or
10 hyperlipidemia or a cardiac risk. I believe it ought to
11 be tied much closer to physician's visits. So I think
12 the claims need to be really nailed down in the labeling
13 and really tightened up before this device is marketed.
14 Is the performance adequate to support the
15 intended use? Many of us have already begun to comment
16 on this. You want to tackle that one, Bob? I think you
17 have already --
18 DR. REJ: I think clearly, at least in my
19 mind, the answer is no. More than half the results from
20 the device relative to the reference lab are outside of
21 the NCEP guidelines.
22 CHAIRPERSON NIPPER: Dr. Everett?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
299
1 DR. EVERETT: I agree. Clearly not.
2 CHAIRPERSON NIPPER: Dr. Manno?
3 DR. MANNO: I concur.
4 CHAIRPERSON NIPPER: Dr. Doumas?
5 DR. DOUMAS: No, it is not.
6 CHAIRPERSON NIPPER: Dr. Janosky?
7 DR. JANOSKY: I agree that it is not.
8 CHAIRPERSON NIPPER: Dr. Lewis?
9 DR. LEWIS: I also agree that it's not.
10 CHAIRPERSON NIPPER: Ms. Kruger?
11 MS. KRUGER: It's not.
12 CHAIRPERSON NIPPER: Dr. Floyd?
13 DR. FLOYD: I agree.
14 CHAIRPERSON NIPPER: Thank you.
15 Dr. Rifai?
16 DR. RIFAI: Yes.
17 CHAIRPERSON NIPPER: You agree, okay.
18 Dr. Kimberly?
19 MS. KIMBERLY: When you say you agree, you
20 agree that it's not, okay. All right.
21 Yes. I agree. I have a lot of trouble with
22 that 95 percent accuracy claim. I don't really know what
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
300
1 that means or where that number came from.
2 CHAIRPERSON NIPPER: Dr. Clement?
3 DR. CLEMENT: It is obviously not an NCEP
4 guideline criteria. It is much too loose. But as I
5 mentioned before, I think for clinical purposes for a
6 home-use device that's used in comparison, as an adjunct
7 to an analytical measurement done at a physician's office,
8 that the total error could possibly be monitored to as
9 much as 20 percent.
10 CHAIRPERSON NIPPER: Dr. Rosenbloom?
11 DR. ROSENBLOOM: I concur with the criticism.
12 CHAIRPERSON NIPPER: As the device stands
13 with the intended use, the intended use that's in the
14 documentation, I have questions about whether the
15 performance is adequate to support that intended use.
16 Unlike our esteemed consultant, Dr. Ginsberg, I do not
17 support widespread screening for the purposes of
18 screening. I think it has tremendous unintended use in
19 lots of false positives, lots of false negatives. It
20 causes people to chase their tails, so to speak,
21 diagnostically, and go off in all sorts of paths that take
22 them on the geolyses type journeys that they shouldn't
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
301
1 be down.
2 I think that if we put a device in the field
3 that's for measurement of triglycerides, it ought to
4 measure it accurately. So therefore, I am assuming that
5 the intended use is what we say it should be. Then the
6 performance needs to be adequate to support it.
7 Okay. We have a specific question about
8 labeling. Is it an acceptable approach to refer lay users
9 to a second professional use package insert for additional
10 information, TG, quality control instructions?
11 What do you think, Dr. Rifai?
12 DR. RIFAI: I have a lot of problems with not
13 including the QC material and everything you need in one
14 package. I think most likely for the general public, that
15 strip that I put in to check for if electronically the
16 device is working properly will suffice. I don't think
17 they are going to go and buy the quality control material.
18 I think if you want to sell something as an
19 OTC device, it has to contain everything that the person
20 needs.
21 CHAIRPERSON NIPPER: Is it my understanding
22 that the quality control material is with the strips?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
302
1 This is directed at the sponsor. In other words, or do
2 you need to buy a separate quality control material?
3 MS. ENRIGHT: We don't package it with the
4 strips, mainly because there is an issue, a use issue.
5 There is not a good balance between the quality control
6 material and the strips. So logistically, what happens
7 with the customer, the consumer, is they end up having
8 a balance -- an imbalance between the strips and the
9 material. So from a logistical standpoint. They have
10 one that expires on them. So based on that, we don't
11 package them together because they will have different
12 expiration dates. So just to be customer friendly to the
13 consumer we do that. We do have the material available.
14 DR. RIFAI: That would further support the
15 argument that they will not buy it.
