DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
This transcript has not been edited or corrected, but appears as received from the commercial transcribing service. Accordingly the Food and Drug Administration makes no representation as to its accuracy.
BLOOD PRODUCTS ADVISORY COMMITTEE MEETING
Thursday, March 13, 2003
620 Perry Parkway
Kenrad E. Nelson, M.D., Chairman
Linda A. Smallwood, Ph.D., Executive Secretary
James R. Allen, M.D.
Kenneth Davis, Jr., M.D.
Michael G. Fitzpatrick, Ph.D.
Jonathan C. Goldsmith, M.D.
Harvey G. Klein, M.D.
Suman Laal, Ph.D.
Judy F. Lew, M.D.
Paul H. Schmidt, M.D.
Robert J. Fallat, M.D.
Temporary Voting Member
Liana Harvath, Ph.D.
Non-voting Industry Representative
D. Michael Strong, Ph.D.
C O N T E N T S
Welcome, Statement of Conflict of
Interest, Announcements 5
CBER Update: Mark A. Elengold 11
Medical Device User Fee and Modernization
Act of 2002:
Mary Elizabeth Jacobs, Ph.D. 19
CLIA Waiver for the OraQuick Rapid HIV-1
Antibody Test: Elliot Cowan, Ph.D. 25
TRANS-Net Pilot Program:
Alan Williams, Ph.D. 31
Open Public Hearing 67
I. West Nile Virus (WNV) Donor Testing
Introduction and Update of Previous
Activities: Hira Nakhasi, Ph.D. 67
Update on NAT Testing:
James L. Gallarda, Ph.D., Roche 77
Christina Giachetti, Ph.D., Gen-Probe 94
Bruce Phelps, Ph.D., Chiron 104
Andrew Conrad, Ph.D., NGI 113
John Callaghan, MS, Tetracore 129
George Dawson, Ph.D., Abbott Labs 141
Charles Tackney, Ph.D., Ortho
Wallace Narajowski, Focus Technologies 167
CDC Update on Investigations of West Nile
Virus Transfusion-Transmitted Cases
Lyle Petersen, M.D., CDC 173
Donor Serologic Studies of WNV 2001 Outbreak
Susan Stramer, Ph.D., ARC 200
Regulatory Pathway for WNV Testing
Guidance for Industry:
Martin Ruta, J.D., Ph.D. 229
Robin Biswas, M.D. 238
Clinical Study Design, Unit and Donor
Indira Hewlett, Ph.D. 244
C O N T E N T S (Continued)
Panel Development and In-House Testing:
Maria Rios, Ph.D. 263
Blood Supply Management and Triggers for
West Nile Virus Testing
Alan Williams, Ph.D. 280
Testing Source Plasma Donations and Clearance
of West Nile Virus in Plasma-Derived Products
Mahmood Farshid, Ph.D.
PPTA Presentation: 304
Dominique Pifat, Ph.D. 311
Open Public Hearing
Dr. Roger Y. Dodd 326
Dr. Michael Busch 336
Questions for the Committee, Discussion
and Recommendations 339
P R O C E E D I N G S
Welcome, Statement of Conflict of Interest,
DR. SMALLWOOD: Good morning. Welcome to the 76th meeting of the Blood Products Advisory Committee.
I am Linda Smallwood, the Executive Secretary of this committee. At this time, I will read the statement of conflict of interest that applies to the proceedings of this meeting.
This announcement is part of the public record for the Blood Products Advisory Committee meeting on March 13th and 14th, 2003.
Pursuant to the authority granted under the Committee Charter, the Director of FDA's Center for Biologics Evaluation and Research has appointed Dr. Liana Harvath as a temporary voting member.
Based on the agenda, it has been determined that there are no products being approved at this meeting. The committee participants have been screened for their financial interests. To determine if any conflicts of interest existed, the agency reviewed the agenda and all relevant financial interests reported by the meeting participants.
The Food and Drug Administration has prepared general matter waivers for the special government employees participating in this meeting who required a waiver under 18 United States Code, Section 208.
Because general topics impact on so many entities, it is not prudent to recite all potential conflicts of interest as they apply to each member. FDA acknowledges that there may be potential conflicts of interest, but because of the general nature of the discussions before the committee, those potential conflicts are mitigated.
We would like to note for the record that Michael Strong is participating in this meeting as the Non-Voting Industry Representative acting on behalf of regulated industry. Dr. Strong's appointment is not subject to Title 18, United States Code, Section 208.
He is employed by the Puget Sound Blood Center and Program and thus has a financial interest in his employer. He also is a researcher for Roche Molecular Diagnostics. In addition, in the interest of fairness, FDA is disclosing that his employer Puget Sound Blood Center has associations with regional hospitals and medical centers that store platelets and could be affected by today's committee discussions.
With regard to FDA's invited guests, the agency has determined that the services of these guests are essential. There are interests that are being made public to allow meeting participants to objectively evaluate any presentation and/or comments made by the guests.
For the discussions on West Nile virus, Dr. Michael Busch is the Scientific Director, Blood Centers of the Pacific. He receives consultant and speaker's fees from firms that could be affected by the discussions. He also has a financial relationship and/or a professional relationship with several other firms that could be affected by the discussions.
Dr. Lyle Petersen is the Deputy Director for Science, Division of Vector Borne Infectious Diseases, CDC, in Fort Collins, Colorado.
For the discussions on the storage period for pooled platelets, Dr. Mark Brecher is employed as the Director, Transfusion Services, University of North Carolina. He has various interests in firms that could be affected by the committee discussions.
Dr. Ruby Pietersz is employed as the Manager of the Department of Research and Education, Sanquin Blood Bank, Northwest Region, Amsterdam, The Netherlands.
Dr. Edward Snyder is employed as the Director, Pheresis Transfusion Service, Yale New Haven Hospital. He has various interests in firms that could be affected by the committee discussions.
In addition, there are speakers making industry presentations and speakers giving committee updates from regulated industry and other outside organizations. These speakers have financial interests associated with their employer and with other regulated firms. They were not screened for these conflicts of interest.
If there are any other declarations to be made regarding proceedings concerning this meeting, please do so at this time.
Hearing none, then, we will proceed and this conflict of interest statement is available for review.
At this time, I would like to introduce to you the members of the Blood Products Advisory Committee. As I call your name, would you please raise your hand.
The Committee Chairperson Dr. Kenrad Nelson. Dr. James Allen. Dr. Kenneth Davis. Dr. Michael Fitzpatrick. Dr. Harvey Klein. Dr. Suman Laal. Dr. Judy Lew. Dr. Paul Schmidt. Dr. Robert Fallat. Dr. Liana Harvath.
I would also like to introduce to you two new members of the Blood Products Advisory Committee, Dr. Jonathan Goldsmith and Dr. Michael Strong.
Members absent for this meeting are: Dr. Cunningham-Rundles, Dr. Donna DiMichele, Dr. Samuel Doppelt, Dr. Daniel McGee, Dr. Sherri Stuver, Dr. Lori Styles, and Dr. Mary Chamberland.
Before we move into our proceedings, I would like to ask Dr. Katherine McComas from the University of Maryland to come forward. Dr. McComas is involved with a project with the FDA regarding a survey, and she will explain that to you at this time.
DR. McCOMAS: Thank you and good morning. My name is Katherine McComas and I am an Assistant Professor at the University of Maryland. I am here today conducting research on Public Perceptions of the Conflict of Interest Procedures that the FDA uses to monitor real or potential conflicts of interest of its Advisory Committee members.
So, I am responsible for all those surveys that are sitting in your chairs of people in the audience, and I also have a separate questionnaire which I will be asking Advisory Committee members to fill out if they have a chance.
What I would like to ask you to do is voluntarily, of course, is to take about 15 minutes sometime today or tomorrow and to complete the questionnaire. There is a box outside of this room where you can drop it in if you have a chance to do it while you are here.
If you don't have a chance to do it while you are here, there is an envelope that is addressed to me, and you can just drop it in the mail also.
I will be here today and tomorrow, so if you have any questions, please feel free to ask me. I would be happy to talk to you more about the survey. Again, I am grateful for this opportunity to hear your opinions. Your opinions are important and they matter to us.
Thank you very much for your time.
DR. SMALLWOOD: Thank you.
I would like to bring to your attention that we have a very full agenda today, that we have designated time allotments for speakers, that we will adhere to those time frames.
At this time, I will turn the proceedings of the meeting over to the Committee Chair, Dr. Kenrad Nelson.
DR. NELSON: Thank you, Dr. Smallwood.
The first items are a series of committee updates. First, is Mark Elengold, who will talk about CBER reorganization or plans or whatever.
Committee Updates - CBER Update
Mark A. Elengold
MR. ELENGOLD: Thank you. My name is Mark Elengold. I am the Deputy Director for Operations for the Center.
The first order of business, I would like to apologize for Dr. Jesse Goodman, who is the relatively new Director of the Center, for not being here to do this update himself, but he had a meeting he had to attend with the Commissioner this morning, so I get the honor of doing it.
I see Dr. Smallwood start the clock on me and rather than anger her, I will talk fast.
It is pleasure to be here. I haven't been at one of these for a few years, and I would like to introduce myself to the new committee members and say hello to some of the old ones. A lot of things are happening at CBER as you may have known, and a full update would probably throw the agenda into disarray, so I will cover several of the high points. I don't see anything, so I will say next slide, please.
The first thing I would like to cover as we have a new Center Director, we also have a new Commissioner. The Commissioner has established five priorities for the agency over the next few years, and we are in the process of developing a strategic framework for initiating them.
Number one is a strong FDA. Number two, risk management. Three, decreasing medical errors and adverse events. Fourth, better informed consumers and, fifth, countering terrorism.
All of these have groups that are working on it. I personally serve on the Strong FDA group and on the Human Capital subgroup of that organization. Over the next few months, you will be seeing more and more.
We are holding, in the next few weeks, a Risk Management Workshop Seminar. For those of you on the committee and in the audience who have an interest in that, I urge you to check on that and attend and participate in it.
What do we see as challenges in CBER over the next year? Well, number one is organizational changes, and we will talk about that a little bit. New performance goals, I will talk about that as well. New technologies, international harmonization, e-business, countering terrorism, something that is taking increasing amounts of our efforts, and a strong regulatory research program, something that has been a keystone of CBER's work since it was founded over 100 years ago.
This is our organization chart. The first thing I will talk about is, number one, the little red box, which you probably can't read, but is in the handouts. That is the Office of Therapeutics Research and Review.
As you probably know, a decision was made last summer to transfer responsibility for therapeutic biologic products from the Center of Biologics to the Center for Drug Evaluation and Research.
These are the key questions everybody immediately asks, what's going? There are monoclonal antibodies, cytokines and growth factors, interferons including recombinant versions that mimic those found in nature, proteins intended for therapeutic use that are extracted from animals or microorganisms except clotting factors, which is of key interest to this group, and other therapeutic immunotherapies.
What's staying and despite some of the things that you may have read in the trade press, the majority of our products are staying as evidenced by the fact there are two slides of what's staying.
Monoclonal antibodies, cytokines, growth factors, other proteins that are used only in the preparation of other biologic products, so products that are not intended for use as therapies, but used to expand cells, expand vectors, create other things, we will regulate as part of the products they ultimately are incorporated in, viral-vectored gene insertions, what the lay press likes to call gene therapy, products composed of human or animal cells or from physical parts of those cells.
Plasma expanders, allergen patch tests; one of my favorite group of products, allergenics; antitoxins, antivenins, and venoms; in vitro diagnostics used in the screening of the blood supply or for the diagnosis of HIV; vaccines, and toxoids and toxins intended for immunization.
Now, those are the categories established in this transfer. It can get a little confusing. For example, one of our most popular products in the lay press, the botulinum toxin, the cosmetic or other treatment Botox is going because it is used as the treatment, not intended for immunization.
Over the next few months, as we work through this process, there will be a date which will occur relatively soon where these products will be transferred for review of responsibility to the Center for Drug Evaluation and Research.
To minimize confusion on the budget side of the agency, for the first few months of that, the people will continue to be employees of our Center, but be detailed to CDER, and at the start of the next fiscal year, which is October 1st, the actual official transfer will take place.
We will be using various means to notify sponsors of a way to tell whether their product is being transferred, and there should be an announcement on the details of that in the next few weeks.
Another big organizational change that is of interest to the committee members and to those following the committee is the creation of the Office of Cellular Tissue and Gene Therapies. That is the bottom red box. Phil Noguchi is the Acting Director of that.
Why did we do this? Well, an increase in regulatory activities in these products and the consolidation of these products which were spread among our other product offices into one place to leverage the expertise and provide consistent regulatory authority.
The mission of it is to review things and regulate things, such as tissues, cell and tissue products, gene therapies, xenotransplantation, unique assisted reproductive therapies, and combination products including those that have living cells and tissues.
Of course, the ultimate goal is to assure the safety, identity, purity, and potency of novel products.
This is the organizational structure. As you can see, everyone is Acting for the time being. We will be in the process of recruiting for these positions.
For those of you who aren't familiar, I will go over these very quickly. These are our proposed approach to the regulation of tissues. The bottom line is the most important. It's a tiered, risk-based approach.
I did not see the final agenda until last night, and I found out that I am being followed by Mary Beth, who is going to be talking about this, so I will talk even faster than I normally do.
These are the key provisions of MDUFMA. I will let you look at those in the handouts or read them, I will let Mary Beth expand on it.
These are the bottom-line numbers for it. What is interesting is in the recently enacted final budget for the fiscal year we are now virtually halfway through, the appropriate amounts were not provided, as well as in the President's budget for the next fiscal year, not all the funds that were planned have been requested.
The first year fees range from $154,000 for a premarket application, to a little under 2,200 for a 510(k). I will just highlight again the bottom line, which is that without reauthorization, the allowable sunset in October of 2007.
Finally, when I started with FDA some 32 years ago as an investigator, I was told never tell people you are from the FDA and you are here to help them.
I like to say this because over the past years, that has changed. We do believe we are here to help you and anytime we can, members of the committee, members of the public, these are the quickest way to contact us.
Thank you. Should I ask if there are any questions, Linda? Okay. I always do what I am told to Dr. Smallwood. She has me well trained.
Are there any questions I can answer real fast without going off schedule? Committee members?
Thank you very much. It's a pleasure.
DR. NELSON: Thank you.
Next on the agenda is Dr. Mary Elizabeth Jacobs to talk about MDUFMA.
Medical Device User Fee and Modernization Act of 2002 - Mary Elizabeth Jacobs, Ph.D.
DR. JACOBS: Good morning. Last December, we had a rather thorough discussion of the law including user fees, performance goals, and all the key provisions, so today I am giving you a short update.
I would like to cover the guidances and the FR Notices which have come out, what the current performance is, implementation, and planned reports.
On February 25th, we posted the guidance everyone was waiting for, which differentiates among all the different types of supplements which have different fees, so that guidance covers the questions. It is interim guidance waiting for feedback, but it is pretty detailed and so far we are not getting too many questions on it.
Secondly, it discusses small business qualifications.
We also have FR Notices. These are on our web site, FDA.gov. Go to Devices, then, to the User Fee page, it is quite easy. These discuss all the user fee payment procedures, and of interest to a lot of people, it announces the establishment of the public docket, which has been since the December BPAC.
You go to FDA.gov. Go to Dockets, to the list of them. You will see it quite easily. You can not only send in your own comments electronically, but you can also read all the comments that other people have sent in.
Briefly, I just want to remind you that for all three types of our submissions, PMAs, 510(k)s, and BLAs, we have cycle goals for all three of them, and for two of them, we also have what are called "decision goals." These change from 05 through 07, but this illustrates for you, for PMA goals, that we have cycle goals and then we have total decisions goals.
This shows the same for 510(k)s. You will see we have what people think of as quite a challenging goal for the total decisions.
BLAs, just as a reminder, under MDUFMA, we now have for resubmissions a new kind of resubmission, which is things that have shorter updates, and we have only two months to review that.
This shows our receipts as of March the 3rd. Our 510(k)s are running a bit higher than they ordinarily do. We are quite pleased because so far we have met all of the FY 05 MDUFMA goals.
Process changes that have been involved in this include a new document courier service among all of our different locations. We have more CDRH scientists serving on our teams, which has helped to speed up our alliance with them.
We have had additional Least Burdensome training, and we are going toward a practice of finishing the reviews by mid-cycle and then working very actively with the companies to resolve the problems before the first letter goes out.
Implementation. We now have more reference materials on the web site. I encourage you to look at those. The Office of Combination Products has been established. Guidance is continuing under rapid development.
We have had a lot of outreach. This is only one example, but all of our roundtables and liaison meetings have had updates, and I want to particularly draw your attention to something that has just been announced.
On April 9th, there is going to be a live satellite broadcast and webcast with staff from both CBER and CDRH answering your questions and explaining the law, so if you are interested, you will see the instructions for this on your web site.
What are the reports that are coming up that are related to CBER? I am not dealing with the ones that are GAO reports on what the appropriations are, et cetera, or the reprocessed devices. We don't review any reprocessed devices at CBER, so there are a number of reports on that, that are just pertinent to CDRH.
But at the end of April, the criteria for third-party inspections will be published. In August, there will be information on devices used just for pediatric patients. In October, we will have accredited third parties using the criteria which are published in April.
October 26th is the report which goes on review by Centers other than CDRH, and we have gotten questions on this. Yes, we are planning to report using the MDUFMA goals on that even though only one goal kicks in this year officially.
Also, there is a report on the Office of Combination Products. Finally, November 30th, there is a report on FDA as a whole and what our performance has been in the first year.
Again, I am referring you to the web site, very easy to find under FDA.gov, either directly from that or going through Devices.
Finally, I would be happy to talk with any of you about any of your concerns or any problems that we need to address.
DR. NELSON: Thank you.
Any comments or questions? Yes.
DR. STRONG: There was a rumor around the hallways that one of the changes might include a change in submissions for instrumentation relative to blood screening assays, that you were going to require BLAs rather than 510(k)s. Could you clarify that?
DR. JACOBS: We are using the same instrumentation rules and software rules that CDRH is using, and they are a little bit complicated because it depends on what your first submission is, so that if you have a platform which is already being used, and now you want to put a new test on it, you then need to supply when you submit for that test, the instrumentation and software that is specific to that test.
But let's say you want to have your very first use of a platform with a test which is a BLA test. Then, you need to supply all of the instrumentation and the software when you supply that test.
We would need to discuss with you what your plans are. It may be that the company would like to have a separate 510(k) for that platform because they plan to have other tests on it. They may prefer to send it in with the BLA.
So, the basic rule is if you are already using the platform, you only send in as part of your test what is new and specific to that test, and we incorporate by reference all the prior ones.
Now, I think what you are zeroing in is the second case that I talked about, is if you want to have a test for the very first time, which is a BLA test on a platform, then, we would need to talk to you about whether you would want to send it in with that or whether you would want to send it as part of a 510(k), which has shorter time frames.
Anything else? Thanks.
DR. NELSON: Next, is Dr. Elliot Cowan, who is going to talk about the CLIA Waiver for the OraQuick HIV Rapid Test.
CLIA Waiver for the OraQuick Rapid
HIV-1 Antibody Test - Elliot P. Cowan, Ph.D.
DR. COWAN: Thank you.
At the last BPAC meeting, I had informed you about the approval of the OraQuick Rapid HIV-1 Antibody Test. Since that time, there has been a significant development that I just wanted to formally inform you about.
The OraQuick PMA was approved on November 7, 2002 as a moderate complexity device under CLIA.
With an intended use to detect antibodies to HIV-1 in fingerstick whole blood specimens, as a point-of-care test to aid in the diagnosis of infection with HIV-1, and a test that would be suitable for use in multi-test algorithms designed for statistical validation of rapid HIV test results when other rapid HIV tests become available and data is supplied to us to demonstrate that use.
OraQuick was approved as a restricted device with the following restrictions: Sale is restricted to clinical laboratories that have an adequate quality assurance program including planned systematic activities to provide adequate confidence that requirements for quality will be met, and where there is assurance that operators will receive and use the instructional materials.
OraQuick is approved for use only by an agent of a clinical laboratory.
Test subjects must receive the "Subject information" pamphlet prior to specimen collection and appropriate information formally known as counseling when test results are provided.
The test is not approved for use to screen blood or tissue donors. Everyone who purchases the OraQuick test will receive a letter which states the following: "By purchasing this device, you are doing so as an agent of a clinical laboratory and agree that you or any of your consignees will abide by the restrictions on the sale, distribution, and use of the device."
To remind you a bit about the performance of this test, the clinical trials demonstrated a sensitivity of 99.6 percent with a 95 percent confidence interval of 98.5 percent to 99.9 percent. Specificity in the clinical trial was demonstrated to be 100 percent with a 95 percent confidence interval of 99.7 to 100 percent.
OraQuick was granted a CLIA waiver on January 31, 2003, as announced by President Bush during a press conference which followed the announcement of his HIV initiative, obviously, a very high profile event.
The waiver was granted on the basis of data that was submitted to FDA, and that consisted of a study that was done at four sites, consisting of 100 lay users who had no prior laboratory experience, who tested a panel of 6 masked randomized specimens, 2 of which were negative, 2 of which were low positive, and 2 of which were high positive.
The bottom line from these studies was that there was no statistically significant difference between the lay user results and the correct results.
Where we are now is that the package insert for the waived device is in preparation, and that package insert will contain details on the waiver studies.
A critical piece of this is that the restrictions that were developed for the device at the time of approval will remain in place for the waived device.
A centerpiece to this, of course, is the quality assurance restriction.
There have been obviously many questions about what constitutes an adequate quality assurance program. To address this, the Centers for Disease Control and Prevention held a working group at the end of January--actually, it turns out it was just days prior to the waiver approval--to bring together national and international experts to discuss what types of quality assurance issues should be addressed for the OraQuick Rapid HIV-1 Antibody Test and rapid testing in general.
The recommendations from that working group are now being collated and a document is being prepared which will be posted on the CDC web site when appropriate for use by their funded programs and also will be available for the public, so that they will be aware of the types of quality assurance issues that should be addressed when running a test like this.
I would also like to let you know that for those who are interested, on Thursday, April 24th, from 1:00 to 3 o'clock, there will be a satellite broadcast by CDC, which is called an Update on Rapid Testing for HIV. It will be live along with some taped interviews to bring together folks who are interested in this and for CDC to discuss issues related to rapid HIV testing and especially related to quality assurance.
Thank you very much.
DR. NELSON: Thank you.
DR. SCHMIDT: You described the customer letter. Is the customer the person who wants to test himself, is that what you mean?
DR. COWAN: No. This is not a self-test. This is for purchase by laboratories, and one of the key restrictions is the testing has to be performed by an agent of a clinical laboratory.
A customer is defined as the person who is purchasing the test. The way that the system is set up, the only people who can purchase the test are those who are registered with CLIA either as a moderate complexity lab or, now, those who would possess a certificate of waiver.
DR. SCHMIDT: I understand. I was worried about the words "agent of a clinical laboratory" and "consignee" as words that would go over the average user. Thank you.
DR. COWAN: Yes, thanks.
DR. FALLAT: What is the purpose of the testing of the lay user if it's not going to be used by the patient? What was the meaning of that waiver?
DR. COWAN: A lay user could be someone who is not formally trained in laboratory science, for example, someone who has a Certificate of Waiver would purchase this test and give it to his or her staff to run. The staff may have a high school education, for example, or may not even be a high school graduate.
One of the key provisions of moderate complexity testing is that there are educational requirements and training requirements. Those are not imposed with CLIA waiver. I need to emphasize it is not designed as a self-test and that is covered by the agent of the clinical laboratory.
DR. NELSON: Thank you.
The final committee update is the TRANS-Net Pilot Program. Alan Williams.
TRANS-Net Pilot Program
Alan E. Williams, Ph.D.
DR. WILLIAMS: Thanks, Ken.
The TRANS-Net program was first presented in concept to this committee I believe in the fall of last year, and we have recently completed two months of pilot work with the program involving 5 transfusion services and 4 blood centers, so it seemed appropriate to give the results of this pilot and some of the progress that has been made to date and looking forward into the future, what some of the plans are with respect to this program and potentially integrating it with the HHS Blood Supply Program and other studies that are out there in the field.
Just very briefly, I think it is quite clear to everyone that the blood supply and ways to monitor the blood supply have been the focus of intense interest particularly over the last year or two in part stimulated by the CJD deferral.
This has reached the level of congressional subcommittees and the New York States legislature and, as described here, the HHS Advisory Committee for Blood Safety and Availability, which at its September meeting last year made some specific recommendations that HHS should actively increase public awareness of the ongoing need for routine blood donations by healthy individuals, the Department should maintain and/or increase funded support for blood supply monitoring, and should also support initiatives to improve management of blood inventories.
I didn't want to go through all the details associated with these recommendations, but, in fact, some were very specific about actions that should be taken.
At this meeting and in several presentations subsequent, FDA has presented a concept of three goals related to blood supply monitoring and areas that should be covered, the first being to monitor the breadth, the impact, and the duration of blood and reagent shortages.
There are programs in place which do this either for various elements of the blood supply system or collect the data, but it is not available in real time, or collect it for blood centers and not transfusion services.
As I will mention in a couple of slides, there is some room for improvement, and this gave rise to the concept of developing TRANS-Net.
The second goal, I think a very important one, is to monitor the long-term blood collection and utilization trends. This was conducted as a biannual survey funded for a number of years by the National Heart, Lung, and Blood Institute. Over the past four or five years, this has been a biannual survey run by the National Blood Data Resource Center, an affiliate of the American Association of Blood Banks.
I think the data from those studies have really been very unique and very important to understanding trends in the blood collection and utilization.
The third goal is to specifically monitor blood center inventory either by capturing inventories aggregated by the major blood collectors or from individual blood centers, and should reserve depots be developed around the country, obviously, one would want real-time inventory data for those depots.
So, those are three major goals to establish a conceptual design for a monitoring system.
I mentioned that there are some opportunities to improve monitoring specifically monitoring related to shortage. I think one area that particularly needs to be addressed is sensitivity to local shortage.
While there are data available, it appears that a program that could capture a very broad-based data input from both transfusion services and blood centers would potentially allow us to focus in at the regional level when shortage occurs as opposed to reading it or hearing about it in the media as typically occurs today.
A system obviously should have national representation. Data ideally should be available in real time because if one needs to make appeals or otherwise conduct interventions, obviously, real-time data is more relevant for doing that.
We feel that blood center information and transfusion information gives slightly different perspectives on supply availability, and we feel that a sophisticated system should be able to integrate those two sets of data and build a model for current supply.
Ideally, this model could then be used to predict situations in the future, such as the holiday situation this past year. There were two mid-week holidays. It shouldn't have been a surprise.
Potentially, had the situation occurred in the past, we would have been able to predict a particularly difficult time in conducting blood drives, and not had the sustained blood shortage that was seen particularly during the January and early February period.
Any system clearly needs fairly broad data access, ideally, even available to the public, so that it can be used to help stimulate recruitment of donors, but clearly the data needs to be available and can be utilized by government and industry to create interventions as necessary to help sustain supply.
Specific to TRANS-Net, I want to acknowledge the team that has been working on the program. The project manager for the program is Captain Manette Niu assisted by Sharyn Orton, Amy El Naggar, Stan Pawlowski, and myself at the FDA Office of Blood.