16 MS. ENRIGHT: That's always an issue. The
17 way the strips are packaged, if they run a package of six
18 strips, if they run it twice a year, we have two quality
19 control materials. If we have even six months dating,
20 six months dating, there is no reason not to run the quality
21 control material because of the dating of the strips.
22 So you have two quality control materials in one test,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
303
1 and run it twice a year. But if you have other suggestions
2 for our packaging, we'll certainly listen to those. But
3 that's our rationale.
4 CHAIRPERSON NIPPER: Thank you. I'm not
5 sure I interpreted that question the same way Dr. Rifai
6 did.
7 I interpreted that question, Dr. Rifai, as
8 referring to the fact that there were two package inserts
9 in with the strips. The first package insert was the one
10 you have on your lap there with your hands on it. Then
11 if you flip forward, there is a professional package insert
12 that has more information on it, two more pages. There's
13 two pages for the over-the-counter package insert, and
14 then right after that is the professional package insert.
15 What the question was referring to was is it
16 acceptable to have two different package inserts and then
17 expect the lay consumer to read the professional package
18 insert for quality control instructions.
19 Did I interpret that question correctly?
20 MR. GUTMAN: Yes. Actually, you have
21 combined, you have sort of combined question seven and
22 eight. If you wish to in fact combine seven and eight
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
304
1 and run through the panel looking at those as joint
2 questions, that might be an appropriate thing to do.
3 CHAIRPERSON NIPPER: That would be fine with
4 me, because the time is rolling here.
5 Nader, did you have anything to add to your
6 previous answer?
7 DR. RIFAI: Yes. I mean I prefer to have them
8 the same, on the same sheet, just small paragraph about
9 the quality control. First, how you do it to check for
10 electronic. Second, how to do it to check if the whole
11 system is working.
12 CHAIRPERSON NIPPER: Yes.
13 DR. RIFAI: Not to have two separate sheets.
14 CHAIRPERSON NIPPER: Okay. Dr. Kimberly?
15 MS. KIMBERLY: Regarding the question about
16 referring to a professional package insert, I don't think
17 that's really acceptable for a lay user. I think the lay
18 user needs to have everything right there in front of them,
19 in language that they can understand.
20 As far as question number eight, I thought
21 that the QC instructions -- is that the question? I
22 thought the QC instructions were a little confusing,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
305
1 especially about how often they should be run. In the
2 same section, they instruct the user to run QC at least
3 once a month, and in another sentence two lines down, it
4 says at the beginning of each day.
5 Also, they instruct the user to take care not
6 to contaminate the dropper tip, but give the lay user no
7 concept of where contamination can occur and how is
8 somebody who has never worked in a laboratory before
9 supposed to understand that kind of thing.
10 CHAIRPERSON NIPPER: Thank you very much.
11 Dr. Clement, do you have comments about
12 question seven and eight?
13 DR. CLEMENT: Yes. I agree with the previous
14 comments that working with patients, Murphy's Law gets
15 a new meaning, because they find even more creative ways
16 to abuse things. All the current table top or home devices
17 that I know of now, the modern ones, have the control
18 solutions in it, which makes it very convenient with the
19 patient.
20 If it's not physically packaged with the
21 control solution in some type of box, then the patient
22 is not going to do it.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
306
1 CHAIRPERSON NIPPER: Dr. Rosenbloom?
2 DR. ROSENBLOOM: Yes. Since I am going to
3 be leaving, I'll address everything I have got in the label
4 highlighted. First, there is under the intended use,
5 triglycerides is a fat. There is a verb number mismatch
6 there.
7 Then under item six, after reading the result
8 down at the bottom, it says "discard result and retest."
9 People will use the same strip and put more blood on it.
10 So it needs to say discard result and strip, and retest,
11 or just discard strip and retest. It isn't a matter of
12 discarding the result. You have got to discard the strip.
13 Then under precautions, it says "This test
14 system is designed for in vitro testing only, i.e. outside
15 the body testing." That is another strange statement.
16 Would people eat this stuff or use it to -- I don't know
17 what the purpose of that statement is. Is there any reason
18 why something like that has to be in there? I think it's
19 just confusing. It confused me, and I have had a little
20 more than a seventh grade education.
21 The term "prior to" should be replaced by
22 "before" which is English. "Prior to" is jargon. "Data
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
307
1 generated" is another jargon term under limitations in
2 the procedure. Then there's some typos like many of all
3 of the lipids under expected values.