We have also worked very closely with Voxiva Corporation, which I think has put a lot of development work into the program, developing the platform and making their proprietary software available for telephone data reporting.
I would also like to thank numerous colleagues who have participated in the program both in the early design phase as consultants, and then subsequently in the pilot phase.
Now, what are some of the design characteristics? I should say that this design was put together under active consultation with 12 hospital transfusion services and blood centers largely done in the second half of last year.
One goal is to keep the system very simple, to utilize data routine compiled by the facility, so that there wouldn't have to be specific efforts made to determine the data that are needed.
The reporting would be daily, on a weekday, reporting with a simple shortage/no-shortage option as the first field of reporting.
Denominators for reporting should be available. We would like a daily report from any site participating in the program, so that if a shortage is reported, we know that the same site would be represented if a no-shortage situation existed there.
We did want to recognize some local variations in a shortage definition. I have a slide that goes into this a little bit more, but while we recognize the need for a standardized datapoint, in fact, institutions differ one from another, and we wanted to be able to allow for that to define shortage in a way that it is a shortage to that particular institution.
The primary system is a web-based system, and I will show you some shots of the actual web pages, but information indicates that about 40 percent of transfusion services probably do not have regular web access available. This could change within the next six months, but that is the figure we are working with now.
Because the intent is to make this a population-based system rather than a sample, we wanted to give even the small and more remote transfusion services the opportunity to report, so through Voxiva, there is an option for telephone reporting of data for sites that don't have web access available.
This is a voluntary system, there is no payment specific to the reporting sites. A question frequently arises what is the incentive for sites to report.
Well, as I will show, the data from the entire system will be available through the web in a real-time manner. The data will be available rapidly, essentially as soon as it goes in and goes through server, it will be available on the map.
Secondly, there needs to be meaningful use of the collected data, not only distributed to those who have a vested interest, but made available to inform policy decisions, made available to inform government and industry, to, you know, tailor appeals as needed, and even to shuttle the information in times of emergency to the disaster task force, or FEMA, or other agencies that specifically need to know where there is a blood situation around the country.
We also feel that on the public side, that the data could perform some support functions for recruitment efforts, such as making available to the local media exactly what the blood supply situation is in a given area.
I think it is fairly common knowledge that the public is willing to donate blood, but you need to ask them, and if you make this a daily reminder of what the blood situation is, it could well help recruitment.
As mentioned, the system is designed to be population based. Given the web information transfer, that is easily accomplished. There are 5,000-plus transfusion services in the U.S., so that potentially, all blood centers and transfusion services would have the capability to participate although we would want them to register for participation, so that we could get predictable responses and data submission.
A brief registration form will be requested for participating sites. This will allow data verification, follow-up tracking of the site for consistent participation so as to develop a denominator, and obviously, would give us contact information if anything specific arises that we would like to get more information about.
Data entry is very simple. If a site forgets to report on any given day, there will be an e-mail sent to that site at a time chosen by them, reminding them to submit their report. That e-mail contains a URL, so it is simply a double-click on the link, and if there is no shortage, a single-click to report "no shortage," and they are done. This actually worked very well during the pilot. It seems to be a very simple and trouble-free way of reporting.
Now, I mentioned that we are allowing a little bit of flexibility in shortage definition, so it can be institution-specific, could change by hospital, by blood center, by geographic region. It could change over time, could be change based on differences in the supply network.
Because there have been some questions about this just recently, so I wanted to give a couple of examples. These aren't necessarily inclusive of all possibilities, but, for instance, transfusion services sometimes need to preserve their Rh-negative units and would give Rh-positive units to male patients because medically, although it is not an ideal situation, the likelihood of causing future problems in terms of transfusion is there, but it is largely a controllable situation.
Now, this varies from institution to institution. Some institutions would never do this unless they really did not have the Rh-negative units available. Some institutions do it fairly commonly, so there is a difference in policy with respect to that.
A second difference between institutions, we know even among the pilot sites, that it lists at least one site maintains an emergency reserve. This happens to be a frozen red blood cell reserve.
So, although they may not meet a patient's support criteria, if they have to go into that frozen reserve to get blood to meet that patient's needs, that would certainly constitute a shortage situation for them. That simply wouldn't be an option for a site that didn't have that reserve.
In terms of data management, the plan is that the electronic data will be received and mapped at a central FDA monitoring site. This potentially could also be a contractor, in fact, Voxiva is the server host for the pilot.
Registration of facilities allows verification and follow-up of selected reports.
Data access will be through software that considers the geo-spatial density and frequency of shortage reports, and I will show you what I mean by this language in a moment. Basically, the shortages can be color coded by geographic area on a given map.
Software can also be set up, so that if any shortage criteria meet a certain cutoff, an alarm can be sounded or an ALERT message at any predetermined threshold.
Clearly, we need to get some experience with the study to determine what that threshold would be, but that is a potential for the system.
The data would be available on an interactive public web site using the GIS software and color-coded maps. These maps would have a drill-down capability, so that although the first map would be, for instance, the U.S. map with states, if you click on a certain state, you could potentially go down to the county level or even less.
We originally set up a rule being familiar with OMB requirements that you couldn't drill down to less than 5 participants, and in the pilot, we also put that rule into place. If we made an error, we wanted to make it an error on the conservative side, that we wouldn't display somebody's data that were identifiable and get into trouble for that.
In fact, the pilot recipients' response, in fact, expressed the opinion that they wanted to see their data, and they didn't really care if it was individually identifiable. So, I think there is a message there that down the line perhaps we could set up a consent option there to override that rule.
This is a screen shot for the data entry for TRANS-Net. This comes up a single click on the URL as the registration by user ID and pin number.
This is the initial data screen. You can see on the left side, there is simply a shortage/no-shortage option. If you click "no shortage," you are done, the data is recorded and you exit the system.
The sites from here diverge a little bit for the blood centers and the transfusion services.
if you click on "shortage," you get a drop-down list of characteristics related to that shortage. The list here is for transfusion services, includes criteria probably familiar to most of you, delay of elective surgery, delay of emergency surgery, delay of medically necessary transfusion, et cetera.
There is a similar list with criteria specific to blood centers.
This is an example of the mapping, clearly, a state map with some color coding related to different shortage situations. Now, we could define shortage any way that was appropriate using the data in the system. One way would be to use the shortage designation by the site, but, in fact, if we wanted to tighten up that criteria and produce a map showing where surgeries were being delayed and in a very severe situation, that mapping function could be modified to do that, as well.
This is a drill-down to the Washington area using the same software. These are not real data, these are artificial data.
Now, just to conclude, I wanted to obviously give some information about the pilot study. As all of you know, we experienced a rather prolonged shortage following the holiday period.
This, in part, was likely due to the fact that we two mid-week holidays, and there was a blood shortage recognized fairly broadly by the 3rd of January, which then resulted in appeals from the major blood organizations, as well as Secretary Thompson.
The TRANS-Net program, as well as the HHS program, ABC, Stoplight, et cetera, showed that, in fact, there was a shortage, recovery was somewhat prolonged, and then complicated by the blizzard several weeks ago.
Now, the scheduled start of the TRANS-Net pilot was the 6th of January, and we fully intended to use mock data for the pilot, the reason being, number one, we didn't expect any problems to be going on and we wanted to see some variation in the data, so we would have some things to work with; and, two, with mock data, we felt that we could produce the maps and test the functionality of the report maps, as well.
Well, what happened when we recognized that the problem that existed following the holiday, by the second day of the pilot, we asked the sites to then report real data, and, by the way, could you go back to the beginning of the week and do that. So, the pilot actually now consists of real data.
This is just a graph of shortage reports shown separately for transfusion services and blood centers, just a very rough cut of the data line graph. You can see the number of sites reporting shortages on the y axis versus date on the x axis.
You can see from the 6th of January through about the 20th or so, there were a larger number of sites reporting shortage. This underwent a gradual decline, so that by the 10th, at least the transfusion services had more or less normalized, whereas, the blood centers were still reporting some spot shortages.
Interestingly, this again went up again near the end of February. I don't know a specific explanation for that, but we did start getting increased reports at the end of February.
This is just a first attempt to deal with some of the criteria. What we did was weight some of these criteria, for instance, delay of elective surgery, delay of medically necessary transfusion were weighted more highly that some of the other factors.
You can see a similar trend of the scale based on the weighted criteria, but I think one thing that is evident is that the amplitude of the graph is considerably higher, so it does give us a little more flexibility to assign some importance to shortages that are reported.
Just a few bullet characteristics about the shortage. Of 191 transfusion service shortage days, 35 percent of the time the transfusion services reported that they had a shortage. Just from a very limited number of sites, it appeared that they recovered somewhat sooner than the blood centers, and again, there was a slight rise at the end of February.
Prevalence of shortage reports by blood center days was a little higher, 48 percent during the period. There didn't appear to be a sustained return to a no-shortage status.
So, a hypothesis which I think makes sense, but it remains to be advanced by data, is that blood centers provide a more sensitive and early monitor for shortage whereas, transfusion services probably reflect more acute need related to that shortage.
Finally, just some characteristics of the criteria that were reported. For transfusion services during that pilot period, there were no delays of emergency or elective surgery. There were 11 transfusion service days where strategic reserve was tapped. There were no delays of elective surgery. There was one instance of a delay of medically necessary treatment, and 24 transfusion service days reported support of Rh-negative patients with Rh-positive blood.
Finally, for blood centers, the critical shortage media appeals were reported on 11 blood center days, use of strategic reserves 10 blood center days, and order were cut, 31 of the 176 days.
So, where does the program go from here? This was first phase, 9 sites pre-OMB review. I think the next step is there are plans to have a discussion with the Department on potential integrating TRANS-Net with the HHS blood supply monitoring system obviously to try to prevent dual reporting and optimize the data collection.
Once that plan is developed, we would like to conduct the second pilot with several hundred participants which will allow testing of the telephone system with a little more depth.
Monitoring blood supplies is more at the regional level, and allow functional mapping of the data, and then ultimately, at a time that we can't predict right now, the third phase, we would like to roll it out to the full population of blood centers and transfusion services.
We feel there are some potential expansions to the program. It could be used as an information conduit for things like blood center inventory, platelet supply, adverse event reporting, reagent and supply shortage monitoring, and, for instance, using the e-mail function, a transfusion service that was short on any particular blood product could, in fact, issue a broadcast e-mail message to surrounding participants and actually have those sites call into that service if they had the product available rather than calling around from the site that needs the product.
So, that is the description of the current status of the program, and hopefully, we can give you another update in a few months.
DR. NELSON: Thank you, Dr. Williams.
Are there questions? Jim.
DR. ALLEN: A comment and a question. This is very nice and I am delighted to see you moving forward on this. My concern, as with surveillance for communicable diseases during critical periods of time, i.e., you know, all of this fancy reporting that is postulated during a terrorism event is likely to break down because when people are the busiest, they don't bother to report.
The same concern here, that we have an event where there really is a critical shortage, you send out the reminding e-mails, people just aren't going to have the time to respond, so that this system tends not to gather the data in real time when it is most critically needed, until we can get to the point that we can get some sort of automatic data mining or computer-based reporting from existing data.
Second, going back to an early slide, the HHS Advisory Committee on Blood Safety and Availability, the point one that DHHS should promote increased public awareness of the ongoing need for routine blood donations, if I remember correctly, and perhaps Dr. Harvath can comment on this, NHLBI had such a program 10 or more years ago. I think it was cut because of absence of funds.
I wondered, is HHS broadly, not necessarily just FDA, planning to reinstitute such a program to work with and support blood centers in their call for blood donations?
DR. NELSON: Liana, any comments?
DR. HARVATH: I can't comment on behalf of NHLBI. I thought that the plan was, as in the past, to approach all of DHHS and talk to the major players that are interested rather than focus only on the Heart, Lung, and Blood Institute.
I don't know, maybe you know more than I do about that, Alan.
DR. WILLIAMS: Well, the Heart, Lung, and Blood Institute program goes way back. It was called NEBRP, National Educational Blood Resource Program. It had very limited NGLBI funding. I think it was staff support for the program, but largely, they were attempting to run it on corporate funds.
But specific to the recommendation, I think there has been activity at HHS. I believe that the major tissue recruitment campaign has now, at least in a modest sense, begun to include blood donation in some of its awareness campaign.
I don't know of any specific funds to help blood centers with recruitment campaigns, but I think there also is another initiative to spur on donation by federal employees and specific campaigns related to that, so I think there is movement at the HHS area.
DR. SCHMIDT: Two comments, Alan. First of all, we certainly need such a system and you are well on the way.
The first comment is you haven't discussed a situation that exists in some part of the country that I am sure most of the members of the committee are unaware of, and you are talking about the blood center and the transfusion service, the transfusion service being the traditional hospital blood bank where blood was sent to be under the control of the pathologist who charged so much per cross-match, and it worked very well.
But there are some parts of the country, as I know you are aware, where there has always been a different system, where it has been recognized that the blood center has the medical experts who know how to take care of patient and nowadays it is called transfusion medicine.
It is certainly true of Florida Blood Services and Seattle, and I understand now in Cincinnati, Pittsburgh, where I guess the system we have had for 60 years in Tampa is that we rent space inside the hospital where we place our employees working for the blood center, and the blood is sent there. They don't work for the hospital, but everything is billed through the hospital, therefore, they have control from vein into vein, as it is called.
So, that really affects the situation here, so we have been doing it for 60 years. We see this picking up. Pittsburgh is relatively new in the system, and Cincinnati, as I understand, and it is maybe because the pathologist can no longer get his two bucks per cross-match.
Also, we have electronic cross-matching, as we call it. So, you have a system that differs from the traditional way of doing things, and may be increasing, and I know it's built into your system, I am sure, but I just wanted the people on the committee, who don't have that in the traditional Red Cross system, and most of the non-Red Cross blood centers in the country just don't have that, but it does exist and it may grow.
Do you have any comments on how you would feed that in?
DR. WILLIAMS: I guess the major comment is we did consider that. In fact, we had some of these combined services in the pilot program, as does the HHS program. After consideration, we felt that it would be most relevant to consider those services as reporting as transfusion services, that those would be the most relevant data from those combined activities.
But, yes, there certainly are a number of them out there, they tend to be large, and they are increasing.
DR. SCHMIDT: The other comment, I think I made before, is this really FDA's job, and maybe someplace in HHS, you are certainly doing it very well, but just drawing some parallel, when we talk about quality, we talk about benefits to the patient, but we are not allowed to talk about the cost-benefit ratio. FDA doesn't consider costs, I mean it's just benefits.
Now, we are talking about not quality, but quantity, and I know there must be something in the FDA law that allows you to do it, but it would just seem that you are spending a lot of good time in which people think you should be working on quality, on working on this quantity problem, and I am sure that has been considered and you may not want to comment on that, but to some of us, its seems pretty strange that you are into this business.
DR. WILLIAMS: I will comment briefly and then I guess Jay wishes to, as well.
There was a discussion for some time as to where a monitoring system should live within the HHS. I would say it largely resides at FDA because FDA had the interest to design a system and put it forward as a potential way to conduct blood supply monitoring.
It could live at FDA, it could live with a contractor, it could live elsewhere. I think this system is not necessarily FDA unique, but I would add that FDA in some sense does have a vested interest in the data because so frequently now, policy decisions are a careful balance between blood availability and preserving safety. That was fully manifest to the DCJD decision, but almost every decision made now that emerges is a factor.
DR. SCHMIDT: The comment on that would be I don't think FDA has great experience in what goes on in hospitals. I mean the things that happen there in relation to the blood supply and everything else, I mean every hospital in time of disaster gets all the blood it can and squirrels it away and all hospital technologists are part squirrel, I think.
So, they have different interest in what goes on. They are really not interested in the rest of the community with a national supply, and you kind of have to know these people to work in this system. They are trying to beat your system as well as they can.
DR. NELSON: Jay.
DR. EPSTEIN: I think Alan hit the nail on the head here, which is that there is a close interplay between considerations of safety and supply. This issue came to the fore when we were considering deferrals for VCJD risk that might have a large impact on supply, and we discussed with the Department, David Satcher at the time, Surgeon General and Assistant Secretary for Health, took the position that we needed to closely monitor the supply impact of those policy decisions.
Roughly at the same time, the FDA proposed adding to the Blood Action Plan, which had become a departmental initiative, an item on monitoring and increasing the blood supply, which had a number of dimensions, which included optimizing strategies for donor recruitment and retention, eliminating barriers to safe donation, and also monitoring of blood availability.
We have come to see shortages as safety concerns. We do have enough latitude in our mandate to consider product distributions as a part of what we can examine, but more broadly, we have simply come to see shortage concerns as safety issues.
We do have shortage officers in FDA dealing with a multiplicity of product areas including blood, also reagents that are needed in blood manufacturing sometimes in shortage situations. Do, for these various reasons we get involved.
Now, the question is whose responsibility is it in the end, and I think that we have evolved to a point of view that it is a shared responsibility. We have pieces of it at FDA, piece of it at the Department, pieces of it historically and some currently at NHLBI, and most recently, as Alan suggested, the combined campaign for organ and tissue donation with blood donation has given a piece of the responsibility to HRSA.
So, there is a shared responsibility and, of course, the blood industry itself has a paramount interest and concerns, and we have been developing better coordination with the blood industry in the whole program of monitoring increasing supply.
I think it is simply an observation that the leadership has migrated from group to group over time based on the interest, the availability of resources, and where the problems seem to lie.
So, my answer to you, Paul, is that it has simply been an evolving situation. I could not tell you that there is any one place where the aggregate responsibility lies. That is not true today. Today, we have a shared responsibility which is distributed to a large number of entities, and most of what we are trying to do today under the broader aegis of the Department is integrate things.
We are trying to integrate the monitoring that the blood industry is doing with the monitoring that the government is doing, and to link it up with the intervention strategies. We have moved further down the pathway of coordinating appeals for blood.
In other words, we now have mechanisms whereby high public officials become aware of shortages, they have press statements ready to go and they can do it on a dime, and they have done it. We didn't used to have that situation.
We are also working toward the possibility of having a better coordinated system for prediction of shortage and addressing it through a more coherent system to identify where blood surplus might be and move the blood.
So, it is simply evolving and I think the name of the game today is coordinating effort in a more rational way, as well as improving the systems themselves.
DR. NELSON: Harvey.
DR. KLEIN: Alan, I would like to commend you. I think you are developing a very helpful system for an early warning and early monitoring system. It looks like from your slides and presentation that you are think about what the efferent arm, the mechanisms for intervention might be.
What I wonder, however, is anyone, to your knowledge, either within FDA or anywhere else, look at the impact of shortages in a quantitative way, looking at morbidity, mortality, the quantity of surgery that is delayed, for example, or Rh-positive blood going to Rh-negative individuals, a number of the other measures, so that we can really know what the impact of shortages has been on the American public's health.
DR. WILLIAMS: I think the attempts to do that have been very limited. Those that I am aware of, the NBDRC biannual survey asks participants how many days in the course of the year there was delay of elective surgery, so that is one parameter.
The HHS program captures some of the same characteristics I presented here I believe as write-ins. Also, there would be some data for the 26 participants in that program. Then, there is the data potential that TRANS-Net would have if this was implemented nationwide and we captured the right variables, I think you would have that information.
Then, there is one additional datapoint which I think is a very powerful one, the American Hospital Association presented data at the advisory committee, and I can't remember the exact numbers, but I remember it was over half, I believe, of the hospitals had presented information indicating a shortage that had impacted patient care at some level, and I think that was a very powerful statement.
So, I think it is important information. I think there are various mechanisms that could capture it. I think this is one, but we don't have the data readily available right now.
DR. STRONG: Just a comment. As a participant in this program, one thing I would encourage is that we try to reduce the number of data collection points. We are involved in all three of the ones that you have mentioned, and the HHS one is certainly the most time consuming of those and probably has been the least productive in terms of data. At least you are demonstrating that there are differences from day to day in terms of supply.
Our experience is that this one has been easier to work with and seems to be giving us some data that is useful, but I would certainly encourage that there be some cooperation to try to consolidate these efforts.
DR. NELSON: How frequently do blood centers and transfusion centers report, is this daily, weekly, or what?
DR. WILLIAMS: This would be a daily report at a set time of the day on weekdays, so it would be a daily reporting.
DR. NELSON: So, five days a week, five times a week.
DR. WILLIAMS: Correct.
DR. FITZPATRICK: Alan, I think it's an elegant way of collecting data, and like we said before. One comment that I have had in the past, and still have, is what is the plan as to what to do with the data now that we have got the mechanism and the tool, and are collecting the data.
One thing you have noticed already, that was a little intuitive, but you needed the data to support it, was blood centers are more sensitive than transfusion services possibly or it may turn out that way.
So, it may turn out that getting 5,000 transfusion services to report may not give you much more data than what you are already getting, depending on what you plan to do with the data.
The other thing is on the management piece. Jay mentioned that, and I would support what Jay was saying, that the cooperation, we are seeing more cooperation between agencies and discussion of how to deal with this problem and how to deal with it and that shortage is a safety issue, but both comments from you and from Jay on if we see an overage in an area and a shortage in another area, moving that blood around is not a mechanism available to HHS or FDA.
So, how is it, do you envision a way of doing that, is that an anecdotal comment, or are you just suggesting that if the data is available, people will see it and do something? I am not sure where that is going and that concerns me a little bit.
DR. WILLIAMS: Well, you are also referring to I think the efferent arm that Harvey mentioned. There is no absence of thoughts as to how blood could be moved around the country. The problem is that is business practice, it is not under control of any single entity.
I think looking toward the future, I know DoD and others have spoken in favor of some strategic supply depots which could be established around the country, so that if you excess inventory either at the major blood centers distributed around the country or at strategic supply depots, that could form a clear source of supply that could be sent into a regional area that was hit hard at any particular time. I mean that is a future possibility, it is not one that occurs today.
I think largely today, in fact, we have a situation where the best thing we can do is make the data available to the entities that are capable of supplying the blood, and that is in the private sector.
DR. GOLDSMITH: I have two questions. One is do you have any plans to extend this program to the noncellular components of blood and plasma derivatives. I know that some of the other agencies have been interested in this.
Second, you don't really have clear criteria for shortage and what does shortage mean, and to look at some of the impacts of shortage. Is there any plan to develop these criteria in a more robust fashion?
DR. WILLIAMS: To answer the first question, within this specific system, there is not an immediate plan to include noncellular components. It is a little different dynamic there, not that the system couldn't measure it, but this was designed to meet a need in the whole blood and blood component area.
With respect to the second component, I guess I don't fully understand the question because I think we are capturing the criteria which define a shortage, and the message I was trying to get across is we are not dealing with a dichotomous variable as an endpoint here, shortage/no shortage.
We are trying to characterize the shortage, weight it in a medically relevant fashion, and then define the shortage according to those weighted criteria. That is the goal.
DR. NELSON: Jay.
DR. EPSTEIN: I just wanted to comment that FDA has already, for several years, been gathering monthly distribution data for the major plasma derivatives, and that the PPTA does make public the monthly report of aggregated data of availability of the derivatives, and we think that that effort has served the patient community, as well, in helping them understand the dynamics affecting product availability.
So, we have been doing it, but it is not done on the daily basis that is I think somewhat more needed, dealing with the transfusable components.
Open Public Hearing
DR. NELSON: The next item is Open Public Hearing if anybody wants to make any comments or ask questions about the presentations to date. Are there any comments?
DR. SMALLWOOD: For the record, I would like to announce that Dr. Matthew Kinard from CDC has been screened to participate as a guest of the committee for these proceedings.
DR. NELSON: The next item on the agenda will really take us the rest of the day, and that is discussion of the West Nile situation.
The first speaker is Dr. Hira Nakhasi, who will introduce the issue and give us an update of current and previous activities of the FDA.
I. West Nile Virus Donor Testing
A. Introduction and Update of Previous Activities
Hira Nakhasi, Ph.D.
DR. NAKHASI: While the projectionist is putting the slides up, I would like to start the session because this session is going to be, as Dr. Nelson said, will be the whole day session. I think by the time you finish this, I am sure you will be exhausted
The topic of the discussion for today is the validation of nucleic acid and serological tests to screen blood and plasma donors for acute infection with West Nile virus.
The issues which we are going to be seeking advice from the committee today are the following. First, is to seek advice on the design of scientific studies which are needed to validate NAT, nucleic acid test, as well as possibly IgM test that is the serological test for West Nile as a blood donor screening test.
Second, we would like to also seek advice on whether available data on the clearance of viruses in the manufacture of plasma derivatives are sufficient basis to obviate the screening of source plasma donations for West Nile virus.
The third is obviously the implementation issue with regard to strategies to limit West Nile screening to particular locations and time are appropriate.
Before we go to asking specific questions, I would like to give you some background information and also what FDA has been doing for the last year since the epidemic broke. The first slide shows you basically what is West Nile virus because I am the first speaker, I would like to give some brief background.
West Nile virus is an envelope RNA virus. It is a mosquito-borne flavivirus, infects birds primarily, but humans are incidental hosts. The majority of the human infections as asymptomatic, some 20 percent show flu-like illnesses, 1 in 150 infections result in meningitis or encephalitis.
As I said last time, also, older age is by far the most significant risk for the severe neurological disease. Viremic period can occur 2 weeks prior to symptoms and last up to one month after the initiation of infection based again also on the previous studies which were publishes a long time ago.
In the 2002 U.S. outbreak of West Nile, besides mosquito transmission there were other routes of transmission identified, for example, blood transmission, transplantation, breast-feeding, occupational uses, as well.
The magnitude of risk for West Nile from transfusion is at the moment unknown how much it can be. Another interesting phenomena of this is that the titers of this blood, you know, virus, is not as high as the other viruses like HIV and HCV, and the important thing is viremia resolves rapidly after the seroconversion to the IgM.
IgM can persist for a long time period, in some cases up to one year, however, no chronic state of West Nile infection has been reported so far.
The current status you will hear more about from Dr. Lyle Petersen from CDC, who is going to be presenting this afternoon. In the last year of 2002, there are 4,000 approximately cases reported of West Nile cases, 263 deaths, and 2,741 cases of West Nile meningoencephalitis.
Thirty-nine states including D.C. is endemic for West Nile virus. The estimated risk for transfusion was calculated last year, before the last year, in 1999, between 1.8 to 2.7, and that prediction came a little bit higher in the last year endemic regions, which is highly endemic regions, it was something like 16 per 10,000 with a mean of 6 to 8 in 10,000.
During that epidemic period of August 28, as of January 3rd, now with the change as of March 3rd, 61 possible transfusion-transmitted cases were reported, 21 were confirmed from 14 blood donations, 19 were not transfusion related, and 21 were incomplete because they could not get the follow-up samples.
What did FDA do during this time? FDA posted a lot of alert notices basically August 17, and before the epidemic, you know, when we knew about this, the epidemic had started, vigilance in excluding symptomatic donors prior to actual reporting of transmission.
October 3rd, we issued a notice for facilitating development of donor screening and supplemental tests.
We had congressional hearings. We issued a guidance document around October 2002, and we have been also consistently talking to CDC, other State Public Health Departments, blood organizations, and other components in HHS about the epidemiological case investigation of transfusion-transmitted cases, providing advice on the deferral of donors and withdrawal of products.