4 Then with the issue of quality control, I
5 agree whole heartedly that people have to know what a value
6 means. It can be put in seventh grade English. For
7 example, when you have a value of such and such, it could
8 be anywhere between such and such. That is plain English,
9 and people would understand that.
10 We have that problem with blood glucose
11 meters. People think because they get a number, that
12 that's gospel. They have to be reminded that those things
13 can be off.
14 CHAIRPERSON NIPPER: Thank you, Dr.
15 Rosenbloom. Have a safe journey.
16 Dr. Floyd?
17 DR. FLOYD: I think the issue of two separate
18 package inserts is a problem, particularly since a lot
19 of emphasis has been made on the fact that the customer,
20 the consumer insert is supposed to be written at a seventh
21 grade level. Does that imply that the professional one
22 will be also? I think the package insert needs to be
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
308
1 rewritten embodying the comments that have been made
2 before so that everything the customer needs to know is
3 in the one sheet and tells them exactly what to do and
4 why to do it.
5 CHAIRPERSON NIPPER: Thank you.
6 MS. KRUGER: I agree. No further comments.
7 CHAIRPERSON NIPPER: Thank you. Dr. Lewis?
8 DR. LEWIS: Likewise. I agree, with no
9 further comment.
10 CHAIRPERSON NIPPER: Dr. Janosky?
11 DR. JANOSKY: I am in agreement.
12 DR. DOUMAS: Same.
13 CHAIRPERSON NIPPER: Thank you, Dr. Doumas.
14 DR. MANNO: Same.
15 CHAIRPERSON NIPPER: Thank you.
16 DR. EVERETT: I agree.
17 CHAIRPERSON NIPPER: Thank you, Dr. Everett.
18 DR. REJ: I think as a general rule, all OTC
19 instructions should be stand-alone and not refer to any
20 other document, professional or otherwise. I found the
21 OTC version very, very difficult, with lots of
22 ambiguities. I think that the professional version is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
309
1 woefully inadequate, particularly the measurement and
2 calibration on quality control.
3 CHAIRPERSON NIPPER: Do you think the
4 performance of the device is properly conveyed in the
5 labeling? That's question number nine.
6 DR. REJ: I asked the sponsor earlier what
7 the definition. Dr. Lewis asked too, and I think Dr.
8 Kimberly, about what the results of this device are. I
9 haven't gotten an answer yet.
10 Did you find that?
11 MS. ENRIGHT: I don't have the calculation.
12 I think based on the input of the statistician, the format
13 that we used was exactly what was requested by the FDA.
14 DR. REJ: But you can't tell me what was done?
15 MS. ENRIGHT: The 95 percent number. I don't
16 have the calculation with me. I will defer to the FDA
17 statistician to help us make that calculation.
18 DR. REJ: I am still just curious how you came
19 up with that number for this publication, and you don't
20 know.
21 The answer is no.
22 DR. EVERETT: My answer is no. I mean I tried
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
310
1 to get it from the statistician, but to no avail.
2 CHAIRPERSON NIPPER: Okay. Dr. Manno?
3 DR. MANNO: No.
4 CHAIRPERSON NIPPER: Dr. Doumas?
5 DR. DOUMAS: No.
6 CHAIRPERSON NIPPER: Dr. Janosky?
7 DR. JANOSKY: Absolutely not.
8 CHAIRPERSON NIPPER: Dr. Lewis?
9 DR. LEWIS: No.
10 CHAIRPERSON NIPPER: Ms. Kruger?
11 MS. KRUGER: No.
12 CHAIRPERSON NIPPER: Dr. Floyd?
13 DR. FLOYD: No.
14 CHAIRPERSON NIPPER: And Dr. Rosenbloom's
15 answer is -- I'm sorry. Give me a second to find it.
16 He didn't answer that question, so we'll skip that one.
17 Dr. Clement?
18 DR. CLEMENT: Not with the present labeling,
19 no.
20 CHAIRPERSON NIPPER: Okay.
21 DR. KIMBERLY: No.
22 CHAIRPERSON NIPPER: Dr. Rifai?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
311
1 DR. RIFAI: No.
2 DR. NIPPER: And I don't understand the 95
3 percent of the time statement either. I believe that the
4 manufacturer can struggle a little bit and come up with
5 something that's very innovative and useful here. But
6 I think they have the work to do to get the performance
7 into the NCEP guidelines first.
8 Number 10. Do you benefits of this device
9 outweigh the risks? That's an interesting question.