Most of you know that we had a West Nile workshop in 2002, where various issues were discussed, basically, methodologies which will be suitable for donor screening, proposed studies on prevalence in donors, and FDA interaction with how the tests will be developed, strategies for inactivation of West Nile in plasma derivatives, and implementation issues.
The summary was presented last year, in December, of Blood Products Advisory Committee meeting, so I don't want to go into detail.
The regulatory pathway, FDA's current thinking is that it is with the donor screening and supplemental tests for West Nile will be reviewed as BLA, the IND process approval, and software and instrument portion of the application will be a 510(k) submission. Dr. Mary Beth Jacobs explained earlier that depending upon what type of the instrument it is, it will depend on whether you need a separate 510(k) or it is part of the BLA application.
We still believe that this is a major level of concern because it is used in the donor screening.
We will also hear today the various regulatory pathways regarding the study design and basically approval mechanism. Dr. Robin Biswas will talk about the various issues related to approval mechanism. Dr. Indira Hewlett will talk about the clinical study design, basically, validation of clinical sensitivity of NAT and IgM assays, clinical specificity, and validation in the absence of reference assays, and reproducible studies.
In addition, she will also talk about unit and donor management, algorithms, and how we need to follow up in those guidances.
We also hear our efforts in the development of reference panels for lot release testing, as well as the qualification panels of well-characterized and pedigree samples to establish relative sensitivity of NAT and IgM assays.
We will also hear from Dr. Maria Rios about development of in-house TaqMan PCR and IgM test assay because we would like to understand the distribution of this virus n blood components, infectious dose, and the viral tropism.
You will hear from the industry folks today about their efforts about the nucleic acid tests, as well as serological tests for whole blood, source plasma, and supplemental tests.
You will also hear from Red Cross and CDC about their prevalence studies and to answer some of these questions, because of the time constraints, I don't want to go through the whole slide because you will hear from Sue and Lyle more detail about their efforts in the prevalence studies.
We will also hear from the source plasma industry, as well as the testing of source plasma donation, how it is to be done, as well as the viral inactivation process. Dr. Mahmood Farshid will also talk about the issues which are relevant to the viral clearance, basically, whether we need to have viral clearance specific for West Nile or we do we need an upfront screening of the source plasma as a source of added safety in the NAT.
We will also hear from Dr. Alan Williams about the supply management and triggers for West Nile testing in the event if the test is not available at the time of epidemic or if the test is available, how this test will be implemented and what will be the trigger, basically because, as we know, it is a seasonal epidemic, it is not throughout the country all the time, and also the issues with the minipool versus individual NAT testing and testing in the presence of other related viruses.
So, finally, the questions which we will be posing to the committee are as follows:
First of all, we will ask the committee to comment on FDA's proposed criteria for validation of West Nile NAT and IgM assays for donor screening.
We will also ask the committee's input whether they agree that the product and process-specific clearance of West Nile agent, as opposed only to marker virus should be demonstrated in order to adequately assure the safety of plasma derivatives.
We will also ask the committee do the committee members agree that screening of all plasma for fractionation for West Nile would add a safety margin in the manufacture of plasma derivatives.
We will also the committee to comment on the scientific validity of possible strategies to limit West Nile screening to particular locations or times depending upon the epidemic surveillance information and test availability.
So, I think with that introduction, I think I would like you to sit back, relax, and listen to the whole presentations the whole day, and by the end of the day, you will be ready to answer some of these questions.
Thank you very much.
DR. NELSON: Thank you, Dr. Nakhasi.
The first series of presentations are some industry presentations on NAT testing. First, is Dr. Jim Gallarda from Roche.
B. Industry Presentations
Update on NAT Testing
James L. Gallarda, Ph.D., Roche
DR. GALLARDA: There are some great cartoons on West Nile virus on the web site, so I pulled one down because it's relevant to what we think is going to be happening in the upcoming mosquito season.
This is a cartoon where a mosquito is on a deli bar sitting next to a patron saying, "I'm the mighty mosquito, carrier of the West Nile virus, spreader of fear, humbler of men. What do you have to say about that?"
Our intention is to provide a technology that will do exactly this for the upcoming mosquito season.
My presentation is entitled West Nile in Style - Roche Update.
I would like to give a review of what we discussed last time in the November workshop for West Nile virus and to compare what progress we have made since that time.
The distillation of our talk last time for our concerns about implementation of a nationwide IND for screening for West Nile by PCR.
No. 1, as it was true in November to this very date, we have been really pleased with the collaboration amongst the stakeholders in trying to figure out how to do this, what normally takes three years, to collapse it down into nine months.
The capital, training, and support costs that we are encountering right now, as we said, are bearing true that this is substantial. The time frame, of course, is unprecedented.
We had, in November, several concerns about how to execute a multi-site IND although these issues have been remedied in discussions with the agency, as well as technology transfer from a research type application to a full GMP-regulated in vitro device.
From that time in November until a critical milestone that Roche reached in the end of January, I would like to just review what we have been doing.
Initially, we had several options to explore how best to provide a PCR system that would provide evidence of safety and efficacy for screening for West Nile, and we are looking at our current AmpliScreen technology very closely, which really has definitely got some advantages and some disadvantages.
The advantages are, of course, it has been FDA approved, the customer can do single unit or minipool testing for AmpliScreen. There is an installed customer base throughout North America.
We looked at a couple of options in the semi-automated systems and compared that to other options that we had available which were fully automated technologies.
The advantages, of course, with fully automated technology for sample prep is that we have the option to do total nucleic acid extraction, which allows us to simultaneously extract, not only RNA viruses, but DNA viruses or nucleic acids from infectious bacteria or other total nucleic acid targets, so this was something that we felt was rather important.
We were looking at the potential to implement our next generation platform, which is the TaqMan chemistry kinetic PCR, which really is a much more advanced state of PCR.
We also were trying to look at automation features that would truly provide for pos ID, that is, a human being does not have the opportunity to, by human error, introduce a mismatch of a donor ID with a particular PCR result.
Of course, we were very concerned about improving the amount of user intervention in a full system technology. What is hidden on the bottom of the slide is that we also are aware of the agency's request to provide what we call "full process controls," which mean that a control in a kit would mimic the virus, that is, from the lysis step through detection step, the control would assess the system's ability to perform as intended.
So, we have decided on January 31st was to commit to the fully automated system. We felt that it was important to take that step now to move into the future, which really is a future for full automation.
The system that I will be talking about, which is really in its infancy right now, but is what we intend to go forward with for screening for West Nile, consists of the current Hamilton MicroLab AT/plus 2 Pipettor, which will create pools of the plasma for extraction and PCR amplification, and a COBAS AmpliPrep, which is the automated device for total nucleic acid extraction, coupled with a COBAS TaqMan Analyze which does the kinetic PCR for West Nile virus.
In this particular iteration, we show what we call a docking station, which allows for samples and results out with no human intervention. Because of the timeline for implementation of West Nile, we are initially not using this feature, so these two boxes will be separate.
High-level overview of what we are calling now the TaqScreen West Nile Virus System is we will use a pool size of six on, as I said, the Hamilton MicroLab AT/plus 2 Pipettor.
The system features ready-to-use reagents. There will be automated sample extraction on the COBAS AmpliPrep, automated addition of the master mix to extracted pools by the COBAS AmpliPrep, so the user does not have to add the master mix to the extract.
Because we are leaving out that docking station which couples the two devices together, there will be a roughly 10-second user intervention to transfer extracted samples from the AmpliPrep into the COBAS TaqMan.
The separation of the two instruments allows for flexibility in the instrument layout. A laboratory can choose to put all devices, the Hamilton Pipettor, the COBAS AmpliPrep, the TaqMan into one room or to separate those into rooms depending upon the structure layout of the individual laboratory.
The other thing that this automation system does is it provides for a rapid introduction of new analytes since there is a generic sample preparation that is applicable to other infectious agents that may be on the horizon.
The throughput of a module, which consists of a Hamilton Pipettor, an AmpliPrep, and a TaqMan is about 1,200 samples in a pooled format per module in a two-shift scenario. All of our clinical trial sites have, at a minimum, two sites, and depending on the volume encountered in the course of a year, we increase that number of modules to fit their needs.
This shows a COBAS sample prep in the context of a human being. This is a relatively moderate size instrument. It provides for full automation through a generic sample prep method. It clearly reduces the potential for human error in manual sample extraction procedures.
It also allows for increased operator safety since the operator is not doing the extraction themselves.
AMPLILINK is the software that will be attached to a personal computer driving the system. It will be a Windows NT interface and it provides for positive identification of the sample with the test results.
What I would like to talk about are some of the features of the AmpliPrep.
The sample preparation kit comes with cassettes that contain the reagents necessary for sample extraction. These are, as I mentioned, ready-to-use reagents, there is no mixing of any type. The operator simply takes a reagent rack, which has a unique barcode associated with that rack, and each of the reagents has a barcode set of information that as the rack is inserted, a barcode reader reads all of these, so there is a link between the individual reagent components and the particular run.
There are LED indicators up here which indicate the status of the run. If it's green, it is in operation; if it is not in operation, it will flash orange. The LEDs allow the operator to feed the instrument with a continuous flow of reagents in an ongoing process.
The reagents themselves have self-sealing caps for increased stability. This allows us to do a run if we need to take the reagents off the instrument to put in the refrigerator for the next day, we can do that.
The samples are pooled in the Hamilton Pipettor into these S-tubes. These are 1.5 mL, small, microtubes that have a special cap on them. Once the samples are loaded into the AmpliPrep, there is no human removal of the cap. This is done automatically by the robot onboard.
Each of the pools has a unique barcode clip associated with it. As the operator slides these racks into the instrument, the barcode reader reads all the information associated with each of the individual pool and is captured in a data management server.
This is where the action is for sample prep. There is a device called a specimen preparation unit, or SPU. It consists of a plugged specimen tip, so to prevent contamination by pipetting, this is a plugged tip. It comes packaged in this SPU.
I mentioned the S-tubes or sample tubes are these 1.5 mL tubes that have a special cap that is gripped by a robotic device within the AmpliPrep extractor to open the tube for sample pipetting.
All of the extraction process goes on in these specimen reaction chambers. Any waste that is generated in the process is deposited in the central chamber. There is a lot of engineering features designed in this SPU to mitigate contamination control.
On the COBAS TaqMan side, we are going to be using a COBAS TaqMan 96. It's a real-time, closed tube, amplification and detection device.
The COBAS TaqMan has four independent thermal cyclers here and four holding spots for K-tubes, which are the tubes that have the extracted sample and the master mix added.
Another view of the COBAS TaqMan. Here are what we call K-carriers or the tube racks for the actual PCR event. Sample racks can be slid in up to four slots, which contain the extracted sample and master mix, and the four thermal cyclers onboard. A robotic arm is responsible for moving these at the appropriate point to the correct thermal cycler.
This simply shows the K-tube or the kinetic tube that is the device that is loaded into the COBAS TaqMan for PCR.
This is a slide that simply shows the technology associated with kinetic PCR. It is called fluorescence energy transfer mechanism. It simply is a mechanism that shows that through ongoing cycles, there is a cleavage event of a reporter molecule that results in fluorescence with each cycle, hence, it is called kinetic or real-time PCR.
If we take the TaqScreen West Nile Virus looking at the CDC lysate, these are the types of growth curves that are generated with each cycle ranging from zero to 60 cycles when you see a growth curve of any type, and that indicates that you have a positive result.
We have 100 percent detection at 2.8 x 10-7 dilution of the CDC lysate, and we see at 107 dilution about a 62 percent detection rate.
If we take the TaqMan portion of the West Nile Virus system and we look at a BBI panel ranging from 1,000 to 25 copies/mL, this is the type of growth curves. A high-titered sample, 1,000 in this case, is shown in red, and the 25 copy/mL is shown in aquamarine.
This is our benchmark to which we are developing now and optimizing a full system from the Hamilton pipetting event through sample extraction, through TaqMan PCR.
What I would like to show you now are some slides that really illustrate where we are or where we have been in the total system development of this, and where we expect to be by the time we go live in July.
Our early full automation dataset or benchmarking against 100 copies/mL BBI panel member shows growth curves that range from this type of event to this type of event. At this point, we are getting a 92 percent detection rate for this particular dataset, and a dataset is consisting of a particular process on the COBAS sample prep and formulations associated with PCR.
We are moving by changing our process and formulation, which is currently ongoing. We are starting to move these growth curves to a higher level of fluorescence. Here, we are getting a nice 6 percent detection rate at 100 copies/mL.
Really, the sensitivity in the hit rate will depend on how a negative population behaves in the TaqMan system. Here is a group of 120 samples, which have a very nice flat line growth curve that will establish our baseline in the future for deriving a certain performance in the system.
Interestingly, of 120 samples, one of these had a profile classic with a West Nile virus positive sample. Since this came from California in the wintertime, we think it is unlikely that this is a West Nile virus infected person.
On root cause analysis, we identified that a technician had used a pipetter that had previously been used for controls in this event, and we believe this is a contamination of it.
Critical milestones going forward. Our plan is to file our IND in April with the agency. We have what we are calling a Phase I program, which consists of training, installation at three high prevalence sites in the U.S., and to initiate Phase I studies in the March to May time frame.
These will focus on unlinked studies that will concentrate on work flow and training of the operators. We will conduct some non-clinical performance studies with the system early on. Mainly Phase I allows us to refine the system, so that for what we call Phase II, we will have a polished system.
Phase II training and installation will include nine additional U.S. sites, five Canadian sites. We will initiate studies July 1, 2003. This will be a linked study and will be used to demonstrate safety and effectiveness in detecting West Nile.
Current issues. This slide should look familiar. We continue to be grateful to the inner stakeholder collaboration that we have had with the agency and the blood screening community.
We have significant capital, training, and costs. Especially now, we are implementing our rollout phase as we speak.
The new thing here is that because of the accelerated timeline, we are in the process of optimizing the entire system which we think really introduces the next generation in PCR screening, which is to provide full automation.
These are our partners in Roche globally. We have instrument labs in Switzerland and Germany, our manufacturing site in Belleville, importantly, our affiliates in both Indianapolis and in Canada, and an outstanding research group in Alameda and development group in Pleasanton.
DR. NELSON: Thank you very much, Dr. Gallarda.
Any questions or comments? Jay.
DR. EPSTEIN: Thank you very much, Jim.
Could you just clarify whether the sensitivity at 100 copies/mL is for a neat sample or one diluted 1 to 6?
DR. GALLARDA: Right now we define the 100 copy/mL sensitivity as an analytical sensitivity. If you have a sample, individual or otherwise, if there are 100 copies in 1 mL, the detection rate is at that level.
So, at a 95 percent detection rate at 100 copies/mL, we would expect a sample coming in at 1,000 copies/mL diluted 6-fold to be reliably detected.
DR. NELSON: Is your HIV-HCV NAT, does that also use a pool of 6?
DR. GALLARDA: No, the currently licensed system uses a pool size of 24.
DR. NELSON: So, do you conceive of this, then, as a multiplex assay or is it separate from the HIV-HCV testing?
DR. GALLARDA: This is an entirely separate technology from the licensed HCV and HIV, where we really want to migrate towards is full automation and ultimately the single unit.
DR. ALLEN: Given that it is a totally separate system, blood centers that want to implement testing for that would need to purchase or at least additional equipment, find laboratory space for that, and integrate that into their routine flow?
DR. GALLARDA: Right. The sites that we are working with, both in U.S. and Canada, those sites are currently users of the AmpliScreen technology and have production setups for HCV and HIV.
Additional space is used for this system, and depending on the laboratory's structure, we can place it in a large centralized room or in multiple separate rooms.
As far as the capital costs, we predict it to be under cost recovery, and the cost recovery would include all costs that would support a large-scale IND.
DR. FITZPATRICK: The IND is for pools of 6 only?
DR. GALLARDA: We will be doing pools of 6 and single unit PCR testing.
DR. NAKHASI: In the Phase II, which you said that there will be linked studies and there will be 9 sites, where will these sites be located, Question 1, and Question 2, which I guess I asked you also last time, in case the epidemic is out someplace else, what are you planning to do, how are you going to get the data?
DR. GALLARDA: That is a very good question. We chose 3 high-prevalence sites for the Phase I rollout, which was primarily designed to allow us to get workflow established, operator training.
These 3 sites include a site in Florida, Indiana, and in Texas. The other 9 sites on top of those include the geographic distribution we have for the current AmpliScreen user base, which ranges from Puget Sound in the Northwest throughout 2 sites in California. We have Kansas City, we have Minneapolis, we have South Bend, Mississippi Valley, Houston, Texas, New York, Orlando, Gainesville. It is a broad distribution across the geographic U.S.
The majority of those are from the Mississippi Valley to the East, which is where historically, West Nile has had higher activity.
As far as trying to implement a faster rollout, I think it boils down to discussions with the agency on what kinds of activities might be acceptable to put what is really marginally being produced into a GMP environment. This is our goal, if there is anything that can be done that is mutually agreeable between ourselves and the agency on now to accelerate that process.
DR. NELSON: Thanks.
Next, is Christina Giachetti from Gen-Probe.
Cristina Giachetti, Ph.D., Gen-Probe
DR. GIACHETTI: Good morning.
This shows the objectives for the program and is to develop and manufacture a TMA-based assay for detection of West Nile virus in blood, plasma, and organ and tissue donor specimens.
The program has two phases and will support a Phase I clinical study to determine RNA reactivity in archived linked samples collected in areas of potentially high West Nile virus incidence rate during 2002.
The Phase II clinical protocol will allow for widespread donor screening, and it is anticipated that this trial will be initiated in July 2003. Dr. Sue Stramer will talk more about the status in her presentation.
These are our goals for assay development. We have analytical sensitivity goal of at least 95 percent detection at 50 copies/mL, detection of genetic variants belonging to Lineage 1 and 2 with similar sensitivity, a specificity higher than 99.5.
We are going to add an internal control to validate each reaction, and the system is going to be completely compatible with the Procleix eSAS (semi-automated) and the TIGRIS (fully automated) instrument platforms.
The Procleix eSAS system is our current licensed platform that is used for screening for HIV and HCV.
Here is a schematic of the Procleix Semi-Automated System. We use TECAN pipetter to pipette calibrators and target capture. The sample processing is done in our target capture system, and then we have amplification and detection that occurs in regular water bath, and finally, the readout is in the luminometer.
This shows our basic technology. I am not going to go into too much detail because most of you are familiar with that, but if you have questions, you can ask me.
For specimen processing, we use target capture/magnetic microparticle separation. This allows for viral lysis, nucleic acid capture, or removal of unwanted specimens, so that we are able to isolate the RNA which is ready for amplification.
Amplification is with transcription-mediated amplification, or TMA, which utilizes reverse transcriptase and T7 polymerase.
The detection is with the hybridization protection assay that utilize acridinium ester labeled probes.
The detection we do with the dual kinetics analysis to be able to differentiate between our internal control signal and the target West Nile virus amplicon.
I will be talking the aspects of performance regarding the specificity in normal blood donor specimens, specificity in the presence of other blood-borne viruses, analytical sensitivity, and clinical sensitivity.
Specificity in normal donors was done testing for 175 frozen plasmas and 100 fresh plasmas. The initial reactive rate was zero percent and initial internal control proliferate was 0.69 percent with no repeat internal control failures. This was set for internal controls very similar to the ones that we observed for the HIV assays.
We tested the specificity and other blood-borne pathogens, and this include hepatitis A virus, cytomegalo, Epstein-Barr, hepatitis B, hepatitis C, HIV, and HCLV-1 and 2. We tested several numbers of specimens for each of them. In all cases, the assay was non-reactive, which gave us a specificity of 100 percent in this type of populations.
Our statistics regarding average single to cutoff for internal control and analyte were very similar to what we observed in the normal population.
We determined analytical sensitivity using different sources of materials. We tested an in vitro transcript. The transcript was quantitative at optical density. Then, serial dilutions of the transcript were made from 300 down to 0.1 copies/mL.
We tested 40 replicas at each copy level to determine our positivity rates at each level, and then we used a probit analysis to determine our 95 percent and 50 percent detection limits.
The 95 percent detection limit is 8.2 copies/mL and the 50 percent detection limit, 2.9 copies/mL.
This shows similar study performed in this case with the CDC viral panel. This panel was provided to us by Dr. Robert Lanciotti at the CDC. This panel is quantitated in PFU/mL. We performed serial dilutions of this panel going from 1.1 down to 0.0015, in plaque-forming units/mL.
Again, we test about 20 replicas for each copy level, and we used probit analysis to determine again 95 percent and 50 percent detection limits with our assay.
The 95 percent detection limits was 0.008 plaque-forming units/mL, and the 50 percent detection limit is 0.001 PFU/mL.
This shows determination of analytical sensitivity using the West Nile Virus BBI panel that belongs to Lineage 2. This panel contains 13 members, which expands from 1,000 to 30 copies/mL and some negative panels.
We tested 4 replicas of each panel member and we were able to detect with 100 percent detection all copies levels including the one with 30 copies. The negative panels gave nonreactive results.
What we did to be able to determine our detection limits with Lineage 2 panel, what we did again was serial dilutions, tested different replicas of different copies levels, and again determined 95 percent and 50 percent detection limits, which are 9 and 0.45 respectively.
So, all along, our analytical sensitivity results correlates well regardless which is the source of materials we are using.
This showed clinical sensitivity data. What we did here, what I am showing is reactive results obtained from the CDC case investigations.
Dr. Sue Stramer, as well as Lyle Petersen, will talk more about the study. Here, I would just like to focus on which is the performance of our assay in the samples.
We received 384 samples of blind specimen from Dr. Sue Stramer, and we did the testing. Here are the different cases, and our results in the Procleix West Nile Virus as duplicate samples or at 1 to 8 or 1 to 16.
Also, we tested the TMA confirmatory assay that uses different primers and probes, and our results are compared with the CDC TaqMan assay. For the first 9 cases, we have completed a correlation of the results between our neat samples diluted 1 to 8, 1 to 16. Similar results were obtained with our confirmatory assays and with the TaqMan CDC assay.
For two specimens, what we obtained was positive results while the initial results for the CDC TaqMan assays gave negative results, but was positive once used a higher sample input.
Finally, there was one specimen that was positive in our assay, but was negative the CDC assay. When we tested the specimen diluted, sometimes the results were positive and sometimes the results were negative. This indicates that this sample has a very low copy level.
Here is a summary of our results for the performance of our assay so far. We have no cross-reactivity with other blood-borne viruses. The initial reactive rate in normal blood donor population was zero percent for an n of 575, for 100 percent specificity.
Analytical sensitivity shows 95 percent detection between 7 and 12 copies/mL for Lineage 1 and Lineage 2 virus, and between 0.005 and 0.042 PFU/mL for the CDC virus panel.
The 50 percent detection rate were between 2 and 4 copies/mL for Lineage 1, 3 and 6 for Lineage 2, and between 0.001 and 0.003 PFU/mL for CDC panel.
Our clinical sensitivity data shows that the Procleix West Nile Virus assay has a clinical sensitivity equal or better than the CDC TaqMan Assay.
All the reactive results obtained with the Procleix West Nile Virus Assay were also confirmed with an in-house TMA assay using an alternative amplification region.
So, our general conclusions show that the performance to date meets design goals for specificity and sensitivity, and that the results support feasibility of pool testing.
Acknowledgments. This project is funded in part with federal funds from the National Heart, Lung, and Blood Institute.
DR. NELSON: Thank you very much.
DR. STRONG: What was your input volume for your analytical sensitivity studies?
DR. GIACHETTI: The input volume is the same volume as is used in our current assay studies, 500 microliters.
DR. NELSON: You didn't test other flaviviruses, such as St. Louis or Dengue?
DR. GIACHETTI: No, we haven't tested them. We plan to test them. It is just a little hard to get the samples. There are a lot of regulations, but we are trying to get them, yes.
DR. FITZPATRICK: Where are you with plans for automation of the method?
DR. GIACHETTI: This is a semi-automatic method and in parallel we are working with our fully automated TIGRIS platform right now. We also have a second plan for this TIGRIS instrument, and we are moving along in that instrument system, as well.
DR. FITZPATRICK: Do you have any plans for an IND by July with TIGRIS?
DR. GIACHETTI: No. This system will not be by July on TIGRIS.
DR. LAAL: Are you considering getting more samples from countries that have a much higher incidence like Israel or somewhere to see if you will get more of that last patient that you had, which you could detect by some assays, not by the others because your sample size is pretty small.
DR. GIACHETTI: Pardon me?
DR. LAAL: Your size is pretty small, your sample size. If you go to countries where disease are endemic, you are likely to get a much wider range of viral titers.
DR. GIACHETTI: I think it's a good idea. We don't have plans right now to get a sample from other countries mostly due to the very short timeline for this project, but this assay is going to be used for linked studies where we are going to look at prevalence during the last year epidemic season in U.S., yes.
DR. NELSON: Thank you.
Next, is Dr. Bruce Phelps from Chiron.
Bruce Phelps, Ph.D., Chiron
DR. PHELPS: Dr. Nelson, Dr. Smallwood, members of the BPAC Committee.
Last November, at the workshop, I provided an overview of the ongoing program at Chiron to produce some reagents and materials that could be used in support of development of West Nile virus assays.
Today, I would like to update you on the status of the program at Chiron in particular with respect to development of supplemental assays for West Nile virus and providing other critical reagents in support of the Procleix West Nile virus assay development ongoing at Gen-Probe.
There are basically four main areas of forming objectives for our program. We are, as I mentioned, developing a TaqMan assay for West Nile virus application as an alternative NAT to support the TMA assay under IND.
We are also looking at propagation of West Nile virus in large quantities for use as controls or calibrators to standardize the assays. In addition, we are looking at development of other assay standards which can be used, not only for nucleic acid testing, but also for developing serological tests. Indeed, we are looking at reagents that may be useful for serological assay development.
As I mentioned, we are developing a qualitative West Nile Virus TaqMan assay at Chiron to support the Procleix West Nile Virus assay development program.
The assay will be full validated by our current plan and may be used as a supplemental alternative NAT during Phase II trials to be initiated at the advent of the 2003 mosquito season in July.
We have used the capsid region of the West Nile virus genome as the target for amplification in this assay. It is a 965 nucleic acid nucleotide RNA transcript from the capsid region is also included as the internal control in the assay.
Currently, the sensitivity is between 10 and 50 copies of purified RNA transcript from our current data.
This is some data that we have generated using the CDC viral panel, viral lysate panel that I think every one of the manufacturers who are currently optimizing assays are using this panel, so it gives a fairly good comparison.
You can see that down to 4 x 10-7 dilution we were able to detect 2 out of 3 of the triplicates in that particular assay with our TaqMan assay. This is about 10-fold better in terms of its sensitivity for dilutional panels relative to the CDC TaqMan assay, and is equivalent to sensitivity at least within a half log sensitivity of where we are with the current TMA assay development.
We are currently evaluating two methods for our target isolation, magnetic silica absorption and magnetic beads for oligonucleotide capture. At this point, we haven't really locked in on either one of these methodologies yet. We are evaluating them both.
In terms of propagation of the West Nile virus, using the New York strain, the 385-99 West Nile virus strain that was originally isolated from New York, we are propagating in Vero cells in BL3 facility in our laboratories.
We have currently obtain viral titers of up to 107 plaque-forming units/mL in several different viral cultures. I think we have about 6 or 7 so far that we have completed in large quantity viral cultures.
We are also developing a quantitative TaqMan assay for use in-house that will be used to estimate the viral genomic equivalents/mL and using supernatants of these virally infected cells.
The West Nile virus inactivation process that we are using is actually based on a validated procedure that we have used internally for HCV virus. It's an heat-inactivation protocol, and we have confirmed that it does inactivate West Nile virus by using Vero cell culture and infectivity studies in Vero cells, and that is how we are providing inactivated virus for use in controls and calibrators for potential use with the TMA assay.
At this point, we have succeeded in producing viral suspensions of heat-inactivated virus, which are the equivalent of 109 plaque-forming units of pre-inactivated virus, and these have been provided to Gen-Probe for use in internal controls or potentially for other purposes in developing the assay.
We have also developed three RNA transcripts of different genomic regions of the West Nile virus, and you can see the three areas that we have looked at here pretty much spanning the entire genome, the 5-prime noncoded region, the core, and the pre-membrane and membrane regions, the envelope and nonstructural regions 1 and 2A, and similarly, the nonstructural 5 and the 3-prime noncoded regions.
These are the nucleotide sizes for each of these transcripts. These have been fully characterized and have also been supplied to Gen-Probe for use in developing their assay.
Our ultimate goal is to provide a genetically engineered non-infectious full-length West Nile virus genome, which could then be used as a standard for development of West Nile virus assays and make this broadly available for anyone who would want to use this as a standard.
We are continuing to pursue this in our program in Chiron. Once it is fully characterized, as I said, this would be made widely available for use as a standard in nucleic acid test evaluation.
In terms of reagents for serological assay development, we are in the process of cloning and providing expression in use of West Nile virus recombinant proteins. We have four that are listed here on this slide, again, the pre-membrane and the envelope regions, envelope protein, capsid protein, NS1 region, all from the New York strain.
Some of these have been purified to the extent that we can do some pretty good evaluations of them ongoing immediately, others are in the process of essentially being scaled up and improving the cloning and expression in the U cells.
These recombinant West Nile virus proteins can facilitate development of supplemental or diagnostic tests to detect IgM or IgG antibodies in the blood and serum of infected individuals, and we plan to evaluate these in multiple formats.
They could also be used to generate monoclonal antibodies for antigen test development if that should be something that we would want to proceed with in the future.
We would like to acknowledge Dr. Laura Kramer and Alan Dupuis at the Arbovirus Labs in Wadsworth Center, New York State Department of Health, for providing the protocols and the much valued assistance to allow us to get culture and propagation of West Nile virus underway at Chiron.
DR. NELSON: Thanks, Bruce.
Comments, questions? Jay.
DR. EPSTEIN: Thank you very much, Bruce, for the presentation.
One question here. You mentioned that you have the goal to develop the NAT assay as a supplemental NAT, but you linked it to support of the Procleix assay development. Of course, we are aware that you have a special collaboration with Gen-Probe, but I guess the question in the long term is do you seek potential approval of a supplemental NAT as a free-standing test, and if so, have you thought about the possibility of making your assay available to other candidate developers of NAT screens.
DR. PHELPS: At the present time, we have no plan to seek approval of this test for a more broad use. We are currently developing only to support the TMA assay under IND.
DR. NELSON: Mike.
DR. BUSCH: Bruce, all three presentations I think looked very good in terms of progress. What I am interested in is the data, now that we have it, molecular transcripts, and especially the comparative endpoint titration of the CDC plaque-forming unit standard versus these more quantitative viral load type assays.
Do we have insights into the relationship between plaque-forming units and copies? Some of the data I think Christina shows suggested that there may be about 1,000 plaque-forming units per genome if you just compare the 50 percent hit rates on the PFU dilution.
I am wondering, with your data, or your sense of what that relationship is, and then also whether perhaps others may comment on the infectious dose that might be evident in animal models, how many plaque-forming units are required to transmit if that is known in animal systems.
DR. PHELPS: Currently, Mike, we are doing the studies to determine exactly what the plaque-forming units versus copy number would be detect in the TaqMan assay. It is in the same ballpark as what Christina was showing for TMA, around 3- to 500 plaque-forming units, and we are continuing to refine that as we complete the studies.
As far as the infectivity level, we don't have any data at this point to respond to that. Perhaps someone else could answer that.
DR. NELSON: Other comments?
DR. SMALLWOOD: Would those speakers following the break please check with the audio-visual gentleman there to make sure that your presentations are loaded and they are the proper order.
DR. NELSON: I have got 25 after, so let's come back at maybe of 11:00. We will take a half-hour.
DR. NELSON: The next presentation is Dr. Andrew Conrad from NGI.
Andrew Conrad, Ph.D., National Genetics Institute
I am going to highlight a few points that have been made many, many times, but they are relevant because they become interesting and part of what we did here.
The most important thing here is that the West Nile virus is obviously an envelope virus. The incubation period is rather short and I will present some data that supports that concept.
The literature talks about the IgM to IgG conversion. When we encountered original patients, when the literature encountered original patients, 37 percent of the patients only IgM positive, but they converted to IgM-IgG within 4 to 5 days.
The IgM, and again this will be some data that is relevant for your later thoughts about the role of IgM in this testing, the IgM was stable with 50 percent of the patients having IgM positivity after several months.
At National Genetics, we have developed both a quantitative and qualitative test, and I won't go into massive detail. Instead, I will give you some data instead of showing you what the test looks like. I will give you a brief description, but then I am going to give you data from serologically positive individuals, and these are important.
These were individuals that came into our Clinical Divisions of the laboratory where doctors suspected them of having West Nile virus. They were febrile or demonstrated other sequelae of infection. Then, I will finish this talk with the results from preliminary screening, and I am going to use the data from preliminary screening of 40,000 plasma donors for West Nile virus. This will hopefully give you some ideas of the proportions and the magnitudes of these issues.
Again, in order to do this, we looked at all sorts of different regions of the virus, and West Nile virus fortunately is conducive for making tests, and I think from the previous speaks in the whole blood arena, that fortunately, West Nile virus lends itself to some testing.
But I thought the most important part about is in our validation studies, we found that we could get tremendous concentration of the virus through ultracentrifugation and in dilutive medium, which meant that we could increase sensitivity almost at will by increasing the sample size. That may be important because there has been a lot of talk about West Nile virus being in fairly low concentrations, and I think I will show you some data which supports that.
The next two slides we can skip through quickly. They are just what we used the same poor snowy owl liver that the whole world has used to generate strain 385-99, made serial dilutions, quantitated those serial dilutions, determined that the amount that we had in our culture supernatant was 1.2 x 109 copies/mL.
This sample was diluted in negative human plasma to create positive quantitative standards. Again, these are instead of transcripts, which we also have, we thought that we were going to use these because it better controlled for whole viral extraction.
This is just what the assay was looking like, and you can skip through these. These are the readouts that the computer sees from the Southern blots.
The next one, as well, you can go ahead. That's the quantitative sort of 40-cycle version. Go ahead, skip through.
We were then blessed with the hard work from the CDC about also having the panel, which was these 9 members. Unfortunately, when we tested all 9 members, they came up positive 100 percent of the time, all dilutions, so we had to make further dilutions down another 1,000-fold, I guess, and if you look at the next slide, we will see some numbers that we translated.
In the 100 microliter reactions that we used to do this, we had 95 percent detection in the smaller reagent reaction volume of 0.01392 PFU/reaction, which turns out in our 1 mL assay to mean that we have a 95 percent detection of 6.63 copies.
We all say these numbers, of course, tongue in cheek, and I wouldn't bet if someone said 6.63 and someone else's was 8 that there is really any difference. These are still based on relatively small numbers. We did triplicates, there different lots of the same Lanciotti dilution and then in replicates that you can see. So, any of these numbers are loose, but we are getting some idea.
Under our calculations, we ranged here from anywhere from 500 copies/PFU to 1,013 if you use those two spreads of those numbers of copies/PFU. I knew that that was a question previously asked, and that's somewhere in the infectious unit in the tissue culture models. Again, I think that may be dependent on the conditions of your tissue culture.
So, this in hand, we thought, okay, well, we have an assay that is fairly sensitive, let's move on to the clinical samples, which is the next slide.
We were given 140 clinical samples that came in to our clinical laboratory environments; 38 of those samples turned out not to be West Nile virus, and they developed no sequelae of the disease, they turned out to be hypochondriacs.
The other 102 were people who obviously weren't and what is interesting, and I think is most important here, is that of the 102 samples that we were able to detect IgM or IgG in, only 10 of them were viremic. This again supports the concept that it looks like there is a rapid--upon the appearance of IgM or IgG--there is a rapid clearance of the virus.
With the average viral load in these 10 samples being 45 copies, and it ranged from 20 to 100 copies, and I think that that is again important, that these were extremely low level viremic samples that 10 out of the 102 people who truly were West Nile virus candidates.
So, I think that that is again an important thing to keep in mind.
With this data in hand, we thought it was now time to move forward into a true asymptomatic population, at which point we began to investigate plasma donors from a very limited source that were participating in sending us samples in through our FDA-approved mechanisms, test mechanisms where we have those pools that I am sure many of you have heard of far too frequently.
Again, we used those standard algorithms. The magic cube that we have allows easy analysis of samples the size of 512 or 64, since both are simultaneously generated on the TECANS, and our licenses are for pools small than 512, so we figured we would take those shots at it. We could do pools of 512, 64, or 8, all without changing anything. We picked these two because we figured we would rather use higher volume test analyses than smaller pools. It makes the amount of work less, and that is what we specialize in, being lazy.
The first effort, we were able to locate 7,107 samples that were blinded, however, we did some geographic information based on how they came in. Of those 7,107 samples, 5,366 were from very high incidence areas.
So, when I begin, I don't want you to start making calculations of incidence and prevalence when you know that this was front weighted to cherry-pick into areas and the time period when there was likely to be higher infections.
The remainder of the samples, the 2,000 or so different samples came randomly throughout our catchment. National Genetics probably sees an awful lot of plasma, but this was a limited supplier who catches, happens to be fairly geographically dispersed.
In that first week, we were able to locate one West Nile virus NAT-positive sample out of the 7,107, and statisticians would be impressed by the power of this calculation, thus, determining 1 out of 7,107 is the incidence.
The titer of the individual sample was determined to be 2 x 105, which was interesting to us because this was higher than what people traditionally thought the viral loads for West Nile virus in humans would be.
We went on to fortunately grab that sample blinded. I can't identify the donor, but we could identify the sample, and found it to be negative for IgG and IgM, and that is important. That was the first true window period donation that we were able to locate.
So, thrilled with our performance, we decided to roll onward and look to see if we could expand that while rallying against people's hassles of having to do all this stuff. In the next slide, we were able to get 33,000 blinded samples from the same systems.
We screened 64, in pools of 64. These samples, however, were collected in late October 2002. In this case, the samples were distributed more widely throughout the U.S. and did have samples that did not represent the high prevalence areas, and the data follows.
Eleven additional samples tested West Nile virus positive from these 33,000, 3,300, for an incidence of 1 in 3,300 is extremely important to note that in the plasma industry, these could represent the same donor repeatedly donating in this catchment period. It is not like whole blood where the donor times would be spread out.
So, what I am talking here is donations, not donors, and it may be that these incidence numbers are dramatically skewed by the fact that a single donor donated three times, four times, whatever number of times in this, because they can donate as frequently as twice a week, and the average that we have gotten in our system typically is more than seven times in a three-month period.
So, 5 of these samples were detectable in pools of 512 or greater, so the other 6 were not, they were only detectable in the pools of 64. The viral loads of these samples were very low with none of the viral loads being above 13,000 copies/mL.
The range was anywhere from 250,000, which was the high sample that we spoke about earlier, down to 100 copies/mL, which were the range of these samples. The serologic status, I am getting today on some of them, but I will tell you that we haven't gotten IgG, that of the low samples, the ones that were below 1,000, all of them have been shown to be IgM positive, I do not know the IgG yet, so that is important to know.
In addition, quantitative tests have been shown to demonstrate the efficacy of inactivation methods against West Nile virus, and I think you will hear more about this afternoon. We are continuing to work with plasma suppliers to continue to perform these pilot studies, INDs, whatever you want to call them. We might go forward if appropriate.
It appears that West Nile viremic samples are pretty easily detectable in pools of 64 and definitely detectable sometimes in pools of 512, which surprised us because it would probably be efficacy in which pooling can be performed and the virus precipitated.
The final slide I have is that in our small study, and it looks like these things have come true, that IgG and IgM-positive, very low titer samples are one class of samples that we are going to get. So, in other words, the appearance of IgM and definitely IgG causes a precipitous drop in viral loads.
There is a second class of samples which are IgG-negative and IgM-negative, and those I am going to call mid-titer samples because they are certainly not as robust of infections as we have seen in some of the other viruses, parvo or HCV. So far we have not produced what we would call a very high titer viral sample, and we will continue to look.
The pace at which we continue to look, we are fully operational to look in any scale we need to because it is just added right onto our already FDA-approved systems. There is no modifications that need to be made other than the primer pairs.
So, as the industry deems necessary to get the inactivation data, we will look if we have to. That's all.
DR. NELSON: Thank you, Dr. Conrad.
Questions or comments? Jay.
DR. EPSTEIN: Thank you very much, Andy.
Can I get you just to clarify, since you used the concentration procedure, what would you say is the sensitivity for detection of the neat sample when placed in the configuration of a pool of 64? Is it 6 copies/mL or is 6 x 64, or something in between?
DR. CONRAD: I will tell you that remember that because of constraints in the amount of material we had in certain panels, that it gets confusing, and I apologize for this. Sometimes we were only afforded enough material to test 100 microliters, and therefore, we would have to extrapolate sensitivities to the mL amount by timesing that sensitivity or actually dividing it by 10 to get a lower sensitivity if you had a whole mL.
That is based on the efficiency of the precipitation, so as you know in our FDA licensed assays, we have licensed both the 100 microliter 1 mL, 2 mL, and you can see that the sensitivities ratchet downward depending on--we always report in sensitivity/mL, but right now the assay that we are representing here is a 1 mL assay, which we are claiming has 6.6 copies/mL in the neat sample.
If we did a 2 mL, it would be something, 8 or 9 times more sensitive, or 3 mL, so it depends on what becomes relevant in the examination, like all the clinical samples we used, we only had 100 microliters on in order to spare those samples. So, we were able to detect 12, and that is I think part of the problem.
We try to normalize against because we believe that in this case, with the proper diluent, that the ultracentrifugation really precipitates this virus better than it does with other ones, and we haven't seen floating that we saw with other flaviviruses like HCV because we took the top and the bottom pellets off, and we haven't seen that as long as you put the appropriate diluent in the plasma before spinning.
DR. GALLARDA: Jim Gallarda from Roche.
Andy, you describe your copy sensitivity in the terms of Poisson statistics. Does that give you an actual copy number of a detectable unit?
DR. CONRAD: Well, obviously, we do it two ways, Jim. I mean we like to use the final thing of the extractable unit, and you are right, because is it a true copy, but we also try to compare that against transcript, which we assume has 100 percent efficiency although there are some RT issues in there.
I didn't mean to imply that we just used Poisson. We have always gone two ways. We have always taken transcript against Poisson, trying to use the Poisson to control for extraction functionality and the transcript, so that is an amalgam when we get those numbers, and we assume the unity that you need for Poisson distribution is a derivative of having both the transcript data and that extraction data.
DR. LEW: I just wanted to clarify that there was absolutely no processing prior to detection of these pooled samples.
DR. CONRAD: I am sorry. "Processing," what do you mean?
DR. LEW: Processing, at this stage, it is just pooled, you know, sterilely for testing.
DR. CONRAD: It is pooled according to our standard FDA algorithms. We take the plasma, we combine it in an algorithm where an automatic robotic device combines the plasma into these pools of 64 or 512, is then lifted off those machines, extracted with standard methodologies that I think many of my predecessors have described in better detail than I, and then amplified in our case using PCR, and that is what happens to it. But by "processing," this is not ready for fractionation if you mean any processing of that kind, none of that, this is just raw plasma in a pool.
DR. NAKHASI: Hira Nakhasi from CBER-FDA.
Andy, I am just a little bit curious to know, you mentioned that the late summer samples of 2002, you had an incidence in the pool of 64, the problems was 1 in 7,000, whereas, you go to the late October 2002, again, at 64 dilution, you had 1 in 3,300. It seems to be much higher.
Summer will be much higher problems than later October.
DR. CONRAD: The problem is remember as I catch more, this is a relatively small supplier of plasma, so as I catch more time, remember I took one week to get those 7,107, and I got 1 guy. That is why I really want to make that clear, that the incidence of 33,000 could have been 1 guy going four, five, six times.
DR. NAKHASI: I see.
DR. CONRAD: So, the problem until we move into IND and unblind this is that I can only give you donational prevalence and incidence, but that in no way infers that there was an increase--there is an increase in the amount of the donations, but that doesn't mean that there is more donors who are succumbing to the virus, which would be a traditional way to look at an outbreak of an epidemic.
It may and most like has to do with the longer duration of this relatively small group who I am now following to garner 33,000 samples. You can see appreciably longer than I followed to take a snapshot of one week.
Does that make sense, because I don't want to engage in some paranoia here that it is going up in October, it is probably going down, especially because how low those viral titers were. These weren't all new incident infections. They looked much more--with the IgM presence--like an ongoing infection.
DR. NELSON: Thank you.
Next is John Callaghan from Tetracore.
John Callaghan, MS, Tetracore
MR. CALLAGHAN: Good morning.
As we saw at our last meeting, we have a real problem with West Nile, we need a test, we need it soon, and this should be in a NAT format and also we need an instrument, platform, and software that is FDA approved.
What we have from Tetracore, we have a NAT assay format, but it is not on a platform that is compatible yet. It could be made compatible, but we have some unique applications for our test that might be interesting to this audience.
What we have is a rapid vitrified real-time RT-PCR for West Nile. This target a 3-prime UTR region and what I mean by "vitrified" is we take the reagents and we dry them and use the test, so it could be stored in a dried format.
The potential uses for this test could be for supplemental or confirmatory test of positive or indeterminate samples identified by the NAT screening test, and another possible use would be for organ and tissue donor screening, and we have data that I am going to show you here today that supports our ability to identify West Nile in heart, lung, kidney, spleen, brain, and other tissues from infected birds, so this is not human data, but we are actively seeking possibilities of using this test on human organs and tissues.
What we were looking at in the design of this test was we wanted a broad specificity among West Nile and Kunjin viruses, and I will show you why we decided Kunjin was important in just a moment, but we wanted a test that had no cross-reactivity with other flaviviruses or clinically similar viral diseases such as the other encephalitides.
Also, we tested some human sera from parenterally transmissible diseases that the blood bank would routinely screen for, and I will show you that data in a bit here. Also, we wanted to have maximal sensitivity. We wanted to make the test quantitative, and we wanted it to be easy to use and have excellent test indices.
So, what we did in the original design is we took a look at the phylogeny of West Nile virus. You can see that it is broadly divided into two groups, and I have heard them referred them to as New World and Old World, and the isolates that we are seeing here in the U.S. are the New World group, and then the isolates from Uganda and Africa, other parts of Africa are classified into Old World group II.
Also, you will notice down here that all of this group here are Kunjin viruses, so if you want to have an assay that is going to be able to detect all of the members of both of those groups, we decided to pick a design with primers and probes that were 100 percent homologous or nearly 100 percent with every member of Gen Bank representing all of the groups here of Kunjin and West Nile, both groups I and II.
This just shows some of the data that we have got from one of the platforms was an ABI-7700 instrument showing the range of linearity across several logs. On this particular instrument, we are able to get down to between 2 and 20 copies/volume tested.
The other approaches we have been using, the Cepheid Smart cycler, which is a portable instrument that has 16 individual wells. These can be linked together and you can get up to six instruments running off of one computer at one time.
So, this is nowhere near the kind of volume that you are going to need to screen blood units, but if you were in a confirmatory situation, you may not have that many positives to analyze, and this particular format might be useful. These instruments are relatively inexpensive compared to some of the other instrumentation, so there would be an advantage to having a system like this.
What we have done with our reagents is we take the primers, the probes, the enzymes, all the perishable materials and we dry them down in the presence of trehalose. Trehalose is a glass-forming solute, so it gives you the appearance of a glass-like sliver kind of on the edge of a slide.
So, this is Cepheid Smart Cycler tube, so we have had a lot of experience in our company drying down RT-PCR assays for RNA viruses for a number of different agents using this particular format.
The way that you process the sample, we are using the Qiagen extraction kit, but basically, you would add your sample to the viral lysis buffer. Our input volume is 140 microliters. The virus is neutralized and the RNA is stabilized, and you go through the RNA extraction process.
For the test itself, we have the dried stabilized mixture containing all the components in the tube, and we have a universal rehydration buffer that we add. We bring the volume up to 25 microliter reaction, you know, put in 22.5 microliters of the rehydration buffer and 2.5 microliters of your template, then add it to the machine and test, and the results are out in about 35 minutes.
So, the test features are it is a single tube method. It is a reduced potential for contamination, a simplified handling of reagents because we have it in a dried form. In the vitrified form, it breaks the chain of refrigeration, so the handling of reagents is further simplified.
We have done shelf life studies on other assays showing that we get up to two years in accelerated shelf life studies, so, you know, leaving the test at room temp and then applying it to other accelerated temperatures.
These results can be transmitted electronically through the Internet, and there is a possibility, I mean even though we haven't approached this, but it is conceivable it could be linked up to a BIS system.
Of course, these types of tests generate much less waste than the standard PCR methods.
We would package it something like this. This is another assay that we have done for veterinary application, but basically, the tubes would be packaged with the desiccant and could be stored at room temperature.
These will be produced in a GMP facility. We are a company located here in Gaithersburg.
Now, this is a panel of specificity that we did on some of these agents. As you can see, we have tested a number of different flaviviruses, and there is no cross-reactivity. We tested several isolates from group I, the New York crow isolate.
We were working with the National Zoo here in D.C., and we got a number of different isolates from multiple species, multiple sample types, so we are able to detect at least several members of group I, and then from group II, this is representative of the BBI panel, the first one that they had available.
We also got an isolate from Romania that we tested, and then a Kunjin isolate from Australia, so these were all positive by our test. Then, we tested some human sera that we got from the University of Maryland, representing HIV-1 positives, HTLV-1, hepatitis C, hepatitis B and A, and then also at the last meeting, we had heard that some of the serological assays cross-reacted potentially with syphilis-positive samples, so we tested that, as well, and there was no cross-reactivity among those.
This is just showing some of the data on the ABI-7700. It has been shown that we can get both groups of West Nile and Kunjin.
Also, for specificity, we decided to take nucleic acid extracts from a number of cell lines that are commonly used for the propagation of viruses, and there was no cross-reactivity with our assay with any of these nucleic acid extracts. I would like to point out it is a very complex mixture of sequences, so it is good that there was no cross-reactivity.
Some of the veterinary samples that we had worked with, we were working with Cornell University, and we had a number of different animal species here and a number of different tissues. One real interesting finding was that we were able to work with cloacal swabs from birds and fecal samples, as well, and detect West Nile in these, but also notably, we had heart, brain, kidney, spleen, lung, liver, and CNS from a horse, and we found that with the assay that Cornell uses, they use a nested RT-PCR, and we had 100 percent correlation between our real-time PCR and theirs on the sample panel tested.
Also, these are samples that we got from the National Zoo here in D.C., and of all these different species, we identified West Nile in 12 different species.
Also, multiple samples, so the plasma that we used from animals, there were represented multiple species, and then also the tissues, there were multiple tissue types. Often we would get like a homogenate, a blend of tissues to test, but we made a lot of positive identifications throughout this past summer with the veterinary samples.
Now, the technical challenge was the quantitation. What we decided to do was take a look at the Armored RNA. Now, you can get this from Ambion in Austin, Texas, but the problem is that you would need an NS5 assay to amplify this and ours is a 3-prime UTR, so we needed a bridge between the two assays to quantitate the positives that we were using, so we developed a second assay that targets the NS5 region.
Then, on that, we tested the Armored RNA, which is a known quantitative standard that you can add. It goes through the whole process of extraction, so it's a known standard that you can put into your test system. You know what goes in, you know what comes out.
Then, along with that, you could compare like, for example, the sample from an infected sparrow, you could compare it against your standard curve and quantitate with that, then, run endpoint dilution in parallel on the 3-prime UTR.
So, we ran those in parallel on the Smart Cycler.
Then, also on the ABI-7700 to compare.
That quantitation showed that the sparrow sample we had came out at 2250 copies/volume tested, was 2.5 microliters.
So, if we run an endpoint dilution, this is just data from the Smart Cycler showing down at log 4. That would be equivalent to about 2.25 copies/volume tested. So, if you converted that back, it would be in the range of 7.5 to 75 copies/mL. We still need a lot more data to show the reliability of these figures, but at least that is a good start, I think.
The specificity so far has been 100 percent among all the panels that we have tested. There has been no cross-reactivity with other members of infectious disease panels that we have selected. There have been no false positive in samples collected where the disease is absent, and the sensitivity is between 2 and 20 copies/volume tested.
We have just recently received the panel from the Red Cross just a few days ago to start testing human plasma, but since we just got it, there just wasn't enough time to finish the work before today's presentation, but that work is ongoing right now. Also, we would like to identify partners to collaborate with for testing human tissues and organs for West Nile virus.
Some of the advantages that we have right now is excellent specificity, a broad coverage that detects both West Nile groups I and II and Kunjin viruses, but doesn't amplify other flaviviral near neighbors, and it so far looks sensitive, that we could get down to less than 20 copies/volume tested, and we can make this qualitative or quantitative. Then, we could supply these reagents in wet or dry form, so that if there were other instrument platforms, they could be used on there.
We have detected virus in multiple tissues and fecal material and plasma from infected birds, and it is very fast turnaround time. On this particular instrument, if you just had a handful of samples, you could do the extraction and testing, and have results in under two hours.
So, basically, that's it from Tetracore.
DR. NELSON: Thank you.
Questions? Yes, Judy.
DR. LEW: Actually, I have some concerns about the sensitivity and maybe you can explain that to me. You are talking about 20 copies/volume tested, but the standard test only requires 2.5 microliters, and if we know that humans tend to have low viral load, you try to do that extrapolations, because with the birds, they have very high viral load, so it would be much easier to detect.
How sensitive is it going to be if a person who has just in the thousands copies/mL or maybe 10,000 copies/mL?
MR. CALLAGHAN: Well, I think we would have a much better feel for that as soon as we are done with the panel that we have received, because those are plasma from human donors that some of them are infected.
So, we have just begun that testing, so we are going to have to do quantitation on those, as well, to get a better feel, and also do many more replicates of the extracted RNA to get more confidence in our real level.
The work is ongoing, so we are not completely finished with that yet, but this is just some preliminary data showing that we have got down to those levels on those types of samples.
DR. NELSON: Thank you.
The next speaker is George Dawson from Abbott.
We are just now moving into serological tests from NAT testing.
George Dawson, Ph.D., Abbott Laboratories
DR. DAWSON: I am going to update you today on some of the work we have done on West Nile virus IgM testing.
We talked last time in early November about some of what our goals were, and it was really to try to understand in a little more depth what is the marker profile of an infected person, what is the overlap between RNA detection, antigen detection, and IgM detection, what are the markers associated with symptomatic acute infection.