10 Dr. Gutman, maybe you could help me with this
11 question because there are some of us who would like to
12 see this device perform a lot better. Should we do an
13 if-then statement or should we do an as-is statement, or
14 should we do both?
15 MR. GUTMAN: You can do both. With the
16 question on the table, is as-is. But you are welcome to
17 make recommendations for what we could do to help the
18 sponsor move forward, if the question is --
19 CHAIRPERSON NIPPER: It's more of an
20 open-ended question than it looks. Does anybody want to
21 start off and tackle this, or do I be a villain and pick
22 somebody? Dr. Rej?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
312
1 DR. REJ: I believe Mr. Connolly closed his
2 presentation with saying that he couldn't imagine any
3 device where the benefit didn't outweigh the risks. I
4 think the risks of an erroneous result are self-evident.
5 Any test worth doing is worth doing right. The NCEP have
6 made some guidelines. From the current data that we have
7 seen, I think that the risks outweigh the benefits.
8 DR. EVERETT: I tend to agree. I mean if this
9 instrument worked, I would have a different opinion. But
10 trying to decide if it really works, and then if it's risky
11 to use in its current status, I would say it's very risky.
12 So again, the risks outweigh the benefit in its current
13 position.
14 CHAIRPERSON NIPPER: Dr. Manno?
15 DR. MANNO: I concur.
16 CHAIRPERSON NIPPER: Dr. Doumas?
17 DR. DOUMAS: Yes, as is. If it's improved
18 in accuracy and procedure. Actually if the reliability
19 of the paper is improved, and is directed to the right
20 people for use, then it will be a different thing. At
21 this point, I see very little benefit and a lot more risk.
22 CHAIRPERSON NIPPER: Dr. Janosky?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
313
1 DR. JANOSKY: As presented to us today, I see
2 more risk than benefit.
3 CHAIRPERSON NIPPER: Dr. Lewis?
4 DR. LEWIS: I agree with Dr. Doumas'
5 statement, and would make the same statement myself had
6 he not.
7 CHAIRPERSON NIPPER: Ms. Kruger?
8 MS. KRUGER: I agree.
9 CHAIRPERSON NIPPER: Dr. Floyd?
10 DR. FLOYD: I agree. It's more risky than
11 beneficial as it stands today.
12 CHAIRPERSON NIPPER: Dr. Rosenbloom said
13 given the need for triglyceride to be interpreted in the
14 context of a lipid profile and the possibility of
15 inappropriate decision making with self testing that is
16 not medically supervised, potential risks far outweigh
17 the benefits.
18 Dr. Clement?
19 DR. CLEMENT: I agree with the current
20 accuracy and precision. It is risky. I think with
21 tightening up the accuracy and precision, it could be quite
22 useful. We know that Type 2 diabetes is as much a problem
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
314
1 of lipid disorders as it is glucose disorders, and could
2 be very, very useful.
3 We would all have a learning curve with this,
4 but I see a future for it if the accuracy and precision
5 can improve.
6 CHAIRPERSON NIPPER: Dr. Kimberly?
7 DR. KIMBERLY: Not in its present form.
8 However, I believe that if the accuracy could be tied closer
9 to the accepted accuracy base, and if the precision was
10 improved, it would be a better device.
11 CHAIRPERSON NIPPER: Thank you.
12 Dr. Rifai?
13 DR. RIFAI: I agree with what Dr. Clement
14 said.
15 CHAIRPERSON NIPPER: Thank you. I think the
16 onus is on the manufacturer, that if he claims benefits,
17 the manufacturer needs to show that in a well-designed
18 clinical study with the present device. That would
19 demonstrate that the benefits far outweigh the risks.
20 We do not have a clinical study. Therefore, I think if
21 you want to say it's equivalent to the predicate device,
22 and therefore it can be used along side any of the other
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
315
1 appropriate lab tests, then you don't have to show that
2 the benefits outweigh the risks. It's the same as.
3 I think we have a conundrum that we haven't
4 done either thing appropriately. We haven't shown it is
5 equivalent to the predicate device in precision and
6 accuracy, and we haven't done the clinical study to show
7 that as-is, you can make hay with this machine. So I have
8 a lot of reservations as well about this particular
9 situation.
10 Okay. We have reached the end of the
11 questions. Dr. Kroll was not able to be here today. Do
12 I need to read his statement into the record? Okay. I
13 would ask the panel to bear with me. It's only a page.
14 "Dear Ms. Calvin: Here are my responses to
15 Polymer Technology Systems Incorporated's
16 over-the-counter triglyceride system. In reviewing the
17 documentation of their device, I find the sponsor has
18 failed to adequately characterize its performance.