So, we have been working on developing an IgM class antibody test for detection of antibodies to West Nile virus. We have obtained two different recombinant proteins from two different laboratories.
We took a lot of time to get the Material Transfer Agreements signed, but we have been working in earnest on that. One of them was signed in early November, the other not completed until early January, so we don't have a ton of data to present today, but I am focusing mainly on data today.
So, we evaluated these two proteins for ability to detect IgM. We also wanted to have capability to do PCR in our lab, so we looked at different primer pairs and their ability to detect RNA in a panel of samples provided by BBI.
Our major goal is to look at the utility of IgM testing in three different settings. One is in blood screening, one is in reinstatement of RNA positive donors, and a third is in diagnosis of acute infection. We are relatively sure that the best test around for acute infection will be IgM.
Here is a depiction of the viral genome. I will call your attention to the different sets of primers that we have developed and looked at their ability to detect RNA. We looked at different sets of primers and we selected, in bolded red there, degenerate primers that were built specifically for West Nile virus detection.
We have been using RT-PCR test and using nested PCR, running agarose gels to detect the amplicons.
I mentioned we obtained two different recombinant proteins. Protein 1 was a truncated envelope protein expressed in eukaryotic cells. The second protein was a pre-membrane/membrane/envelope gene expressed in eukaryotic cells and actually secreted into the supernatant fluid as vesicles believed to be secreted as viral-like particles and to allow the envelope protein to be folded in its correct conformation that would more simulate the native virus.
This protein 2 came from CDC. We actually received from CDC the protein itself, as well as the cell line that produces this protein.
We looked at two different assay formats with these two proteins. With protein 1, we use an indirect assay format. This assay entails coating a solid phase directly with antigen, reacting the antigen-coated solid phase with sample. IgM, if present, would bind to the antigen-coated solid phase, and would be detected in the second step with a conjugate directed against IgM or against IgG because we also wanted to look at, at the same time we are looking at IgM detection, we wanted to look at IgG detection.
For protein 2, we used an IgM capture EIA. This is a three-step assay, uses a solid phase coated with antibodies to heavy chain of human IgM, so in the first step, IgM is specifically captured on the solid phase. After a washing step, antigen is added.
After an incubation step, the antigen antibody complexes will be bound to the solid phase, unreacted antigen would be washed away, and we then subject the complexed solid phase with a conjugate specifically directed against West Nile virus envelope protein.
We have done more testing than I am showing today, but here is what I will focus on for today's presentation. We tested volunteer donors from areas that were not endemic at the time that the blood donations were taken.
We looked at some panels that were categorized as IgM only or IgM and IgG only by a CDC assay run at a state health laboratory. We also obtained paired human samples from acute West Nile virus infection.
We received the CDC proficiency panel and we are in progress with an experimental infection in Rhesus monkeys, and this study is one that I proposed last time that we would get involved in. We have made good on that, and we are in the midst of doing this testing. I will show you just a little bit of data from that study as I conclude this presentation.
With protein 1 for both IgM and IgG, we found a lot of bounce in the population, and we found reactives IgG and IgM with it, with this assay format, indirect assay with protein 1. These were RNA-negative. The IgM-positive was IgG-negative, the IgG-positive was IgM-negative, so there was no co-detection of antibodies.
With the second protein, which we felt is more specific, these samples were negative. In any event, we had to cut a cutoff somewhere and what we did is we just took a bottom line of looking at about 7 standard deviations from the population mean, set the cutoff there, and then proceeded with our studies.
With protein 2, with the three-step IgM capture assay, we tested 241 volunteer donors from a non-endemic area, and found that none of the 241 samples were repeatedly reactive. We set a cutoff which nicely kept the samples that we assumed to be negative as running negative.
We then began to look at some of the samples that came to us, categorized as being IgM-positive and IgG-positive, and we found that if you looked at this, a dotted line running across the graph here, and we have sample number on the x axis, S/N values on the y axis, and you can see that all 8 are positive with protein 2, only 5 of the 8 are positive with protein 1, so we are beginning to have some concerns about protein 1's ability to detect IgM.
Then, we looked at samples that were IgM-only, and here there are 9 specimens that are IgM-only. Protein 2 detected all of these. Again, you see the dotted line running across is where the cutoff would be set. All 9 positive with protein 2, and only 3 specimens were positive with protein 1. So, we are seeing some very clear differences between these two proteins and ability to detect IgM.
A second set of pedigreed samples we got were 12 pairs of specimens that were diagnosed as IgM-positive for West Nile virus, and we see here the 12 patients aligned on the x axis, S/N values on the y axis, comparing this time it is not protein 1 versus protein 2, it is protein 1 alone here, and the number at the top of the bar indicates the days post onset that this sample was tested.
You can see that a couple of them are very nicely positive, patient 2 and patient 9, patient 7, the first bleed positive, et cetera, but a lot of these were negative, again another concern that protein 1 may be missing true acute West Nile virus-infected individuals.
With protein 2, however, we found that all 12 of the first bleed specimens were positive for IgM, and 11 of the 12 of the second bleed specimens were positive for IgM, so again we are finding that the second protein that we are evaluating, the CDC protein works very nicely in detecting the acute phase infection.
You will note that in some cases, patient 1, the IgM value goes up between day 11 and 21, other ones, as in patient 2, remains the same. Patient 3, the IgM values goes down, so we see a lot of different types of responses. For those experienced in serologic tests, this is what you see. There is not a monolithic type of response. You will see different things with different individuals.
Both assays here are using protein 2. We received a 20-member panel from CDC, and we tested these. We have the panel member on the x axis, on the axis, the S/N value obtained.
We have, in the lighter blue, the CDC values that we received from Fort Collins. On the dark blue is the Abbott IgM S/N values, the point being that we detect correctly all the samples that were sent, and, in fact, our sample to negative values are substantially higher in our hands. It may be because we have a lot of experience developing solid phase tests, and we are actually in this assay that I am talking about today, we are using actually an IgM-coated solid phase that we manufacture for other IgM tests, so we have taken all the bugs out of how to coat that solid phase, how to keep the background very low, and how to make the solid phase consistent and reproducible and able to detect IgM.
I mentioned that we have an ongoing study. This is just a schematic, so I can address what we found, in general, is that we are seeing RNA anywhere between 5 and 10 days, from day 2 through day 5 in some animals RNA positive, and other animals, it goes out as late as day 10 post-inoculation, but there were no RNA positives beyond day 10.
IgM in the five different animals we look at came up in every individual, ranging between day 9 and day 11 post-inoculation. IgG follows about 4 or 5 days later. It seems to be coming up around day 14.
So, we are seeing some overlap of RNA detection and IgM detection. One of the animals we clearly saw IgM-positive, RNA-positive bleeds.
Just go back there for one more second.
We are working with collaborators at another site, and we are doing the IgM and IgG testing. We are going to have hemagglutination inhibition data, plaque reduction neutralization tests. We have RT-PCR in my lab, and we also have infectivity being done. So, we will have a pretty nice profile of available tests on this experimental model.
It doesn't depict it on the x axis, but actually, these animals were bled every day for the first 14 days post-inoculation.
So, we have looked at two different recombinant proteins and we have developed IgM tests. Protein 2 seemed to do a much better job in detecting antibodies of IgM class than protein 1.
I don't mention here, there is just not enough time, but we did do also some IgG evaluation. Protein 1 works very nicely for IgG, and protein 2 also works for IgG, so we don't see a big difference in performance with protein 1 versus 2 for IgG detection.
So, we are continuing to look at viremia and the relationship with IgM detection. We have a cell line in-house that we have obtained from the CDC, that we are generating internally. We are producing antigen. This antigen is showing to work similarly to what has been working with the CDC antigen.
We are looking at prototype assays first on polystyrene beads, has probably a more rapid assay development cycle, and also microparticle-based assays.
We are also expressing additional proteins internally from envelope, I mean 1 in 3 of envelope protein. We are looking at NS3. We will be looking at core.
We have studies planned or in progress. Here, we are continuing to optimize the assay. We are looking, how are you going to confirm a positive, and this something you can spend quite a bit of time talking about because probably the best confirmation is plaque reduction neutralization tests, but it is very tedious, time consuming, and it requires very specialized training and equipment.
RT-PCR will work probably on early acute phase samples, but it is probably not going to help you out on most of the clinical samples coming in, in an asymptomatic, in a diagnostic lab.
Another confirmatory strategy is to look at IgG detection with protein 2 or to look at both IgM and IgG on alternate proteins. I mentioned NS3, core, and other envelope proteins we are looking at and expressed in different host cells.
External studies, we are not really embarking on anything right now, but we are looking forward to participate in the study on quarantined units to find out what we see with IgM and RNA and IgG, and possibly to participate in the donor reinstatement studies, and we will be looking to align our tests with the testing done at different state health labs.
Thank you for your attention.
DR. NELSON: Thank you.
Questions or comments?
DR. LAAL: How many animals do you have in your macaque study?
DR. DAWSON: There is five animals.
DR. LAAL: And they are all infected with the same dose?
DR. DAWSON: They were infected with the New York strain, intradermal inoculation of 105 PFUs/mL.
DR. LAAL: And in all of them, you see this fall in titers?
DR. DAWSON: That is a schematic, it doesn't represent any reality. Some did fall, some did not fall, so it is hard to generalize.
DR. STRONG: Are you using a blocking reagent for IgM assays? I would be a little concerned about specificity.
DR. DAWSON: With protein 1, in an indirect assay, we used different, typical types of blockers, animal sera, detergents, and, as well, we are using goat antihuman IgG or whatever it is, antihuman IgG, so we can make sure that we are detecting the IgM and it is not being masked by IgG.
In the second assay format, the IgM capture, we used some of the diluents that we use for hepatitis A and hepatitis B, which have different types of protein blockers to optimize specificity, so I think we are in pretty good shape for that.
DR. STRONG: And you are using sort of the classical format. Have you looked at chemiluminescence, as well?
DR. DAWSON: No, we have not, but we will. Like I said, we didn't get materials that were going to be of any use to us until mid-January. Some of these assays require using monoclonal antibodies, especially that IgM capture, you are going to have monoclonal antibodies that you purify the antibodies and label with your chemiluminescent tag or whatever enzyme you are going to have, so we are working along that path.
DR. EPSTEIN: Thank you, George.
You mentioned interest in the potential development of IgM as a donor screen. I am just wondering whether you contemplate having any products available for IND use in the 2003 summer months.
DR. DAWSON: I think that is a difficult question for me to answer because, you know, it would require a pretty large volume of tests, and we would really need to meet with CBER to talk over how we would implement that plan. We think for this year, we could probably participate in some external studies to look at is there value here.
What I am seeing so far is that I would love to get the 20 or 21 samples that were linked to transmission to find out how many of these are IgM-positive. What we have heard, what I have heard at other meetings, with IgM testing, is that most or all of these samples were negative for IgM.
So, we are here if, you know, the community believes there is a need for an IgM test for screening, but I think current consensus--and please correct me if I am mistaken--is that NAT will be done. IgM could be there for donor reinstatement, or if there was a turn in thinking that IgM is required, then, we need to discuss that.
DR. NELSON: it was my understanding that the transition early, in the first few days, that IgM appears, that there still may be PCR-positive infectious virus, so that theoretically, you might pick up low-level viremia using an IgM.
I just wondered when IgG appears, in other words, when IgG appears, does that mean that the unit would not be infectious or that the person would not be infectious at this point. Is there any data on that?
DR. DAWSON: I don't have data on that. I would like to ask a couple things of everybody here. When you find a NAT-positive, when anybody finds a NAT-positive, I think they should be withheld in some manner, made available for IgM and IgG testing, and for virus isolation, because there are labs that do this routinely.
Let's get to the bottom of this thing and find out. You know, there has been all these positives identified in those 33,000 samples, let's have some cooperation and see, you know, maybe make those available to people that test for IgM, IgG, to experts in the field, send them under code, and won't have to second guess.
DR. NELSON: Thank you.
Charles Tackney from Ortho. He is substitution for Steven Alexander.
Charles Tackney, Ph.D., Ortho Diagnostics
DR. TACKNEY: Thank you. Steven is off on another jaunt.
I am going to show you today some early data on a prototype assay that we put together a few months ago after the November meeting to try to see what sort of tools we could mobilize to answer some of these questions that were coming up.
If you noticed, I use the word "user-friendly" here, and I hope when we get done, maybe I will leave you with the thought that the assay is rather facile to do, it has certain benefits of performance, and even though it is a very complex subject, potentially, it's even deployable.
I would like to just reiterate what George said because we seem to be bringing up the rear on immunoassays. It is a very complicated topic, and we do have a research tool here, it is a very good tool, but it is a very complicated medical question that we are trying to answer.
I think what we want to do is say that we have put together some things, we have some observations and some new tools of reagents that might help answer some of the critical epidemiologic and immunologic questions about this virus which are really not clear, I think, today.
It is not your typical flavivirus using the HCV paradigm. So, take this as a contribution towards the knowledge really and be aware that we are perfectly amenable to discussing any scenario around these tests because we really don't know where to go with it.
The test uses a 10 microliter sample, it's a microplate format, and lab people prefer to call this right now a hand-cranked assay. It does not take advantage of any of the process and the excellence that we can put into a full-blown assay, but I think you will see there is enough data to justify that it is working.
It is a homogeneous reaction in solution that involves a biotinylated anti-mu capture reagent that is mounted on a SAC well, either a microplate or a cup. It has a less than 2-hour time to result. That time can be pushed down below an hour when it is brought onto one of our automated platforms, and it uses the chemiluminescence readout typical of the ECI platform.
Very simply, the patient sample is diluted down in liquid to the anti-mu capture reagent. The reaction there takes place almost instantaneously. It is incubated in a SACwell for 30 minutes in this format on the SACwell, it is washed. A conjugate is added, and the term "conjugate" here we are going to put in quotations. I will show you what that means, a "conjugate" is added, substrate and chemiluminescent readout, and time to result 30 minutes.
In the typical predicate assay that this comes from, which is our HAV or HB core IgM assay, these incubation steps are 15 minutes each, so that is why I say a hand-cranked assay.
What you have is a conjugate which, in this case, you should keep in mind is either a monoclonal antibody to a ligand, and that ligand can be a recombinant antigen, a peptide, or a viral lysate, or it can be a directly labeled analyte, an HRP-labeled material.
In the assay I am going to show you, we are going to use an envelope recombinant, and we have a very substantial effort here in cloning. We have basically cloned all of the relevant proteins, the structural and some of the non-structural proteins, in both bacteria and, where applicable, particularly for the envelope protein in mammalian systems, and we have a monoclonal in this case to that envelope protein. That is what I will be showing you here.
The biotinylated anti-mu capture monoclonal in solution captures the human IgM, captures down onto the SACwell in a rather seamless manner, and the assay basically is pre-formatted, so there is no coating, if you will, here.
We have not invested a lot--and we will--we have not invested a lot of energy, and this was put together for the purpose of this meeting, in getting the kind of numbers of patient samples and things that would be statistically relevant. We have very, very few actual West Nile patient samples available to us, and that is a big problem, and we have only looked at a small cluster of normal patients that we just pulled out of the archive.
So, the concept of cutoff here, I want you to give me the benefit of the doubt that I am not declaring a cutoff. This is a proof of principle assay, does the reagent format, as it is put together, detect things that we can monitor against the CDC format assays right now, and then we would move on from there.
The assay would be generally very, very quiet in terms of light unit output had it not been for one outlier, so I have not tried to manipulate the experiment to choose my best. Out of 41 samples, this is what the spread of ALU units looks like when we read it.
I am going to leave that number in there and declare a 225 ALU average approximately, just give myself 2 standard deviations, which would not be typical of one of our assays, and say that the cutoff is going to be 400 ALUs.
If you do that, you will see that there is a nice bundling down here towards the left except for this one outlier, but I am going to stay with the 400 ALU for now.
These are samples that were provided to us very generously by Dr. Meyers at the Maryland Health Lab, and I have to say--I don't think he is here--he has been donating a lot of counsel time to us, helping us to understand the medicine and get reagents and recruit reagents, which is not trivial, and echoing what George said, it is really difficult for us to get things. I know everybody is trying their best, but we are not getting everything that we are really seeking in a really timely fashion, especially these new challenge samples that we believe are coming through.
If you look at Bob's data here, they have all been PRNT-positive confirmed. They were sent to us in the blind with the following commentary as per his interpretation of their lab data. This one was told to us to be a West Nile IG equivocable with a Mac ELISA of 3.7, which would put it right on the cusp of positive at the Mac ELISA.
This is an IG-positive who is negative in the Mac ELISA and then a series of IgMs including one that scored negative in that ELISA, the Mac format, but is PRNT-positive.
This is one generation of our assay using that 400 cutoff and clearly, we are going to call this one positive right here, assuming it's above 1, and most likely it is a positive.
This one statistically would become a negative sample very easily if I were to spend the time formatting this out properly, and all the say down here, you see we have a very, very nice signal range of these positives including this fellow here, who is clearly positive.
We have done this several times, and we have had these samples assayed several times here at the Health Lab, and we are seeking more samples like that especially those infectious window samples if we could get them.
This is just a very, very minor piece of data just showing that the assay is highly dilutable, it has got a robust dilutional capability for what that's worth. It's alive, if you will, and it would help to get some quantification of these results, which is not really trivial with the current assay.
We identified a patient down in Texas, we just call it the Texas patient, that was scored positive in the CDC assay. It had hemagglutination titer, specific titer over here. We put it through our assay, and we drove it down. We stopped at that endpoint unfortunately. You could see here again we have got a nice dynamic range picking up that patient very clearly.
We do have in our archives a whole array of exotic all sorts of agents. We popped out a bunch of dengues for which we don't have a lot of background pedigree information, but we do know their basic serotypes. You could see here these are really quite hot samples and may form the basis for some sort of a calibrated panel for us, and also reinforcing the fact that I shouldn't be calling this a West Nile assay, this is a flavi assay, and that could be a good thing or a bad thing depending on where we want to go with this assay as a clinical tool.
What are we going to do in the future? Well, the answer is really we are not sure. We want to participate, we want to help. We really enjoy the subject. The staff is pretty excited about it. You know, you get tired of the old hepatitides. We certainly plan to transition this assay onto the vitreous format if possible, pick up some of the advantages of that platform, make it more portable.
We need to optimize assay performance. We have no doubt that the small amount of time we have spent we can quiet that down to make it typical of some of our other assays where these ALUs get down below 100 reproducibly.
We can get the validation down around the reps, the precision, and so on. We want to do lots of studies with experimental samples, challenge samples, interference samples. We are doing that now. We are seeking those from other people, and we really want to be involved in someone's trial, looking at these interesting retrospective studies and seeing what it means today given that data that we could accumulate with this assay.
So, we are here to help, but we really don't have a decision on where we will be next summer, as I think George eloquently put it, I think all the organizations want to do something.
DR. NELSON: Thank you.
Any questions? Yes.
DR. FITZPATRICK: On the IgG-negative, IgM-positive sample, do you have nucleic acid test results on that?
DR. TACKNEY: No, we do not, and we have requested that. We do not have that in our hands although it may very well exist in the laboratory, and we have requested that. We are also waiting on the CDC proficiency panel, the panel that has been described, as well as we are also out in the field surveying our contacts particularly in the American south, doctors who send us patients, send us interesting things that we can look at, and talking to several principals here currently about what opportunities might exist to get that information.
DR. NELSON: Jay.
DR. EPSTEIN: Do you care to comment on the recombinant and antigen, on its origin?
DR. TACKNEY: Sure. This is a mammalian-based recombinant antigen. It is similar in performance to the mammalian antigens that have been discussed earlier at other meetings. I will tell you that our feeling is that it is a confirmational structure. Working out a bacteria is not suitable for this antigen, we can assure you of that.
We have a cell culture system which is a suspension system, a mammalian-based system similar to a myeloma, if you will, that is putting out this antigen. We also have the very, very earliest--and I won't be presenting any information on it--hints that we have a fledgling antigen assay, have no idea what that means in terms of the progression of the disease, it is purely Ph.D.-speak.
We have NS1 cloned, NS3 cloned, capsid cloned. We have an array of peptides because we had sort of a naive dream that we would just pop out a West Nile-specific assay right out of the gate. That's not dead in the water yet, but is proving very, very difficult, and actually, maybe we don't want that if you are going to screen.
The antigen is critical, as George said, not all antigens are made alike regardless of their sequence.
DR. NELSON: Thanks.
Next, is Wallace Narajowski from Focus Technologies.
Wallace Narajowski, Focus Technologies
MR. NARAJOWSKI: Good morning. I am Wally Narajowski with Focus Technologies. For those of you that are not familiar with Focus Technologies, we are a small company, a reference laboratory, and a diagnostic product company. We have been dedicated to infectious diseases for over 25 years, and our headquarters are in Herndon and our laboratory is in Cyprus, California.
Some of you may know us formerly as MRL, Microbiology Reference Lab. We are now called Focus.
I am here to give you an update on our West Nile IgG and IgM ELISA product. I have to thank Dr. Hogerfe and Dr. Prince, who are really the technical people behind this. They are back at the shop making sure we can get product out the door actually starting within a couple weeks.
My purpose here today is to give you an update on our products for West Nile virus. Currently, our goal is to work with the FDA to have an FDA-cleared clinical diagnostic product available this season.
Starting next week, we will actually be supplying the West Nile antigen for the Public Health System. We are manufacturing that for them and we will be starting to ship that next week. We are targeted to have an ASR for both IgG and IgM available starting in April.
The key reason we can do this is last year, in our reference laboratory, we performed over 36,000 serologic assays for West Nile. As you can see, this chart, this is the distribution of when these were actually run, and our goal is to have the FDA-cleared product with the FDA by August.
Currently, we are naming this product Flavivirus West Nile, and the reason we are doing this is because we got the recombinant from the CDC, and this recombinant is targeted for West Nile, but it also does cross-react with other flaviviruses.
Our IgM assay is about a 2 1/2 to 3 hour microtiter standard ELISA format for a capture type, 96-well, microtiter format, and our IgG is about an hour and a half assay, very similar type of format that I think everybody in here is familiar with.
The actual protein that we received from the CDC, and that is going to be part of this, is the recombinant protein, and it is derived from the pre-membrane and envelope genes of the West Nile. CDC last year supplied this recombinant to all the Public Health System. This year, we are going to supply it to both CDC and Public Health.
As I mentioned, the protein retains both specific epitopes and cross-reacts with other flaviviruses.
We utilized samples from our 2002 season to establish a cutoff, and this is really the basis for establishing our cutoff for this assay. We started obviously with the 36,000 serologic specimens.
We took 700 of those and used that as the basis for the cutoff, 150 of the 700 had Public Health results, and 100 of the 150 had plaque reduction results associated with them, as well, and that is how we established our cutoff. What I am going to show you next are the results of the assay that we are going to release in April utilizing this cutoff.
This is for our IgG. We first did a Southern California where we did normals, and we did a mini-seroprevalence, and it was very similar to what we found with dengue many years ago, 95 percent negative, and there was an underlying IgG positivity rate.
Specificity compared to our reference lab was 95 percent. The four outliers were in the equivocal range. Sensitivity was 97 percent and again 3 of the 4 in the equivocal range.
Our IgM assay, we have 100 percent concordance for negative sera versus our reference lab, and 96 percent concordance on positive, 24 of the 24 Public Health IgM, plaque reduction were positive with our IgM assay, and 118 of the samples from Public Health, either ELISA or plaque reduction were 100 percent concordance, as well.
In summary, our goal is to introduce an ASR format with the products that you see here, probably starting the end of April, and we hope to have an FDA-cleared product in time for August. We did meet with the FDA actually Tuesday of this week, and they are onboard, if we give them good clinical results, they say they can make it happen.
DR. NELSON: Thank you.
DR. GALLARDA: Jim Gallarda from Roche.
Are you providing the antigen only or are you providing all reagents for the test methodology?
MR. NARAJOWSKI: We are making available both to Public Health, the antigen, as well as the ASR reagents, but most of them are going to probably perform their own assay like their protocol last year, but they do have the option to purchase the reagents from us, as well.
DR. GALLARDA: I guess my question is related to the quality control from a manufacturer providing uniformity in manufacturability as compared to distribution of the components to multiple labs, how that might affect the reliability of the assay in the setting.
MR. NARAJOWSKI: The way that it is configured is there is a reagent pack, as we call it, a convenient pack that has the plate and the conjugate and the cutoff as part of that convenience pack. Then, all the other reagents, they can purchase from our catalog, and that should take care of any of the uniformity issues.
Does that answer your question? Okay. Any others? Okay. Thank you.
DR. NELSON: Thank you.
Next, is Dr. Lyle Petersen, who is going to give us an update on the transfusion-transmitted cases.
C. CDC Update on Investigations of West Nile
Virus Transfusion-Transmitted Cases
Lyle Petersen, M.D., CDC
DR. PETERSEN: Good morning.
What I am going to present is work from a number of different topics.
What I will first do is talk about the transfusion transmission investigations, first talk about the characteristics of the recipients, characteristics of the donors, and talk about comparison of symptoms of implicated and non-implicated donors.
Then, I will go on to discuss the estimates of risk of transfusion transmission during the 2002 season, talk a little bit about a viremia study which I think Sue Stramer will follow up on that after me, and then talk about some data we have from our experiences from IgM antibody testing in our laboratory.
To go on to talk about the transfusion transmission cases, we wanted to determine the number and type of components transfused to recipients in the four weeks before illness onset. Basically, what happened was is that through our national surveillance system for West Nile virus, possible transfusion-transmitted cases became known, and these were people who had received transfusions in the four weeks before developing West Nile virus infection.
So, we tried to retrieve what we called initial donation samples, which are retention segments and then transfused components that were there from the donor at the time of donation. Then, for all of the donors, we then followed up them with a follow-up questionnaire and got a serum sample for IgM antibody testing.
These were all from donors of components received by the index recipient, and we also went back to try and find co-component recipients of donors who we identified with transfusion transmission related West Nile virus infection.
So, the criteria we used to define a transfusion-transmitted case was actually quite strict, and we required either a donor to be PCR-positive and/or culture-positive from an initial donation sample, and IgM-positive at follow-up, or to be PCR-positive and/or culture-positive and have a co-component recipient that was IgM-positive.
The third category we considered as a definite transfusion-transmitted case was a PCR equivocal donor who was IgM-positive at follow-up and also had a co-component recipient that was IgM-positive.
The last category of a person we considered was a co-component recipient of a person proven to be, by the above case, by the above criteria, and that co-component recipient had to be IgM-positive, as well.
So, we received reports of 61 possible transfusion cases as of March 1st, 2003. We determined that 19 were definitely not transfusion related, 8 didn't have West Nile virus infections, 3 had their transfusions after symptom onset, 2 had symptom onsets more than 28 days before the transfusion, and 6 had West Nile virus infection, but all donors were IgM-negative.
Now, that left 21 additional cases that we called inconclusive due to incomplete donor follow-up, and we have pretty good evidence that many of these 21 cases could have been transfusion related, but we just didn't have the data to categorize them according to the scheme that I mentioned on the last slide. That left 21 confirmed cases of transfusion transmission of West Nile virus, and this will be our final tally, I think, for the last season.
In looking at the recipients, this slide depicts the underlying medical conditions or reasons for hospitalization, and you can see that these were a pretty sick bunch of people, to begin with, 6 had stem cell or organ transplants. Nobody got the infection from the organ transplants in this series, but these were post-organ transplant patients who were on immunosuppressive drugs.
Four had cancers, mainly hematological malignancies, 8 had obstetrical or surgical procedures requiring transfusion, and 3 had other medical problems.
Of these confirmed cases, 12 of them had West Nile virus meningoencephalitis, and 1 had West Nile fever, and the rest were asymptomatic.
Of the 13 symptomatic people, one of the curious things was the incubation periods among these people were rather long, and they had incubation periods from 2 to 21 days, the median 11 after the implicated transfusion, and this was an unexpected finding.
These people received blood from a median of 36 donations ranging from 2 to 274, and of the outcome, 6, or 30 percent, died, 1 due to West Nile virus meningoencephalitis. Among these people, it was very difficult to really determine if their underlying condition killed them or the West Nile virus killed them, or both, to this outcome data is a little bit unclear.
So, all three types of the major blood components were all implicated, red cells, platelets, and fresh frozen plasma, and the maximum interval between the date of donation and the date of transfusion was 33 days for red cells, 5 days for platelets, and 44 days for fresh frozen plasma.
In looking at the donors of these 21 people, we had a total of 726 donors returned for follow-up testing, and 92 percent completed an interview.
Out of the 21 people, there were 14 donors implicated. A couple of the donors had transmitted to more than one recipient, so that is why there aren't 21 donors, there is only 14. All had evidence of viremia and importantly, all were IgM-negative on their initial donation samples, and all were linked to at least one recipient with West Nile virus infection.
Seven of the 14 donors were linked to more than one recipient with West Nile virus infection, and actually, one recipient had received transfusions from 2 viremic donors, and it was unclear which of the viremic donors had transmitted to the recipients.
So, this slide summarizes the 14 donor labeled A through N. You can see that their dates of donation corresponded to the peak periods of the epidemic last year, in late August and September mostly.
One thing that is very interesting is the symptom onset and to the time of transfusion. The ones with no data are the asymptomatic donors who had not symptoms at all, and there were 5 of those, and the minus figures are people who had symptoms before their donation, and plus are people who had symptoms after their donation.
In some cases, there is a range where the donor couldn't really remember.
Now, I would like to express some extreme caution in trying to interpret these data because you have to remember that these donors were asked to recall a very mild illness that had occurred months before, so the timing of their illness onset relative to their donation was a bit unclear, and a couple of them, like donor A, had some kind of sinus infection before their West Nile virus infection, and it was really impossible to exactly determine when her symptoms began.
The last two columns are the PCR test results, both for retention segments and for plasma. The retention segments turned out to be very difficult to test and somewhat unreliable. We both had false positives and false negatives associated with testing the segments, but the plasma samples were held under good conditions, and these are the results in plaque-forming units/mL. As you can see, all of the viral levels were quite low.
So, the next thing we wanted to look at was symptomatic versus asymptomatic donors. What we found is the symptomatic versus the asymptomatic donors had a median of 44 and 45 years respectively. The mean viremia levels in plasma were higher for the symptomatic donors compared to the asymptomatic donors.
The viremia levels seemed to decrease as the time from illness to donation increases.
So, if you look at the chart there in yellow, going from donors who reported symptoms before donation to donors who reported symptoms after donation, you can see that the donors who reported symptoms onset right about the time of donation had the highest levels of viremia, and this was statistically significant.
So, we looked at the implicated donors versus non-implicated donors, and as I mentioned before, we had 14 implicated donors, and we had 654 other donors who were IgM-negative, who we used as a comparison group. Three of the symptoms of these donors were independent predictors of West Nile virus infection, and those were fever, eye pain, and new rash.
Now, looking at the fever column here, you can see that 9 out of 14 of the implicated donors reported fever compared to 3.7 percent of the non-implicated donors. These 9 donors were all of the symptomatic donors. The other five people were completely asymptomatic and had no symptoms at all.
So, we tried to look at what kind of symptoms could discriminate the implicated donors from the non-implicated donors, and as I mentioned earlier, fever caught all of the symptomatic implicated donors, but 3.7 percent of them, of the non-implicated donors, reported fever, so that, by itself, would not be a very good screening tool, so we tried to look at combinations of symptoms.
You can see that fever and eye pain, or fever and new rash seemed to be pretty good predictors of acute West Nile virus infection.
So, what I tried to do then is look to see what the sensitivity and specificity in predictive value positive of these combinations of symptoms would be. If you look at the first column there, I have incidence per 10,000 donors, and I just came up with a number. There is some rationale for this number, which I will explain in a minute.
But looking at, for example, fever, if you assume that the incidence is 10 per 10,000 donors, if you assume the sensitivity of fever as 64 percent, the specificity as 0.963, the predictive value would be 0.174, or put in another way, for every West Nile virus viremic donor, you would exclude 57 normal donors.
If you look down at these combinations of fever and eye pain, and fever and rash, the predictive value positive is actually quite good.
Now, I would like to go on to the next topic and talk about the estimated risk of transfusion transmission in the United States during the 2002 period. These results are based on a model that we published in Transfusion, in the August issue I believe it was, last year, so I won't go into the methodology.
But what we found was that the average risk in the United States last year was 0.36 per 10,000 donors with a maximum risk of 1.55 during the peak of the epidemic. Now, I have also presented on this slide, state specific estimates of risk, and you can see that among these higher incidence states, the risk was considerably higher than the national average, with a high in Michigan of about 4 per 10,000 donors overall during the epidemic.
In this slide, I just showed the same data showing the actual curve of the estimated risks for the same states, and as you can see, during the peak of the epidemic, in Michigan, we estimated that around September 1st, that the risk was about 10 per 10,000 donors.
In some high-risk cities, such as Detroit, the risks actually were up to like 20 per 10,000 donors.
So, then, we decided to try and validate this model with the Red Cross and Sue Stramer, and did a West Nile viremia study in which we took samples from Cleveland and Detroit during the first three weeks of September 2002. According to our model, we estimated that the risk would be about 8.2 per 10,000 in that population.
We tested all of these samples by TaqMan PCR in our lab, and the very preliminary results--I want to emphasize very preliminary--of the first 5,761 tested to date, the point prevalence of viremia was 6.9 per 10,000, and that was based on 4 positives. But what I learned this morning, walking in the room, there was actually 6 positives, so the point prevalence is actually higher than that, and it is probably going to come very close to our estimated point prevalence of viremia.
So, I think our model for transfusion risk is probably going to be pretty close.
Now, one of the things I am always asked is what are the trigger points in trying to predict future epidemics particularly in this season for making decisions about blood donor screening.
We have a national system called ArboNet, which is a national surveillance system, which we look at birds, mosquitoes, horses, and humans, and West Nile virus infection in those. It is a real-time, web-based reporting system.
The first measure I call epizootic measures, which is dead birds. What we found is that dead birds are extremely sensitive to detecting West Nile virus epizootic activity in an area, but not very specific for human disease, so as a discriminator to look for future West Nile hot spots, it is probably not very good.
What people have tried to do is quantitate dead bird surveillance to get a better measure of West Nile virus risk. What we found is that in certain places it does correlate with human risk.
If you look at indexes like dead birds per square mile, you find in places with higher measures of epizootic activity, it does appear to be correlate with future human risk, but this is very difficult to evaluate because the national surveillance is not completely standardized across the whole country.
So, certain states will test a few dead birds in a county and quit, others will test hundreds, so it is very difficult to try and translate these data into a national or a local estimate on a national level of risk.
The other things we look at is mosquitoes which are not timely and not universally done. It is very hard to do mosquito surveillance. Horses, it is the same problem with birds except horses are less evenly distributed. Dead horses aren't particularly useful in Manhattan, for example.
So, if we look at all these epizootic measures, probably none of them are terribly useful for determining trigger points for making decisions about blood donor screening.
The thing that is probably more useful is human illness surveillance. The one thing we do not encourage or standardize is surveillance for West Nile virus fever because there are probably 500,000 cases of West Nile fever last year in the United States, so it is just impossible to do surveillance for something of that frequency, so we really try and do West Nile virus meningoencephalitis, which is universal across the country, and from those data, we actually can estimate transfusion risk, and those data were used to give those models the data that I presented earlier, and we have come up with a simple formula that people can use to estimate transfusion risk in their area.
We can get a good idea of the size of the future epidemic from early season human cases. So, for example, if an area has got human cases in June, we know that there is going to be a huge problem later on.
The one problem is that there is a short lead time between human cases and transfusion transmission, and among these counties where we found transfusion-related cases, the donors of these cases often donated only a couple of weeks after we had recognized the first cases in those counties.
So, human illness surveillance is probably a good trigger point, but the lead time is rather short.
I would just like to end with a little bit of information about experience in our lab with IgM antibody testing and in the sera submitted to CDC from 1999 to 2002, IgM was detectable in serum and CSF on the date of onset in 99 percent with 6 exceptions in serum of 800 tested and in 10 exceptions of CSF in 800 tested.
I would like to put one caveat to that, is that when somebody submits samples to us and writes the date of onset, we don't actually go back and try and verify that that is the case, so the date of onset may not exactly have been correct.
The other thing is most of these people that got submitted samples were from people with meningoencephalitis, and not West Nile virus fever, but the bottom line is that at least in people with meningoencephalitis coming in your lab on the date of onset, most people are already IgM-positive.
Almost everybody is IgG-positive by day 7 post-onset, we also found the P to N ratio was 3 to 5 times higher to West Nile virus than SLE.
The other thing is IgM persists for a long time.
This is the slide that shows P to N ratios in days post-onset up to 400 days, and you can see that with cutoff of 3, that a large number of people are still IgM-positive 4- and 500 days out. This data was published this month in the Emerging Infectious Disease Journal, which has an updated chart to 500 days out.
I would just like to say that all of this work, it was really a tremendous amount of work, and there were a lot of people involved in collecting it all, and I would particularly like to thank Sue Stramer and the others from the Red Cross and the other blood collection agencies, and FDA, for helping out with all of these investigations.
DR. NELSON: Thanks.
When the donors, the 600 or whatever, were questioned, and you looked at those that were implicated and those that weren't by its symptoms, did either the interviewers or the donors know that they were linked to the transfusion case at the time of the interview or was this blinded?
DR. PETERSEN: They all knew that there was a possibility that they were linked to a transfusion case. Each of these cases had an average of 36 donors. Nobody knew which one it was until they got questioned and then the samples went to the lab and they were tested.
We had no clue which one it was at the time of the questioning.
DR. NELSON: A few of them had a fever at the day of donation. My understanding of the way that the transfusion services work is that usually, temperatures are taken. Actually, when I have donated, the temperature is taken prior to collecting the unit. Did that occur just when they went home or something, after donating, or why were some febrile at the day of donation?
DR. PETERSEN: I think there is two things to consider. One is we didn't ask them specifically what time of day, we just asked them what day, so we don't know if they were febrile before or after. The other thing is keep in mind the big caveat here. You are asking people about a mild illness two months earlier, so the reliability of this data are a big questionable.
DR. WILLIAMS: I just wanted to probe a little bit on the morbidity and mortality of the recipient cases.
Of the 12 meningoencephalitis cases, one case died. Of the remaining 11, can you say anything about long-term medical impact of that? A second question. Six of the recipients died. I gather from what you said earlier you couldn't either rule in or rule out West Nile involvement. Could you just say anything more about that potentially?
DR. PETERSEN: All I can say is it is really hard to determine what people die from, because we have looked through these charts just in incredible detail, and a lot of these people were extremely sick to begin with, and particularly where it got difficult is with somebody with an unknown febrile illness, some of these people were septic before they got their implicated transfusion, so it was just really difficult to figure out killed them for one thing or if they even had West Nile virus illness.
Some people it was very clear-cut where they just got the transfusion and developed the meningoencephalitis. Others, particularly if they had milder symptoms, it was very hard to sort out and some of these people we could never sort out to any kind of degree of certainty.
The other thing I might like to add is that there was a certain selection bias of how we got these people, because the ones who had, let's say, the index cases, the ones who were identified with possible transfusion-related West Nile virus infection, by definition, had something that alerted the doctor that they may have had West Nile, so most of those people had meningoencephalitis, but if you looked at the co-component recipients that we went back and identified, most of those people were either asymptomatic or had mild illness that we couldn't separate out from their other underlying illness.
DR. ALLEN: I think that is an important point. Obviously, the volume of infectious blood that one gets from the mosquito is very different from what one gets in a transfusion, so that is interesting that they still could have gotten that as a co-component recipient and still had very mild illness. That says a lot I guess about the host response.
DR. PETERSEN: Right.
DR. ALLEN: The other major question that I am struggling with is this whole issue of viremic states continuing despite the presence of IgM antibody. It would nice to be able to parse it out along those lines if you could in terms of accepting or rejecting donors.
From this study, I just want to reiterate, as far as you were able to tell based on the segment analysis of the donors, there were no donors who were IgM-positive at the time that they donated that blood, is that correct?
DR. PETERSEN: That's correct. We did have one case where there was somebody that was IgM-positive and viremic, had a very low level of viremia at the same time. This was a person who was actually a blood donor who subsequently got West Nile virus infection and called the blood collection agency. The blood collection agency pulled the plasma. That person was both at the time of donation.
Unfortunately, the recipient of that person's blood died, was very sick and died before we could go back and investigate the person, so we actually do not know if that person transmitted infection to the recipient or not, but everybody else was IgM antibody negative.
DR. DAWSON: George Dawson, Abbott Laboratories.
I had two questions. One is from some of the meetings I have been at, it has been stated that only about 1 percent of individuals infected developed CNS involvement, meningoencephalitis, and we have only seen 20 cases of transfusion associated and 13 of those were encephalitis or meningitis.
What is the total number of transfusion-associated transmissions do you think occurred? That is one question.
The second question is--and it will be a difficult one to answer, but it certainly wasn't 20, it might have been some number with a thousand, you know, with a comma--I have it heard it discussed also that people or birds or animals, et cetera, may be negative in their blood for viral RNA, but the virus may be harbored in tissues.
I don't know if you have any evidence on, you know, maybe some of the individuals who have developed encephalitis or meningoencephalitis, did they become RNA negative in serum and positive in their tissues, either brain, CNS, or other.
Does that pose a matter of concern for this committee in looking at using RNA only as a criteria for reinstating donors because if you are non-viremic, but it's at 1 virus/mL and you can't detect 1 virus/mL, could it be harbored at low level and periodically come out in serum.
DR. PETERSEN: I think your second question is of real interest particularly for the organ transplant community. There are data that do show that you can find live virus in organs of people when they are not viremic, and the best data, there is a 1954 paper by Severne and Gore [ph] where they experimentally infected people as part of some experimental cancer treatment. These were not normal people, these were people with cancers.
There is a table in that where you can actually see that some of the people are not viremic at least by the measures you could determine in 1954, but they could isolate virus from various organs. So, that is of concern.
I think as far as the issues for blood donor community, I think you just have to go with your data, and the data show that of all the known transmitters that we have, all were IgM antibody negative, and it is probably the best data you are going to ever find.
As far as the risk goes, there is probably hundreds or maybe, I don't know how many, I haven't done the calculations, but based on my model, you can actually do the math and figure out how many transfusion transmissions probably occurred.
Our system was probably extremely insensitive for determining transfusion transmission because we were first focused on trying to find the sickest people, to begin with, and second of all, all these cases came to light way late in the season, after the initial organ donation cases occurred in early September, so we really didn't have good surveillance for this all through the season, we were just mainly picking up people towards the end.
DR. DAWSON: Thank you.
MR. CAVANAUGH: Dave Cavanaugh from the Committee of Ten Thousand.
I wanted to follow up. Dr. Williams asked you a two-part question and one part didn't get answered, which was the long-term effects, and since there was something in the press from Canada in late February about 80 percent having chronic neurological damage of those hospitalized, I was wondering if you could comment on that story and then on the findings that you found.
DR. PETERSEN: As far as the long-term effects of West Nile virus infection, there are no good long-term studies looking at the long-term effects of West Nile virus infection. We are in the process of doing those right now.
Some of the best data come from the New York City, people identified in the 1999 New York City outbreak. These people were followed up a year later. About 50 percent of them reported, and this is self-report, some kind of residual neurological problem, but well done, long-term studies are lacking and we are in the process of doing them right now.
DR. STRONG: I was interested in your opinion of the progression of this epidemic. We have seen mostly cumulative data. What is happening on the back side of this, everything is moving west, but what is happening, say, in New York with the surveillance data?
DR. PETERSEN: That is actually a good question. What has happened is that in places where there has been West Nile virus infection in humans, that has persisted in the same places. The story has kind of got buried because there was a huge epidemic in the Midwest this year, but if you looked at New York State, New York State, who has now had four years of experience of the virus, had more cases last year than they have ever had.
So, where it has been, it appears to stay, and despite the huge epidemic last year, in which my estimates are about 500,000 people became infected, that is only about 2 percent of the population in that area, so there is stili a huge susceptible group of people.
DR. STRONG: And the bird population, are the numbers there about the same, as well?
DR. PETERSEN: The bird population, we don't exactly know which species are the primary reservoir species, but we suspect house sparrows are one major reservoir species, and what we found last year in Louisiana was that about 30 percent of the house sparrows got actually infected in that area where they have the huge epidemic and epizootic.
The problem with these birds like house sparrows is they turn over every year and so you have a whole new fresh susceptible crop the following year, so it is a little bit hard to generalize from levels of birds what is going to happen in people.
DR. SCHMIDT: It has been said that as far as testing blood donors, it won't be worthwhile in a few years because everybody will have been exposed by mosquitoes, but what you just said, that is not true.
DR. PETERSEN: I can't predict what will happen. If you look at what has happened with St. Louis encephalitis, which is a related flavivirus, very closely related flavivirus, what you see with St. Louis encephalitis outbreaks is that is focal epidemics involving hundreds of people, there are occasional regional epidemics involving thousands of people like what occurred with West Nile last year, and in some years there is nothing that happens, there is a handful of cases, and trying to predict where and when all these epidemics will happen is very, very difficult to determine.
Now, is West Nile virus like St. Louis encephalitis? In some ways, it is, and, in fact, there was a very large epidemic of St. Louis encephalitis in the Midwest in 1975, that actually, interestingly enough, affected the same states, the same cities, and, in fact, the same neighborhoods as the West Nile virus epidemic did last year.
So, we think St. Louis encephalitis may be a good model for West Nile. The caveat is that West Nile virus produces viremias in birds that are hundreds and millions of times higher than St. Louis encephalitis virus.
We think its epidemic potential has been much, much greater than for St. Louis encephalitis virus.
DR. NELSON: It also kills birds, which St. Louis is unlikely to or has not been reported to do.
DR. PETERSEN: Yes, West Nile kills birds and St. Louis encephalitis doesn't, but that is actually another interesting story, but the bottom line behind that story, which is quite fascinating, is that it appears that it is not all strains of West Nile virus that kills birds, it is very particular to the strain that is circulating here and the identical strain that is circulating in Israel.
As far as its killing birds, the species that are probably the most important for maintaining the virus like house sparrows, which get huge viremias and there are lots of those kinds of birds, the mortality rate is rather low, and it kills 99 percent of American crows, but there are relatively few crows, so it is probably not a species that maintains the virus.
DR. NELSON: I think we need to move on. Thanks, Lyle.
Sue Stramer from the American Red Cross.
D. Donor Serologic Studies of WNV 2002 Outbreak
Susan Stramer, Ph.D., ARC
DR. STRAMER: Thank you.
First of all, I would like to thank all the collaborators who I couldn't fit on the slide, both from the Red Cross and outside of the Red Cross, who provided information and data for this presentation.
I will cover two basic areas in this presentation, preclinical studies really taking off where Lyle and Christina Giachetti left off just previously, cover the case investigations that Lyle outlined, and prevalence studies that Lyle called the viremia study.
The two types of viremia are prevalence studies that I will talk about. One, the CDC study that Lyle has already mentioned; and two, a Gen-Probe study or the Phase I study that will be part of the IND for Gen-Probe.
Then, I want to talk about our Phase II studies, which is really systemwide IND implementation of West Nile virus testing, and some of my comments I believe will be very helpful to this afternoon's discussion as the FDA goes through guidelines for implementation.
Lastly, I will talk about contingency planning, what happens, although we say we are going to be testing by July 1st, what if the mosquitoes aren't quite as compliant and we have West Nile human activity prior to that.
The first study involved the case investigations from CDC. Those were the possible transfusion transmission cases during the last calendar year. We were involved in the investigation of 32 cases, and from those 32 cases we were able to obtain co-component frozen plasma from which the other products from the donation were transfused and possibly associated with a West Nile-positive recipient.
So, many of these cases were described in the 21 confirmed cases by Lyle and the other 21 potentially linked cases. In those cases, as was described, red cell segments were tested from implicated donations and follow-up samples from implicated donors.
We also were able to obtain 383 frozen plasma unit samples from these 32 cases. They were tested by CDC TaqMan RT-PCR assay, donor follow-up was tested by the IgM ELISA.
CDC has two protocols for testing. One is the standard input assay which was done on all of these samples, however, if a donor was found to be IgM-positive and negative on the standard input assay, a second assay was used, which is the high input TaqMan assay, which has about a 10-fold increase in sensitivity.
One other comment I would like to make about these 383 plasma, they were in a panel, so I will not only show CDC data, but I will also show panel results from Gen-Probe. We have also sent the panel to National Genetics Institute, and we will have those data soon. They have also been sent to Tetracore as has been referenced, and any manufacturer who is interested in testing these samples may do so.
So, the 383, and many of these were already described, but you will see their context a little bit differently, we had 11 total RNA positives from either the retrieved frozen plasma unit or the stored red blood cell segment.
Ten were plasma reactive and 2 of these 10 were also reactive in the segment. Two of the 10 were nonreactive by the standard PCR assay and required the high input PCR assay for reactivity.
The 11th, or one additional, was plasma equivocal, but segment was reactive, and this result possibly was false positive. When we look through the data, you will see that.
Eight of the 11 had follow-up samples collected and in all 8, IgM was demonstrated upon follow-up.
Now, these are the Gen-Probe results for the same 383 specimens. There were 15 initially reactive and 2 within a 10 percent negative gray zone. So, those were all retested. Three, the initially reactive, were IR, that is, initially reactive only the reactivity did not repeat.
The 2 gray zone samples also did not repeat. Of the 12 that were repeat reactive, 11 of 12 had very high signal-to-cutoff ratios and had reactivity in replicates at 1 to 8, and 1 to 16 dilutions, also with very high S to CO values.
One of the 12, and this is the case Lyle ended with discussing, with the equivocal IgM results, had a moderate S to CO value with plus/minus reactivity at both the 1 to 8 and 1 to 16 dilutions.
Eleven of the 12 were the ones I mentioned that were reactive at CDC, so in comparison with Gen-Probe, 100 percent of the CDC reactives confirmed, and here we had 9 of 12 donors with follow-up samples available. All 9 demonstrated IgM positivity on follow-up.
This is a line listing and I will show three slides of this, of increasing complexity. So, we will start with the simplest first.
These are the cases. These are the CDC-positive results. Here you can see 2 that were positive in the red cell segment and the plasma. These are very interesting samples that have the strongest reactivity, as you will see in the next slide, and represented the points closest from collection to transfusion that Lyle showed in the previous presentation.
These were the two samples that were initially negative, but then were positive following high input PCR. these represent means of either duplicate or triplicate testing in the Gen-Probe assay. Each color is a different dilution. This shows you the IgM data on the last column.
These three were not considered confirmed cases according to the CDC definition because they did not have IgM seroconversion in the donor, and this is the IgM equivocal case.
So, how what I have added are the CDC plaque-forming units/mL or way to get an idea of how much virus was present in all of these samples. The PFU was derived from a standard curve of previously calculated viral PFU/mL, and the standard is run with each TaqMan assay.
So, here are the two samples that were positive in both red cell segments and plasma. You can see here they have the highest viral loads at least by PFU/mL, and the red cell segment, as Lyle pointed out, was not as reliable, and there are lower viral loads from the red cell segment.
Here is the sample I referred to that is potentially a false positive since the plasma result only gave us an equivocal result.
This was the center column that were color coded. These are the replicate signal-to-cutoff ratio in the Gen-Probe test. I show you this for the reason that these samples that were linked to transfusion transmission, and were the most important, these were the IgM-negative samples also had great dilutional capacity and that they were detected, except for the one sample which we will get back to, by all dilutions, and you don't see the dilutional strength decreasing at a 1 to 16 dilution. If I just show you plus or minus results, you don't know from the plus value how strong that signal is and the confidence of that positive result.
Here, you see the one weakly reactive RNA-positive sample, replicates were reactive, but then as you go into further dilutions, we did have discordant results.
Moving into the next study, we have created at the Red Cross, a repository of 89,000 samples, and these were collected from 6 high-risk regions during the West Nile virus epidemic last year. These have been stored frozen, and the samples have been saved for West Nile virus study with IRB approval using a waiver of informed consent for these studies.
Samples have been retained in a database with basic demographic information. 8,000 are in the process of being tested by CDC using the high input PCR assay, and this is the viremia study that Lyle mentioned.
They are composed of the first three weeks of September from Cleveland and two weeks from Detroit. We selected these by metropolitan zip codes, and the preliminary results to date have already been discussed. We have six positive of 5,761 and many more with equivocal results or marginal reactivity seen at long cycle times, so we will have to do additional testing to determine how many of these equivocal and other borderline reactives are true positives.
So, what we are doing with the bolus of these samples from this repository is we will do a Phase I preclinical study. The test samples will be tested individually and we are targeting about 35,000 to 50,000. The purpose of the study will be to define the West Nile virus RNA prevalence and risk during the 2002 U.S. epidemic, determine the efficacy of pooling. From positives, we can determine the viral loads, we can perform 1 to 16 dilutions or 1 to 8 dilutions, and be able to see their reactivity.
This will also serve to begin clinical sensitivity validations for the Gen-Probe assay. This will be a linked study, so we will be able to go back to donors and identify IgM seroconversion to confirm the test results, that is, the RNA test results, and we hope to go back to the consignees, be able to identify the recipients and determine the infectivity of transfused products.
These two slides just show the samples as we collected them by week beginning with the first week of September. This is when I requested samples be retained, and they really kicked in at about September 5th. So, these are samples that we collected from St. Louis, Detroit, and Cleveland.
Memphis, Gulf Coast. These represent our Mississippi collections. We have fewer collections in the Chicago area, but whatever they are will be tested.
The next two slides are older slides of Lyle's, but it was the information we had at the time that we created the repository and served as the basis for some of the site selection. So, these are five of the cities that we have selected and at that point in time, what their estimated average risk was. Mississippi is not on here, but we know the epidemic occurred a little bit earlier and Lyle showed about 2 per 10,000 is the average risk.
At the time, this has been more fine-tuned in the graph that Lyle showed, but it just shows the relative risk at the time we pooled the repository samples for the various areas that appear on the repository.
How will we do the study? The testing was slated to occur in March. Now, we have slipped to April. Our FDA inspection team in the Southern California area helped us move the study out a little bit by doing an inspection of the laboratory, but they are now gone fortunately.
Anyway, so we are targeting now to begin the study at least instrument validation and training mid-April and towards the end of April we will start testing the first samples.
Study duration will be a minimum of nine weeks. It will involve staff from all of our NAT labs and that will facilitate their early training on the assay. The testing will occur by week starting with samples from the first week in September and continue until two weeks of NAT-negative results are obtained.