19 Further, the device fails to meet NHLBI lipid
20 standardization program criteria, which are for the most
21 part, are easy to meet. This requirement not only is
22 reasonable, but also necessary. An error of 30 milligrams
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
316
1 per deciliter of triglycerides translates into an error
2 of six milligrams per deciliter of LDL. That's a critical
3 value of 100 milligrams per deciliter for LDL. That error
4 of six milligrams per deciliter represents six percent,
5 which is fairly large."
6 "Adding the other allowable biases in, gives
7 us three percent from cholesterol and five percent from
8 HDL for a total of 14 percent bias. A diabetic patient
9 with an LDL cholesterol of 115 milligrams per deciliter
10 (a treatable category), could easily be read as 99
11 milligrams per DL, which is the goal for treatment."
12 "The imprecision is too high for their device.
13 There is no control near 200 milligrams per DL, the typical
14 value for triglycerides. They did not evaluate turbidity,
15 nor glucose as sources of interference. For the
16 over-the-counter claim, they did not have people with
17 limited vision, coordination, nor older diabetics use the
18 device. Varying humidity and storage conditions of the
19 reagent strips was not evaluated. Their linearity claims
20 are unsubstantiated because the imprecision is too high
21 in the test group to detect a non-linearity."
22 "They claimed their method agreed with the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
317
1 CRMLN lab at Pacific Biometrics, but I evaluated the
2 predicted values they would obtain in the following
3 table."
4 I don't have a way to project the table. But
5 at 100 from Pacific Biometrics, the device reads 80. The
6 bias is 20. Pacific Biometrics, 200, device 176. Bias
7 12. At 300, Pacific Biometrics, device 272. It sounds
8 like he is plugging numbers into the regression equation
9 and looking at bias from the device. Four hundred, 368,
10 eight percent bias. Five hundred, 464, seven percent
11 bias.
12 "All these biases are well over the allowable
13 five percent. The package insert fails to state the
14 intended medical use of this device. It understates the
15 necessity for using a fasting sample. Martin H. Kroll,
16 M.D."
17 Okay. The hour is getting late, but we still
18 have time for some questions and answers and final
19 recommendations. Are there any questions for either the
20 FDA or the sponsor from any of the panel?
21 Seeing none, does the sponsor have anything
22 that you would like to present at the end of the panel
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
318
1 now?
2 MR. CONNOLLY: Yes.
3 CHAIRPERSON NIPPER: We would appreciate
4 brevity. Thank you.
5 MR. CONNOLLY: I think I agree. I definitely
6 agree that the product should be near the guidelines.
7 I didn't realize how far away from the guidelines it was.
8 We have talked about this at lunch at great length, or
9 as much time as we had. I think this product can and will
10 meet or come very close to those guidelines.
11 Whole blood is a tough issue. I think
12 calibrating against the CRMLN method, for some reason,
13 there is some bias here. I'm not sure what it is, as to
14 why we got it, but it's definitely there. We have seen
15 that. That needs to go away. If that bias goes away,
16 that takes away a good part of that failure to meet that
17 15 percent number.
18 The precision issue, I'm not too sure about.
19 I can't really address that right now. It definitely
20 needs more high end samples. I don't think there were
21 more than three or four in this area that we are looking
22 at, not necessarily as a cutoff, but there just needs to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
319
1 be more data in that end.
2 So I think we can nearly meet them. The guy
3 that came up with 17.5, loved him. The guy with the 20
4 percent, I thought I liked even better. But I think we
5 can get very close to those numbers, in that the 400
6 milligram, I think it's probably incorrect to say that
7 that was close. There are not enough numbers there to
8 accurately justify that 15 percent or 14.9. But I am sure
9 the product can meet them.
10 It is embarrassing to stand up here and say
11 you can meet them, and then obviously you didn't, but that's
12 why I get those big dollars.
13 The labeling issues were extensive and that's
14 something that needs a lot of work, but that's something
15 that is fixable, I think, in most of those things. The
16 drug interferences are easy enough to test. This is no
17 more than a day or two's work to find out what drugs someone
18 has in mind, get a list together, or go back through some
19 other 510Ks for triglycerides used in other parts of the
20 market, see what drugs they did, and just do the same thing,
21 assuming the question would apply to them.