If you remember, the previous slide I showed you at the beginning of September, the peak was the highest and we believe with the first couple of weeks of testing, it will rapidly reach two weeks of NAT-negative results because the curve declined very rapidly.
We will use IND reagents and a draft package insert. Final assay software, we will not have available. I mentioned this is a linked study, so positive donors and recipients will be contacted for samples. We will use a similar questionnaire to that, that was used for the CDC case investigations, and the data will be submitted to FDA in support of pools of 16 and the use of the Gen-Probe reagents.
Then, our implementation for Phase II will be our full scale, as I just said, implementation, which has been approved through management at Red Cross. We are targeting, as the rest of the industry is, an implementation date of July 1st.
We will use the same NAT labs within the Red Cross that do HIV and HCV testing, using the same pool of 16, but using a Gen-Probe West Nile assay in addition to the HIV and HCV assay. We will have a "companion" IND with Gen-Probe, and the study has already been submitted to our IRB for review.
During the first three months of testing, we will have to report results from the testing labs to the regions using a manual reporting process that is fairly comparable to what we used when we launched HIV and HCV NAT in 1999.
Our protocol for testing will be again pools of 16. If a pool of 16 is reactive, they will be resolved to the individual NAT reactive donation. All nonreactive pools, the donations from those will be released and nonreactive individual donations from reactive pools will also be released.
I want to mention, because FDA will discuss this, this afternoon, NAT negative donations in NAT reactive unresolved pools, that is, if a pool is reactive, but all 16 donations test negative, will be released based on the efficacy of the model we have for HIV and HCV NAT, and that model efficacy was based on the fact that reactivity of NAT reactive unresolved pools could never be reproduced, and many studies were done trying to reproduce unresolved NAT pool reactivity.
Also, at the Red Cross, we have done extensive follow-up studies demonstrating no NAT reactivity or seroconversion upon subsequent donation from donors in NAT reactive pools, and this included 15,689 donors.
We all know and recognize that the undiluted sample represents the gold standard, and this is a quote from the FDA document. "The test result on the individual donation is considered to indicate the infectious status of the donor."
So, we believe that those donations from NAT reactive pools that don't resolve are acceptable for use.
We will confirm NAT reactivity by an alternate NAT method, most likely we will do PCR once we have NGI sensitivity, that is likely the procedure we will use. We will confirm NAT in both the source tube and the retrieved plasma unit, which we will try to collect from each region as we do for HIV and HCV.
We will also do investigational IgM testing assuming we find an IgM test that will be available in the July time frame. Otherwise, we will have to use CDC or a reference lab. We will have to plan contingencies for that, but that will also be done in the source tube and the retrieved plasma unit.
Our policy for retrieval and lookback under the IND period will be donations for the prior 28 days from NAT reactive donors.
NAT reactive donors will be temporarily deferred until follow-up testing has been performed, and this is the criteria of our follow-up test results. If the donor demonstrates West Nile RNA clearance by individual NAT, which is how we will do the follow-up testing, and they produce IgM, the donor can be reinstated as long as 28 days from index has elapsed.
If the donor is RNA positive, another follow-up sample is required until they demonstrate clearance of virus.
If the donor is RNA negative, and IgM negative upon follow-up, we will yet request another follow-up if the donor has been confirmed in index by alternate NAT, meaning we know that this is a true positive sample. The reason we want another follow-up sample is we want to eliminate the possibility of intermittent viremia, which we don't believe occurs with West Nile, but that has never been proven.
Lastly, if the donation is RNA negative, IgM negative, and has not been confirmed by alternate NAT, no further follow-up of the donor is necessary because that would be classified as a false positive on the test of record screening.
At the end of the study period, donors not enrolling in follow-up will be automatically reinstated and we will also use a questionnaire for positive donors.
Why follow West Nile virus NAT-reactive donors?
Well, under an IND, and actually in a complex operational setting, a single protocol leads to the fewest errors. Also, the additional data provides donor the complete information on the meaning of their reactive investigational test results.
Another reason is we cannot mandate when a donor will provide a follow-up sample, and we need to wait for all of the follow-up results prior to reinstatement, so I just give you a little example to ponder.
If the donor provides a follow-up sample on day 27, and then if we are allowed to reverse deferral, the donor could donate on day 28, the unit could be transfused, the components could be transfused.
We then subsequently, based on the follow-up sample, could get NAT-positive results and then be in a very sticky situation.
Really, the major reasons have to do with the performance of a test under IND, in which the performance characteristics of the investigational test and viral dynamics in West Nile virus human infection is not well established. Without the follow-up data including all the IgM data, we will never be able to prove this.
True confirmation of infection is IgM seroconversion.
What will we do for contingency planning? Well, this is currently what we have in the mill. We are trying to create a three-month target of FFP, however, production and storage is logistically difficult.
We also, as I mentioned, in the prevalence study, we will have one lab up and testing, so in the event that there is a small outbreak or there is a small number of donations that need to be tested, we will at least have one lab up with limited capacity.
Lastly, we will create a systemwide 30-day inventory of NAT tubes and perhaps associated pools. We have to work out the logistics for testing once we bring the test up systemwide under IND.
These 30-day inventory tubes can be used as needed, depending on reports of human West Nile cases for the entire system or just, as I said, as needed, if there is specific focal points for testing.
DR. NELSON: Thank you.
Questions or comments?
DR. STRONG: You have got Phase II. This is not Phase IIA, right?
DR. STRAMER: This is Phase II, yes, correct.
DR. STRONG: So, you will be doing pre-release testing.
DR. STRAMER: Yes. What Mike is referring to, in the beginning of NAT, there were multiple phases of the rollout of NAT, and many blood centers including the Red Cross only released frozen products based on the results of NAT, and fresh products were released, labile products, even before we had the NAT results because of turnaround time issues, et cetera.
When this test is released, I believe the entire industry will do real-time testing, and no product will be released without the benefit of a NAT-negative West Nile result.
DR. STRONG: Do you have any information on the stability of the sample? You have got segments that are not showing as a high a viral load as the unit. Is that because of degradation or is that just because it's a small sample, do you know?
DR. STRAMER: The red cell segments are highly diluted, so it could be a combination of instability to the red cell degradation and then viral degradation, or it can be purely a dilutional effect of the way we have the red cell segments saved.
DR. STRONG: What about plasma versus serum? Are your archived samples--
DR. STRAMER: Yes, all of our archived samples are EDTA plasma stored in Becton Dickinson plasma preparation tubes. Stability studies, unlike for HIV and HCV NAT, where we were very aggressive in participating in the stability studies, we are not doing any stability studies. We are leaving those for the manufacturer, but our assumption is we are already processing NAT under worst case storage conditions, because HCV is very labile, and I would anticipate that West Nile is not going to differ from HCV.
DR. DAWSON: I have two questions. One, in your initial screen of the 5,700 samples, you found I think 6 positives. Were any of those tested for IgM?
DR. STRAMER: Not yet, no.
DR. DAWSON: That will be interesting to see that.
The second thing is you mentioned using IgM testing in your testing algorithm and that you are going to five or six sites. How many tests are you planning on that, approximately, over the season to need, and how many would be ideal to have for IgM? Just ballpark.
DR. STRAMER: It depends how many reactives we get. I mean that is really unknown. If we have no prevalence, we will not use a single IgM test.
DR. DAWSON: So, when you said IgM as needed, what you meant was that if there were NAT-positive, you would be testing those donors and follow-ups for IgM.
DR. STRAMER: That's right, and it is part of our confirmatory.
DR. DAWSON: You weren't saying that a co-screening with IgM and RNA?
DR. STRAMER: That is not our plan, and I also want to say the plan I presented is really what we can do, our system is at saturation. If we hope to be doing anything by July 1st, we have to make the systems as analogous to our current systems as possible, otherwise, we don't have rapid implementation.
DR. DAWSON: Would it be possible for you to send us under code those six samples that you found positive, so we can test them for IgM, IgG, whatever else we have and any excess you have of the 21 transfusion-associated cases, I would like to try our hand again, send them under code, let's see what we find serologically, so that I can know with a clear conscience and data that we are not shunting any IgM because we do want to participate as needed in IgM testing, and screening included.
DR. STRAMER: For samples we have units from, such as the 32 case investigations with 383 frozen plasma units, you know, everyone is interested in having volume of those. We certainly will be willing to share some of that volume for IgM testing.
Now, regarding the repository samples, the ones that were first tested by CDC, those will be tested also by the Gen-Probe assay first, and then positives will also go on to NGI for confirmation and quantitation. We also want to perform dilutional studies on those, and we would like to, in addition to that, perform IgM. Just how far can you stretch a 2 mL PPT will be the question.
DR. DAWSON: Okay. I am still interested, so keep us in mind.
DR. STRAMER: All right. I know your number.
DR. KLEIN: Sue, do I understand that the first dozen specimens from donors that you showed, that 3 specimens tested positive for RNA in two separate laboratories by two different technologies, but negative for IgM, and are considered false positives, is that correct?
DR. STRAMER: No. The 12 samples that I showed, 11 of which were positive at CDC, 12 were all positive at Gen-Probe; 9 of the 12 were positive for IgM in donor follow-up, but negative for IgM at index.
The one index sample that was equivocal, it had a variable set of test results at CDC, but when all was said and done, I have like several e-mails that say, in all capital letters, this was not IgM positive in index.
That sample demonstrated IgM at index, demonstrated IgM positivity on follow-up, but let's say very weak RNA reactivity at index and very weak reactivity of IgM at index. So, it was kind of a cross-over sample, you know, so you can imagine RNA coming down and IgM going up.
DR. BUSCH: I just wanted to clarify. The one-month inventory that you talked about at the end, is that going to be something that you are going to systematically do, henceforth, you are going freeze down and retain both the pool and the individuals for all your NAT screening, or is this just a contingency that will be put in place transiently to potentially need to test, and would that testing be done only in hot sites or if testing were delayed, you don't intend to back-test the entire inventory for a month once testing is licensed or IND available?
DR. STRAMER: What we are doing now, if we don't plan our contingencies now, they are not going to happen, especially in a system like ours that is complex.
So, in order for us to be ready for any event, we can store probably 30--what we talked about is a 30-day inventory. Those PPDs would be frozen down, and in order to even do this, you know, freezers have to be obtained, they have to be temperature mapped, all the validation, so it is not an easy undertaking.
We would store those samples in the event that there are pockets of West Nile human activity, and then we would have to decide within Red Cross, within the blood industry, do we just test those pockets that had West Nile activity, or do we test the entire United States, I mean our entire inventory.
I think we don't know that as of yet, but if we don't have the samples, we won't have the ability to make that decision.
DR. EPSTEIN: Sue, I just want to focus again on this issue of low-titer samples and possible correlation with IgM. It seems to me that we have two pieces of data that are pertinent.
Lyle has showed us that in the transfusion transmission studies, all of the index samples in case of transmission were RNA positive. On the other hand, Andy Conrad showed us that if you take community-based surveillance samples, it is not at all uncommon to find low-titer viremic samples that have a positive IgM.
What we don't know is whether samples of the latter sort (a) transmit, and (b) whether they would be picked up or missed in a minipool screening assay. So, I think that it becomes very important for us to design the Phase II trial based on the result of the Phase I trial.
In other words, when you run through the first 35 to 50,000, you presumably will find some positives. If there are false negatives at 1 to 8, or 1 to 16, I think we then have an open question whether we need to look for IgM when we move to Phase II.
It would be very informative, of course, whether those samples that are false negative at 1 to 8, or 1 to 16, have positive IgM. If they don't, we don't expect to find anything in Phase II, but if they do, then, they are more like the samples that Andy Conrad showed us, and I would think that even though we haven't proven that such a configuration is associated with transmission, we wouldn't want to overlook that possibility when we do the screening studies prospectively.
So, I am just a little bit uncomfortable that we know what to do in Phase I until we see what we get in Phase II--I am sorry, the other way around--
DR. STRAMER: Right, the other way around.
DR. EPSTEIN: --what to do in Phase II until we see what we get in Phase I.
DR. STRAMER: Phase I, all the samples will be tested neat, so we will be able to look at dilutional consequences, and we will be able to look at IgM in those samples.
DR. EPSTEIN: A second point. I was a little bit unclear, when you talked about further testing of the NAT-reactive donors at 28 days. Not all donors will be retested. I mean I agree fully if you get a reactive test on follow-up, you want to not re-enter the donor until they have ended up retested or negative, so they remain deferred pending further outcome.
But what about the donors that don't present for their follow-up, is the plan to hold them indefinitely in abeyance? What you said is that they would be re-entered at the end of the study, and I guess the question is whether the non-returns could be re-entered at day 28 automatically or day 56 automatically.
DR. STRAMER: We have had many discussions internally about this, and that is why I went through that slide specifically. It is very difficult to design a study operationally that allows automatic reinstatement and follow-up to occur at the same time, because errors will result and we have become very error-averse.
So, the easiest way for us to do the study and to guarantee that we have IgM data is to require the follow-up sample, and if the follow-up sample clears virus, demonstrates IgM seroconversion, the donor is automatically reinstated.
Now, what happens with those other donors, really, the way we have the study set up now, it depends on the length of the study. If the IND period goes one year, those donors then would have to be temporarily deferred for that period of time.
If we learn more about the study as we are progressing through the study, we can change policies prior to that, but at this point, we don't know the performance characteristics of the test. The only way to prove the performance characteristics of the test is to make sure we get the IgM data and demonstrate seroconversion.
DR. PHELPS: What is very important is that the way that Sue designed the study, is that she will be looking at the individual sample, because we believe the comparable sensitivity would have caught all the viremic samples that I presented in those clinical cases, and those clinical cases were people who again reported that they were febrile.
So, I think that the role of IgM will be much more clearly sorted out under Sue's paradigm where since she is testing an individual, if they remain viremic, wait a period of time, then, become non-viremic, the importance of IgM will sort itself out under her algorithm, because we really chose a different paradigm, in a sense, she will be testing individually, at least we will track that data.
Now, when you jump up to pools of 16, that may become relevant, but I think we will learn that data in Sue's study.
DR. GOLDSMITH: Did you address the issue of the turnaround time, that you will be able to test the more labile cellular components?
DR. STRAMER: I didn't show all of those logistics. When the five laboratories were polled as far as increased workload and delay in turnaround time, so we know what to communicate to our hospitals, there will be a one- to two-hour delay in turnaround time due to testing, but that will be absorbed in our normal process, so at the end of the day, there really won't be an ultimate delay in turnaround time. At least that's what they tell me. We will track that and let you know or you will let us know.
DR. NELSON: Thanks.
We are going to have to break for lunch now.
[Whereupon, at 12:30 p.m., the proceedings were recessed, to be resumed at 2:30 p.m.]
DR. SMALLWOOD: We are a bit behind, but hopefully, we can make up some time.
The next topic is Regulatory Pathway for West Nile Virus Testing, FDA Update. The first speaker is Martin Ruta, talking about Guidance for Industry.
E. Regulatory Pathway for WNV Testing (FDA Update)
Guidance for Industry
Martin Ruta, J.D., Ph.D.
DR. RUTA: Thank you, Dr. Nelson.
I want to thank everyone for participating today particularly Lyle from CDC, who interrupted his vacation, and device manufacturers, who came and presented a lot of interesting data this morning.
What you are going to hear now are a series of talks, first, starting off with mine, about the guidance document that we issued last fall, last October, plus some industry action that was voluntarily taken after that.
Then, you are going to hear some additional talks from folks at FDA about our thinking about managing donors and units in the face of testing and raising some questions and options that we have as we enter into the next season that Dr. Williams will present.
What I want to do is go through a guidance document that we issued on October 25th, 2002, but as part of background, last August, we were dealing with West Nile virus as a theoretical risk to the blood supply and very quickly it became a real concern in organ transplantation and soon thereafter in blood transfusion.
Since August, blood transmission by West Nile virus was confirmed, and that is the first time it was confirmed in this outbreak, so this is probably the most studied investigation of West Nile and West Nile epidemic that there has ever been, so we are learning a lot of new things.
Part of the assumptions that we made in coming up with our original guidance, you know, we have to look back and see are still valid. So, in order to assure the safety of the blood supply in the face of the risk of West Nile virus, we took two initiatives.
One is guidance on the management of donors and blood components, and the second is to facilitate rapid development of donor screening tests.
So, there were some assumptions going into the guidance and some that came up afterwards. One of the assumptions is that human infections were expected to occur seasonally during periods of mosquito activity, and we weren't sure if this would continue to occur year-round in some parts of the country. We thought in the southern states there might be ongoing infection, but my understanding is that there hasn't been a case reported since last December.
About 80 percent of infected persons are asymptomatic, and many of the 20 percent of individuals will have mild symptoms that are nonspecific, or flu-like illness, and again, as we have heard, 1 in 150 develops more severe symptoms.
The ratio of the viremia can be up to 29 days. This is based on a 1950 study that we have heard about of injecting cancer patients with the virus.
Continuing with the assumptions, viremia resolves rapidly after seroconversion to IgM, and we have seen that IgM positivity can persist for going on two years.
We know of no chronic carrier state for this virus, and the last bullet is something we didn't know about prior to the guidance, that we have learned about since, and you have heard this morning, is that there can be donors who exhibit symptoms prior to donation, are apparently healthy on the day of donation, and can transmit West Nile virus.
What I want to do is walk you through the guidance document that we issue last October, and then you will hear further thoughts later on from Dr. Williams.
So, we issued a guidance on Recommendations for the Assessment of Donor Suitability and Blood and Blood Product Safety in Cases of Known or Suspected West Nile Virus Infection.
The recommendations included that donors or potential donors with the medical diagnosis of West Nile virus infection be deferred until 14 days after the condition is considered to be resolved and at least 28 days from onset of symptoms or diagnosis, whichever is the later date. At the time, our current thinking was that in the absence of current or recent symptoms, an IgM positive antibody test result alone should not be grounds for deferral.
Donors who report an otherwise unexplained post-donation febrile illness suggestive of West Nile virus infection in the setting of West Nile virus transmission in the community. Again, the trigger here that we told people for last fall was human cases in the community, those donors should be deferred for 28 days from the onset of illness or 14 days after the condition is considered to be resolved, whichever is the later date.
Donors whose blood or blood components were received by a patient with a possible case of transfusion- related West Nile virus should be deferred for 28 days from the date of potential transmission. Bear in mind that there was a very intensive epidemiologic investigation that was going on last fall, in which CDC was doing a lot of work to investigate whether transmission was actually occurring through blood and, in collaboration with the blood organizations, were identifying possible donors who may have been implicated in transmission.
FDA recommended that blood establishments actively encourage donors to report post-donation illnesses that might potentially be associated with West Nile virus, that is, flu-like symptoms that include a fever, occurring within two weeks of blood donation in the setting of active West Nile virus transmission in the community. These are post-donation reports.
In terms of managing the units from the donors, we recommended that in-date components from current, prior, and subsequent collections be quarantined and retrieved if a donor later reports a medical diagnosis of West Nile virus.
The product quarantine and retrieval should cover a time period dating back to 14 days prior to the onset of illness and 28 days subsequent to the onset of illness. Again, our thinking at the time is that in the absence of symptoms, an IgM-positive antibody test result should be grounds for product quarantine and retrieval.
Medical directors should exercise judgment when a donor reports an otherwise unexplained post-donation febrile illness in the setting of active West Nile virus transmission in the community.
Again reminding everyone that there was a very intense epidemiologic investigation and that donors were considered to be potentially associated with transmission of West Nile virus if the infected recipient received the donor's blood components within 28 days before the onset of symptoms in the recipient.
For each associated donor, product quarantine and retrieval should occur for the in-date components that were collected in a 28-day window period prior to, and subsequent to, the suspect donation that might have been implicated in that transmission.
We also recommended that when a blood establishment received information that a donor has a medical diagnosis of West Nile virus, the blood establishment should notify transfusion services to permit lookback recipient tracing and notification.
If the post-donation illness was not diagnosed as West Nile virus infection, actions to identify prior recipients were not appropriate.
In the context of the epidemiological investigation, if there was a suggestion that a specific donor was a likely source, you know, again, there were cases where many donors were being investigated, but if there was one there that might have been a likely source of transmission of West Nile virus to a transfusion recipient, the blood establishment should consider lookback notification of other recipients.
That is the end of the guidance document that we issued. There were a number of other events that occurred after that, and I just want to highlight one of them.
That was an action taken by industry, and this was a voluntary action to withdraw products from the market. On December 12th of 2002, the AABB, ABC, ARC, and DoD put in place a voluntary market withdrawal of selected frozen transfusable in-date products collected in areas that experienced West Nile virus during 2002, as a preventive measure, to mitigate against the risk of transmission of West Nile virus through blood transfusion.
States were identified by the CDC and time periods identified as to when there were peaks in transmission, and the risk period was defined as seven days prior to the onset of symptoms of the first encephalitis case, and seven days after the onset of symptoms of the last encephalitis case in the respective state.
That is the end of my presentation and it kind of brings us up to date where we were as of last fall in our thinking that led to the guidance. Again, we are dealing with a very intensive investigation, the most intensive investigation that has ever occurred for this virus.
You are going to hear from Dr. Hewlett some thoughts on donor and unit management in the context of testing under IND, and from Dr. Williams about some of our current thoughts or possible options for donor and unit management in case testing might not be on line for the whole country if there is a West Nile transmission again this season.
DR. NELSON: Thank you.
Are there questions? Okay.
Next, is Dr. Robin Biswas, who is going to talk about the approval criteria for West Nile virus testing.
Approval Criteria - Robin Biswas, M.D.
DR. BISWAS: Good afternoon.
The next three presentations are closely related and I will go through my slides quickly to save time and also because a lot of what I will say will be gone into in more detail by Indira Hewlett and Maria Rios, the following speakers.
In regard to the approval mechanisms for assays used to test donors for West Nile virus, what you should know is that tests used in the manufacture of blood and blood components are reviewed by FDA as biologic products under the PHS Act.
This means that a manufacturer needs to present to FDA, an Investigational New Drug application, which is sort of a trial plan for the device, and after FDA approval of the IND, i.e., that plan, the manufacturer goes out, gathers the data, and then presents the data to FDA as a Biological License Application.
This is just the list of the things to consider.
In regard to clinical sensitivity, what we mean by that is the reactivity of the assay in individuals with known West Nile virus infection. So, you can test samples from persons with clinical West Nile virus infection, and another way to do that is to test samples in epidemic West Nile virus regions.
In regard to analytical sensitivity, we mean by that the reactivity of the assay in samples containing the analyte of interest, for example, West Nile virus RNA.
To do that, to get a handle on that, you can test serial dilutions of samples with the RNA or test serial bleeds from individuals with the disease; if possible, samples collected closely together very early in the infection. Of course, for this to be useful, you need standard reagents with defined units, and Maria Rios will be going into that in more detail.
Now, early estimations, we were told that the viral load in samples associated with the transmissions, as Lyle Petersen has described this morning, that the viral load varied from about 3,000 to 5,000 copies/mL, so our current thinking is that the sensitivity of the NAT should be targeted at least 100 copies/mL in the individual donation to ensure 100 percent detection of 1,000 copies/mL in the individual donation.
Of course, as data is gathered, and it is being gathered at a great rate, and as technology improves, this target might change.
Clinical sensitivity refers to the reactivity of an assay in individuals without West Nile virus infection, and you would like that to be low in the blood bank setting, of course. Otherwise, you are not going to use many suitable units. You can study this by testing healthy individuals, such as blood donors, persons in winter or in non-epidemic regions. One needs also to test individuals with other diseases.
I have included up there the other related viruses. We said earlier that we would encourage manufacturers to design their tests, so that they could pick up these, as well, but, of course, there is a problem in getting samples. I have sample scarcity there. It seems to be more a problem of getting ahold of those samples, they may be available elsewhere in other parts of the world.
In regard to clinical specificity, the assay should be tested in the end-user setting with a U.S. donor population, in geographically separated donor sites, at least three lots of the assay, and confirmatory additional and supplemental testing needs also to be addressed, and Indira will go into that in more detail in a moment.
In regard to analytical specificity, this is the reactivity of the assays with samples not containing the analyte of interest, and obviously, it overlaps with clinical specificity.
You can test samples with other disease markers, and you can test samples with potentially interfering substances to assure that the assay has very good specificity.
This slide deals with controls during the actual manufacture of the kit. Controls and calibrations during the kit manufacture are needed to assure lot-to-lot consistency of assay reagents and the finished kits.
The control reagents need to have varying degrees of reactivity, and the endpoint titration curves from testing the assembled kit with panels of samples should have slopes and midpoints that fall within validated acceptable limits.
In regard to assay stability, studies on at least three lots are needed to demonstrate stability claims for whatever time periods or under what conditions that the company claims.
In regard to reproducibility and operator proficiency, a panel of plasma and sera, composed of positive, negative, and weakly reactive sera, should be tested in at least three sites with different operators with at least three lots of the assay.
Each study site should demonstrate operator proficiency with the panel before screening donors.
in regard to instrument and software, the instrument and software portion of the application should be included in the BLA, in the Biological License Application.
If it has been approved or cleared for another use--that mean another assay--you need to submit data demonstrating that it meets the new intended use of the assay that you want to now get approved.
There is a CDRH guidance for this, and I have it up there.
Additional guidance documents, I have listed here. There is a guidance in regard to the HIV NAT tests that came out in 1999, which is very helpful, and considerations in this document are applicable to other gene-based tests.
Another guidance document, a rather old one, but still very relevant, is this one that I have here from August 8th, 1989, about anti-HIV test development.
The way forward. We will continue to work with other components of the Department, with NIH and CDC, and the manufacturers and blood organizations to facilitate assay development.
If necessary, FDA would allow widespread study of appropriate tests under IND. I think Indira will be going into that in more detail.
Sponsors are asked to seek FDA's guidance and to submit a pre-proposal before initiating studies to support an IND.
Thank you very much.
DR. NELSON: Thank you, Robin.
Questions or comments? Okay.
Next, is Dr. Indira Hewlett, who is going to talk about clinical study design, unit and donor management.
Clinical Study Design, Unit and Donor Management
Indira Hewlett, Ph.D.
DR. HEWLETT: I am going to be discussing FDA's current thinking in regard to clinical trial design for validation of West Nile assays and some proposed approaches to unit and donor management based on test results.
By way of background, all reported transmissions by blood transfusion has occurred in the acute, viremic phase, indicating that this phase of infection would be the ideal target for intervention.
Therefore, NAT may be the most appropriate strategy to interdict infectious donations.
Mostly likely, NAT on pooled donations will be implemented first due to practical considerations and since platforms for HIV and HCV NAT on pooled plasma have already been licensed.
But there are concerns about the impact of pooling on the sensitivity of NAT assays and the ability to detect specimens with low levels of virus.
Also, it is known that viremia and detectable IgM co-exist in the acute, late phase of infection--there has been a lot of discussion about it this morning--but lack of data on West Nile virus transmission by donations, that are minipool NAT negative and IgM positive, suggests that it may be useful to explore whether a strategy involving a combination of NAT and IgM testing would need to be considered in IND studies.
Finally, although there is lack of evidence of West Nile virus transmission by plasma derivatives, screening of source plasma collections may increase the margin of safety achieved by virus removal and inactivation methods.