22 The precision data, it is impossible to do
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
320
1 this day-to-day. So I would recommend a better way of
2 doing this. That is to not only take more patients,
3 definitely more patients, but you just can't get somebody
4 to cough up a 200 milligram triglyceride daily. So I think
5 one patient running -- and it's tough on your fingers,
6 because we do this for calibration. One patient running
7 maybe 10, maybe 15 tests in a 15-minute time period, and
8 let that patient do that a number of days in a row, maybe
9 10 days in a row, then you get some idea of what the
10 precision is on a particular patient. There is just no
11 other way to do that for any test. They are just not
12 static.
13 So I suggest we replace that precision data
14 with whatever the number is, 10 patients, 20 patients,
15 and let them do as many finger pricks as their sore little
16 fingers tolerate.
17 The minimum sample size, we couldn't address
18 that because we don't have the right person here to address
19 that. Then I realized the minimum sample size you were
20 talking about were statistical issues instead of sample
21 volume. So disregard that.
22 Again, the frequency of testing and where some
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
321
1 of these numbers come from, I think we're in uncharted
2 waters where over-the-counter consumer tests are. You
3 take things from other package inserts and use them. Some
4 of them are not applicable. That is what we did. Cut
5 and paste is neat. It still needs a lot of work, but I
6 think it was -- we didn't know what to put there, so we
7 used something from something else. Now we know how you
8 feel about it.
9 I understand the issues with the two inserts.
10 I am not sure that I understand how to resolve this.
11 I definitely think they should be better, as you pointed
12 out in several areas. But as to having one for everybody,
13 I think is kind of -- I don't think that is going to work
14 either. Some of the professionals are going to be --
15 CHAIRPERSON NIPPER: Let me interrupt. I
16 think I can give you the sense of the panel, that two may
17 be okay, but one shouldn't refer to the other. In other
18 words, they should be stand-alone, and the
19 over-the-counter insert should be the same, should be
20 stand-alone, and not refer over to the professional. That
21 is what I heard anyway.
22 DR. REJ: I think there's a model in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
322
1 over-the-counter pregnancy testing, where the technology
2 and the tests are very similar for professional use, but
3 there is no reference to the professional use product in
4 the over-the-counter.
5 MR. CONNOLLY: I'm not sure what happened, but
6 that's easy enough curable. We can fix that part of it.
7 I had to listen to someone say I learned a
8 lot not too long ago. I guess I learned a lot today.
9 We are looking forward to bringing this back here with
10 the -- I don't like saying things that aren't correct,
11 especially when I know they can be true. This doesn't
12 reflect any of our other products, as the reviewers know.
13 They have not seen this kind of data on our other products.
14 So this one will have some more work done on it, and it
15 will be back. Thank you.
16 CHAIRPERSON NIPPER: We appreciate your
17 constructive response.
18 Is there other committee discussion?
19 Final recommendations? By going through the
20 questions, I believe we have made lots of recommendations.
21 We have heard the sponsor has listened to those. Are
22 there other recommendations that those members of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
323
1 panel who are still here have to make to the sponsor or
2 to the FDA?
3 Dr. Gutman, do you have any remarks?
4 MR. GUTMAN: No. We are just grateful for
5 all of your input. We'll look forward to working with
6 the sponsor to try and see if we can come up with a better
7 product.
8 CHAIRPERSON NIPPER: Do you have any remarks,
9 Ms. Calvin?
10 MS. CALVIN: No. I just wanted to thank the
11 sponsor, all of the FDA staff, and the panel of course,
12 and any public attendants that are here.
13 Also, I just wanted to make a note that the
14 next Clinical Chemistry and Clinical Toxicology Devices
15 Panel meeting will be held on December 6 and 7 of this
16 year.
17 CHAIRPERSON NIPPER: I would like to
18 reiterate Ms. Calvin's thanks to all the panel members,
19 the FDA staff, the sponsors for their hard work. These
20 days are expensive to do, both for the Government as well
21 as to the sponsor. I think I can speak for all the members
22 of the panel when we say that we hope that there has been
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
324
1 benefit and payback for the work required to get this
2 meeting together.
3 I would like to thank you for your hard work
4 and your attention. I am glad we're as close to the time
5 as we are. The Chair will entertain a motion to adjourn
6 if there is no further business.
7 DR. REJ: So moved.
8 CHAIRPERSON NIPPER: Second?
9 DR. EVERETT: Second.
10 CHAIRPERSON NIPPER: All in favor? We are
11 adjourned. Thank you.
12 (Whereupon, at 4:35 p.m., the proceedings
13 were adjourned.)
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com