FDA recognizes the need to implement testing in a timely manner and would therefore allow large-scale studies and widespread use of tests under the Investigational New Drug Application if necessary, similar to HIV and HCV NAT implementation.
This approach would facilitate introduction of testing in blood and plasma centers while evaluating test performance.
FDA would recommend routine use of licensed screening tests to detect acute donor infections with West Nile virus when tests are available.
As part of clinical validation, FDA has traditionally required that a test be evaluated for clinical specificity, clinical sensitivity, and reproducibility.
I will be discussing clinical specificity and sensitivity. Reproducibility has already been discussed by the previous speaker, Robin Biswas.
For validation of the clinical specificity of a NAT or IgM assay for West Nile virus, the test would have to be evaluated in a low-risk population using specimens from low-prevalence areas.
In general, FDA considers a sample size of 10,000 tests, pooled or individual, to be adequate for clinical specificity determinations. Data on clinical specificity could be obtained by identification of negative cases during prospective studies conducted at various clinical sites under IND and retrospective studies of repository specimens identified in clinical and donor settings during previous epidemics.
For investigational NAT assays, clinical specificity of NAT false positive or reactive specimens could be established by additional testing with an alternate NAT, that is, a NAT assay which uses primers from a different region of the viral genome, IgM, and follow-up testing.
Investigational IgM test results could be confirmed using a NAT and a second IgM test and discordant IgM results resolved by plaque reduction neutralization assays.
At this time, testing of all donations by both NAT and IgM during clinical trials is something that may need to be considered since testing of minipool and IgM assays are not well established.
For clinical sensitivity determinations, limited numbers of samples exist, therefore, multiple approaches may need to be considered.
Clinical sensitivity for both NAT and IgM could be determined using repository specimens including transfusion and West Nile virus illness related, community-based specimens from donor and community settings.
In addition, identification of positive cases during prospective IND studies conducted in blood bank and community settings, and testing of seroconversion panels, if available, could provide useful sensitivity data.
For sensitivity determinations, reactive screening NAT and IgM results could be confirmed by an alternate NAT and IgM assays including follow-up testing with NAT and IgM.
For a diagnostic indication, specimens from cases of West Nile virus illness in community and clinical settings could be tested. This intended use, of course, is regulated by the Center for Devices.
If the investigational test is intended for detection of various strains of flavivirus or JE serocomplex or the West Nile virus strains, the ability to detect these strains would have to be demonstrated.
These studies could be conducted with repository specimens from previous epidemics and using well-characterized panels.
There are no existing licensed or well-standardized assays for NAT or IgM testing. Therefore, FDA is considering some additional approaches for test validation.
FDA has initiated efforts to assemble an in-house qualification panel composed of pedigreed specimens as an additional tool for clinical sensitivity evaluation.
FDA believes that testing of a common set of pedigreed specimens by all candidate investigational tests may be a useful additional approach to determining whether assays have equivalent sensitivity.
Also, testing of all reactive samples identified in clinical studies by all manufacturers seeking licensure of West Nile virus tests would help further establish the relative sensitivity of candidate assays.
This combination of approaches should provide adequate data to evaluate clinical sensitivity for donor screening.
For this purpose, however, collaboration between blood, plasma, and clinical centers, test kit manufacturers, and governmental organizations is necessary in sharing of samples which will further facilitate test evaluation and expedite licensure.
I will now move on to discussing supplemental assays which are intended to confirm reactive screening NAT or IgM results obtained with the individual donation.
Supplemental NAT would confirm results for screening NAT, and a supplemental IgM assay, of course, would confirm either screening NAT or an IgM test result.
Clinical specificity for supplemental NAT assays could be obtained by testing a minimum of 500 repository or prospectively collected specimens from low risk and low prevalence areas.
Clinical specificity could be confirmed by testing with an alternate NAT and IgM assays.
Clinical sensitivity for supplemental NAT assays could be obtained by testing samples with reactive NAT results on the individual donation, and clinical sensitivity could be further confirmed by follow-up testing with screening and supplemental NAT and IgM tests.
Similarly, clinical specificity for supplemental IgM assays could be obtained by testing a minimum of 500 repository and prospectively collected specimens from low risk and low prevalence areas, as with supplemental NAT assays.
Reactive IgM results could be confirmed by a second IgM and an alternate NAT. Discordant IgM results could be resolved using plaque reduction neutralization tests.
Clinical sensitivity of supplemental IgM assays could be obtained by testing samples with a reactive NAT or IgM result on the individual donation.
Reactive results of the screening NAT and supplemental IgM test could be confirmed by a second IgM assay and follow-up testing with an investigational and the second IgM test.
Discordant results again could be resolved using a plaque reduction neutralization assay.
I think this has already been stated by Dr. Biswas, but I will just say it again, that FDA's current analytical sensitivity standard for West Nile NAT assays is 100 copies/mL for the individual donation, and that this standard may be revised as tests become more sensitive or as additional data obtained on the levels of viremia and infectivity in future studies.
I will now switch to the topic of unit and donor management on the basis of test results.
West Nile virus detection systems will most likely be minipool NAT based, similar to existing licensed platforms for HIV and HCV. Consensus algorithms for unit and donor management based on test results may be needed across testing platforms.
During the clinical trial, FDA will consider strategies for unit and donor management in order to gather data regarding which of these strategies would provide maximum assurance of safety against West Nile virus transmission by blood and plasma.
This approach may be necessary since the performance characteristics of investigational tests and viral dynamics in West Nile virus infection of humans are not well established. However, at this time, FDA is considering some interim approaches for unit and donor management, and we may revisit these strategies, if necessary, when IND studies are completed.
Similar to algorithms currently in place for HCV and HIV NAT, reactive NAT results on a master pool could be confirmed by testing individual donations with the same NAT.
Non-reactive units could be released, reactive units would be quarantined and destroyed or labeled.
Donors with reactive NAT results would be deferred for 28 days. This is a temporary deferral consistent with the longest known duration of viremia that has been described so far in the literature.
Reactive results in the individual donation could be confirmed using an alternate NAT and IgM test.
During clinical trials, reactive investigational NAT results could be confirmed by follow-up testing. Donor may be retested prior to 28 days. Follow-up testing would include the investigational NAT, alternate NAT, and IgM testing.
If the follow-up sample is NAT reactive by the investigational NAT or alternate NAT, the donor remains deferred for an additional 28 days.
The donor may be eligible for reinstatement if follow-up sample prior to the 28 days is negative by the investigational NAT, alternate NAT, and IgM positive.
However, the donor would continue to remain deferred for the 28-day deferral period.
This slide is just a schematic of what I just said to you, so we will skip it and move on to the next slide.
If individual donation NAT is used, non-reactive units may be released. Reactive units could be retested in duplicate by the investigational NAT, and reactive results confirmed by an alternate NAT.
Non-reactive units may be released, reactive units destroyed or labeled appropriately, the donor is deferred again temporarily for 28 days, notified and counseled.
The sample could be tested by alternate NAT, IgM, and/or plaque reduction neutralization test to confirm investigational NAT results.
Follow-up testing during the 28-day period would include investigational NAT, alternate NAT, IgM. Again, if a NAT is reactive, donor is deferred for an additional 28 days as with minipool NAT testing.
Next slide, please.
We will skip this one, too.
Reactive IgM results could be confirmed using a second IgM and NAT. If reactive on NAT and IgM, the donor is deferred for 28 days, notified and counseled.
Non-reactive NAT results again could be confirmed using an alternate NAT.
If the sample is NAT negative and IgM positive, the donor may be reinstated.
If testing was not performed during the 28-day deferral period, the donor could be reinstated after this time.
In summary, I have tried to outline FDA's current thinking regarding approaches to validation of West Nile virus NAT and IgM screening and supplemental tests, and some considerations for unit and donor management.
Clinical validation could include a combination of retrospective and prospective studies to increase sample size in clinical trials. Of course, this would involve the collaboration between industry and government organizations to facilitate timely introduction and licensure of West Nile tests for blood safety.
I will close with that and thank you.
DR. NELSON: Thank you, Indira.
Questions or comments?
DR. SOLDANA: John Soldana from CBS.
Indira, I am slightly confused about the detection limit for NAT. I think you said it was 100 copies/mL for the analytical sensitivity, and I think the previous speaker said 100 copies/mL, so that 1,000 could be picked up every time.
I think that makes the 100 about 50 percent detection limit, so are we talking about a 50 percent detection limit or a 95 percent?
DR. HEWLETT: As I understand it, we are talking about a 95 percent detection rate at the 100 copies/mL, such that the 1,000 copies can be detected in the original donation.
DR. SOLDANA: Thanks.
DR. GALLARDA: Indira, I have two questions. One is regarding validation of IgM assays with regards to clinical specificity. You touched on clinical sensitivity for both NAT and IgM, but you didn't or at least I missed maybe the clinical specificity requirements to validate an IgM assay. That's the first question.
DR. HEWLETT: I actually did have it on one of my slides. I can go back to it, but it is basically confirming with a second IgM test and resolving discordant results using a plaque reduction neutralization test.
DR. GALLARDA: Thank you. The second question I had, had to do with you mentioned that if you have an initial NAT reactive on individual testing, you could go and then repeat it in duplicate by NAT or do alternate NAT or IgM.
If the unit is tested reproducibly at two times after initial positive result, and those two replicate test points are both negative or one is positive, what does that mean with respect to donor management and unit management?
DR. HEWLETT: If you have got a discordant repeat test, obviously, you would want to run an alternate NAT. The IgM is really more as an additional tool. I think we are more focused on confirming your NAT test result with another NAT assay, so that is how the unit would be handled.
Then, of course, you would reflex into follow-up testing in the event you have got two NATs that are positive, you would want to do follow-up testing to confirm the NAT result including seroconversion to IgM.
That is how you would deal with the unit. The donor obviously is going to have to be notified and counseled. There will be deferral triggered by the NAT reactive test result that you obtain on the second test, on the duplicate test where you have either one or both as being positive. That would trigger the deferral.
The alternate NAT, of course, is to further confirm and to validate the test result of the original NAT test.
DR. GALLARDA: If the initial NAT screening test result is positive, and the two replicates for the same screening NAT assay are negative, what would you do with that unit?
DR. HEWLETT: You would still want to test using an alternate NAT at least during the IND phase, so we can collect that data to be able to say, well, you are done once you have a negative result on your duplicates, so it sort of back-validates the assay.
DR. GALLARDA: And that donor would be deferred until the 28-day follow-up period.
DR. HEWLETT: Right.
DR. PETERSEN: Just one question. If I understood you correctly, to look at the specificity of the IgM antibody assay, you were contemplating testing with a second IgM antibody assay?
DR. HEWLETT: Do you have other suggestions? There are no reference assays, there are no really reference materials for these assays, so you have to work with the tools you have, and that would essentially be testing it with another test, another test for the same marker.
DR. PETERSEN: I think probably you want to do a neutralization test.
DR. HEWLETT: Yes, that is actually part of the algorithm.
DR. PETERSEN: But not another IgM test, because the problem is that because of the cross-reactivity of the flaviviruses, you are going to get yourself into real trouble because even a low prevalence area, you may have somebody that may have had yellow fever vaccination, dengue, et cetera, previously, and you are going to end up with some problems sorting it all out without a new test.
DR. HEWLETT: I think the plaque reduction obviously is going to have to play a role in sorting out results of the IgM tests, there is no question that that is going to have to be invoked and have to play a role. Thank you.
DR. LEW: I think the question was asked a little earlier, but slightly different, but I am still concerned about the 100 copies/mL, because someone had also presented, and I think it was American Red Cross, saying that they are going to look at pools of 16 versus I think the way the design was, if you have a 1 to 6, or pools of 6, then, clearly, the level of 100 should detect that 1,000, you are quite comfortable.
But when you start moving on up, and you are getting larger and larger pools, is that something you all are comfortable with?
DR. HEWLETT: Yes, you are right, I think that is a very, very important point because obviously, we are looking at a disease where virus levels are not as high as what you would see in HCV or in HIV infections.
The more sensitive the test, obviously, it is more desirable. So, the 100 copies is just to set that limit as being the most desirable or the minimum level that we will accept at this point based on what is achievable by assays that are being developed at the moment.
I don't know if that answers your question.
DR. LEW: I guess I was just concerned if I was correct that the American Red Cross are looking at pools of 16, I honestly can't remember that presentation.
DR. HEWLETT: Yes, I think that is what we heard this morning, but Sue is going to get up and clarify it for us.
DR. STRAMER: There are multiple factors that you have to look at for sensitivity. Pool size is only one of those factors. Certainly, the sensitivity of the test, if the test is ultra-sensitive, it gives you more flexibility to do a pool of up to 16 donations.
So, the data that Gen-Probe showed at 50 percent hit rate was 2.9 copies/mL, which is considerably south of 100 copies/mL, so we feel, with that sensitivity, it will allow pools of 16. From the case investigations that I showed, the Gen-Probe test, even in a pool of 16 and replicates of 3, had S to CO values greater than 30 in all the CDC-confirmed samples.
So, we have some feasibility data at least that shows pools of 1 to 16 work on clinically relevant samples.
DR. HEWLETT: One does need to bear in mind that although these are diluted samples, there are ways to concentrate the virus and that is what manufacturers are doing. They are trying to find ways to concentrate what is in the diluted samples, so they can get as much of it bound to the plate and to enhance sensitivity of detection.
DR. NELSON: Dr. Maria Rios is going to talk about panel development and in-house testing.
Panel Development and In-House Testing
Maria Rios, Ph.D.
DR. RIOS: I will be talking about the panels for regulatory use and standardization of West Nile assay.
The need for panels for West Nile virus is quite obvious from what we heard this morning, there are multiple newly developed assays for blood screening and it is the concern of transmission of West Nile by blood transfusion.
However, there is a lack of reference reagents for the assay standardization in terms of viral quantification and sensitivity evaluation, and what was what we just heard about when Indira was questioned about how many copies and how to detect that.
There is also an acknowledged scarcity of human clinical specimens. The availability of animal specimens was made by CDC, and the most available was the flamingo strain from NY99, but the human isolates for evaluating these assays in standardization is still a need.
There is a lack of consensus in the viral load titer, as you have heard several times here, that the viral titer has been referred as plaque forming unit. Just to clarify what plaque forming unit is, it is now many plaques you get in one plate that is covered with the cell in the plates, are corresponding to viral activity lies in the cells and making empty holes in the plate.
So, we don't really know how many virus are required for forming a single unit or single plate. So, it is estimated that the viral particle in a PFU can be as broad as one infective particle to even 1,000 if they are localized in a single spot.
Therefore, there is a need for correlation between RNA copies, actual copies in the PFU unit. Because the non-infectious particles that is known to be present in flavivirus, and specifically in West Nile, those defective particles are not detected by PFU, but they are detected by PCR. So, the viral RNA is present, it is detectable, but it cannot be seen or identified in plaque forming unit.
Copy number determination is necessary for us to proceed with a definition of analytical sensitivity and therefore fulfill the regulatory requirements as we have been discussing about what is the real sensitivity, what is a PFU, what is a plaque number.
FDA and CDC have been working jointly towards the development of standards, and initially, CDC have made available the flamingo NY99 isolated in New York from a flamingo in 1999, which was titrated based on PFU units.
The lysate of flamingo 99 was made available to the companies that were developing assays. One has to bear in mind that, first of all, the units in PFU, when you do dilutions to try to tell your assay sensitivity, you are not really talking how many virion is there, how many RNA copies are there, but how many infective detection can be seen, therefore, you don't know exactly your RNA copy number.
So, as I said, CDC provided the isolate and then provided FDA with the NY99, the viral seed, so that we could grow, do titration, identify plaque units number, and try to correlate to RNA.
The CDC also provided to the FDA the first FFP unit that was implicated in the West Nile transmission by transfusion from which the co-component, the red cell, was proven to be the transmitter of West Nile infection. The virus was isolated and propagated, expended for manufacturing of panels.
These panels would be used with the purpose of lot release and for the licensure and post-market surveillance for NAT. We also plan on using panels for IgM tests.
The qualification panels for evaluation of relative sensitivities of investigational NAT and IgM assays are necessary for the post-market surveillance.
Both NY99 provided by CDC and the FDA-Hu2002 isolates are in the process of being characterized at molecular level. We are doing viral infectivity determination by performing PFU assays both at the FDA and in collaboration with the New York Department of Health.
We are further doing cytopathic effect and titration of the virus to know how many infective virus are there. In parallel to that, the same expended virus will be used to determine the RNA measurements.
By doing that, we would know the actual RNA copy numbers by measurements of the RNA by fluorescence and optical density determination with TaqMan compared to in vitro transcription RNA of known concentration. Then, we can calculate based on the size of the virus how many RNAs and therefore, how many viral copies are there.
Once this is done, we can correlate fairly accurately the PFU unit with the RNA copy numbers.
The final panel specification will be only established after we have done this and we have done these two procedures in collaboration with the companies, and they are performing these assays in other laboratories that have some excuse in doing that, and we will break down, correlate the results, and come to a common ground of correlation PFU and RNA titer.
For the IgM panel, we would like, FDA would like to have clinical specimens, and we urge here that you please share your specimens with us, between yourself and with us. We need clinical specimens that will contain varying titers of antibodies for West Nile and some members that are also NAT positive, as we heard here. Some are IgM and NAT positive. That will be needed for the validation.
The panel will be evaluated in collaborative studies, as well, using various candidate IgM assays.
Specifications for NAT and IgM panels has to be established based on results of collaborative studies.
Our plan is to have at least 100 pedigreed clinical specimens, and these 100 samples might include RNA positive only, IgM positive only, and dual RNA and IgM positive, and please help us to make this happen.
FDA also recommends that all reactive specimens identified in the IND clinical trials be made available to all manufacturers through sharing of samples, so that the assays can be cross-compared.
The sensitivity qualification panel, animal strain has been used, as I mentioned to you, the NY99 was available during early test development, but the desirable source material that it is the same for that one that we are testing, the human panel.
The fact that CDC provide FDA with the sample and the isolate will allow us to use human isolates for developing these panels.
FDA plans to formulate a panel for sensitivity evaluation or qualification of the assays using both the animal, because it was the original strain that was made available for the people developing the assay, as well as human strain.
The panel will include members with very high or high copy numbers, a broad range and very low, so that we now, if we are not quite clear about how high one viremic stage can be, we don't miss anything.
In summary, FDA is preparing lot release and qualification panels. Panels will be made available when adequately standardized, and FDA recommends testing of qualification panels when the investigational INDs are performed for all applicants.
DR. NELSON: Thank you, Maria.
Questions? Yes, Judy.
DR. LEW: I was just going to make a comment and maybe a question, as well.
When you are talking about comparing the PFUs to actual viral load, are you going to try to do that concurrently? Just because of the experience with HIV, we know freeze/thaw can really affect your RNA copies, so if you send it off separately, that may not work.
DR. RIOS: No, what has been done is that we expanded our most wanted mail off the viral isolate, so the same unit or the same volume of 500--actually, it was 750 mL--was pulled together and was shipped simultaneously in dry ice from the same stock.
I am working, at the FDA, the group is working with the same, as well as the New York Department of Health, was provided to NGI, Chiron, Gen-Probe--not Chiron, I have to send it to Chiron--Gen-Probe, Roche, and all the players in the game, so that we would have, each one of us have one type of measurement, and try to come to a common ground.
Then, with the PFU, we establish both FDA and New York State Department of Health, we can compare it, but it is the actual same isolate, it is the same bulk material.
DR. LEW: It may be the same bulk material, but I am just saying I know how things are shipped around, and if you do your PFU at FDA, and then you have to freeze it down and send it to various places--
DR. RIOS: No, it was frozen before. I asked them to report to me whether or not the sample came there in good condition, otherwise, don't use it. The only place that is performing PFU is the New York State Department of Health that has been doing this consistently, and we want at least two laboratories performing and correlating data.
DR. GALLARDA: A comment and a question, Maria. The copy number determination, in working with the NIBSE and W. Cho, on making international standards, I think John Soldana can attest to the difficulty of saying this is a copy because technologies have different ways of defining copy. So, it sounds like it is a work in progress, how to do this, is that correct?
DR. RIOS: Uh-huh.
DR. GALLARDA: On how you establish a commonly defined term copy number. I think you have already started the ball rolling, right, by sending out your initial panel to testing laboratories.
DR. RIOS: I can give you just what our approach has been there. Like I was explaining, the same bulk material was extracted, making sure that you only use RNA as carrier, and then all the measurements were taken in fluorescence measurement using fluorescence RiboGreen from molecular probe, and they were correlated.
Then, the concentration of RNA was translated into molecules per volume using a number based on the size of the virion, and try to correlate back. It is very hard to define it, to say what is one virion, but it is the best that we can come close to what we can call a unit.
That is why I would like to have from all of us. I am planning on talking to Soldana.
DR. GALLARDA: The question I had was OD-260s would be influenced by TRNA, which has aromatic compounds.
DR. RIOS: I am not using TRNA.
DR. GALLARDA: Oh, I thought you said there was--
DR. RIOS: No, a void.
DR. GALLARDA: Thanks.
DR. NELSON: Mike.
DR. BUSCH: Maria, just a couple comments on the PFU versus copies. What you are doing is using a tissue culture expanded strain. You know, with HIV, there has been a long-term controversy over the relationship between copies and TCID versus infectious units.
DR. RIOS: Absolutely.
DR. BUSCH: I think the same kind of work you are doing could be done as well with actual plasma from primary material rather than expanded tissue culture.
DR. RIOS: The difficulty, Mike, is that we don't have even a small volume to do. Imagine to have for the same people to get a handle on, I don't know, a couple of units or something to do that.
Also, when you have to put in the counter, your plasma, you are counting with a lot of factors. Not every plasma, even though it is positive, you can get a good isolate, and when you grow and you have a standard, then, you can work more homogeneously and getting normal material like natural.
DR. BUSCH: I think in the units that Sue has identified and that may be identified from some of the other studies, there should be large volumes of viremic units that could be subjected to parallel studies.
DR. RIOS: Sure, but this is ongoing, it is almost done. I needed to have the unit a month or two ago.
DR. BUSCH: The other comment, it is actually more important perhaps than copies per PFU, is copies per human infectious unit, and this is where I think the kind of studies we heard from Abbott in appropriate relevant animal models where we could do titrational endpoint inocula and correlate that with titers will give us the answer whether we really need ID-NAT.
DR. RIOS: Absolutely, but when you run a PCR and you say that it was, I don't know, 104 or 103 or 106, you are not knowing whether or not you are measuring defective particles in there, that may give you a higher signal, and then that is when the infectivity will play a role.
DR. EPSTEIN: I just want to comment that, you know, the big picture here is to establish a well-characterized reference material that can then be provided to a wide number of candidate manufacturers to enable meaningful comparisons, whether the copy to PFU ratio is or is not characteristic of a clinical sample is perhaps less important than having a single, well-characterized material.
DR. FITZPATRICK: I am just trying to correlate Indira's talk to yours. In the clinical specificity, she talked about investigational IgM test results could be confirmed using NAT and a second IgM, and then discordant IgM results evaluated using the plaque reduction assay.
From your talk and the discussion, I am a little concerned about the evaluation of the plaque reduction as--are you evaluating discordant result to find out if you have an infectious donor, or are you evaluating discordant result to see if they were infected with West Nile, and it doesn't sound like a plaque reduction or a plaque forming unit would tell us that.
DR. RIOS: I think you are looking at both. First, you are looking at the infectious unit when you do the plaque reduction or if the antibody was really specific or if you are getting some cross-reactivity.
I don't know, we have never looked at flu. I am sure that in the middle of next epidemic, when everybody is going crazy, somebody that has headache and fever, we will think they have West Nile. If we don't have a sensitivity that is such that we know that flu is not cross-reactive, somebody that goes somewhere, we know the world is a global village, so we need to use both parameters.
What I was referring mostly regarding constructing the panels, is to make sure that when we say we would like an analytical sensitivity or a qualification assay or validation assay, we can give a panel and say you test this and you give it to us, and we know where we are standing.
IgM is okay, it is confirmation whether it is specific or not, but again, we can't rule out a previous infection a year ago if we cannot identify positive by NAT test, and so on, and so forth.
I forgot to mention that we are developing house tests to cover most of the whole viral genome because most of the assays won't use the same viral genome, and we want to make sure that the panel we are generating we will be reliably detecting any of the genome segment of the virus.
DR. FITZPATRICK: I am not sure the PFU is going to help you evaluate a discordant IgM result, though.
DR. RIOS: Not a PFU, no.
DR. EPSTEIN: I think the confusing that is being engendered here is whether we are using a culture for virus titer or we are using a neutralization titer of antibody. What we are saying here is that if you have an investigational IgM assay and you get a positive result, how do we know it is a true positive.
The most definitive available test is the plaque reduction assay to confirm that is a West Nile specific IgM, so we see that as the ultimate test. The question is how available will that assay be, and short of that, what we are saying is that if you have a positive IgM result of an investigational test, corroboration that it is a true positive could be obtained by other IgM assays unrelated to the particular test that was used to screen and/or if there is a concordant NAT positivity because the odds of having, you know, a true positive if you have also a positive NAT, we would think are very low.
But we did hear the comment from Dr. Petersen earlier that we should be wary of attempting to confirm an IgM test with an IgM test, but the reason that we have put that forward, that's statistical confirmation rather than analytical confirmation, but in the back of our mind is that if there are several thousand putative positive IgMs that come up in the various studies, we may not be able to do plaque reduction assays on all of those samples, so what else can we do.
Is that clarifying for you?
DR. FITZPATRICK: That is helping.
DR. RIOS: Are you trying to comment about what Mike mentioned about using actual sample to do PFU for validation, because I think it is two different--one is a tool that you can use a reference to quantitate that unit. You don't have to do the PFU in the unit if you have a viral count or that you have standardized and you have a big volume and you know one PFU correspond to X, Y, or Z. Of course, we are going to talk about log, but it won't be 1 to 1,000.
DR. FITZPATRICK: That helps. Thanks, Maria.
DR. DAWSON: Just one more comment on the IgM alternate testing. I think that if you saw the data that Sue Stramer presented this morning, that perhaps 1 in 1,000 donations during peak season may be positive, and there is about a million donations per month, that would mean you would possibly get 1,000 RNA positives in September-October time frame, doing plaque reduction neutralization tests would be very, very difficult. In fact, my understanding of the epidemic last year, that CDC stopped doing plaque reduction neutralization tests because there were so many tests to run, they couldn't do it.
So, it is very time consuming, so I think that second IgM test will be of value especially if you go into a second host cell, which will not have the same type of cross-reactivities, and you would have to validate that second IgM assay with plaque reduction or some other suitable supplemental test.
DR. NELSON: Thanks. According to the schedule, next is lunch. So, presumably everybody has had lunch.
I think we will move to Alan Williams, who will talk about blood supply management and triggers for West Nile virus testing.
F. Blood Supply Management and Triggers for
West Nile Virus Testing - Alan Williams, Ph.D.
DR. WILLIAMS: Thank you again.
If tests with adequate sensitivity and high levels of predictive accuracy are available nationwide before the first human West Nile case occurs, we will be in pretty good shape in terms of preventing transfusion transmission of the agent, but if not, and, in fact, the first anniversary of the first 2002 human case comes in 87 days from now, we potentially need to think about contingency plans.
The focus of this talk will be to discuss interventions, both some different levels of testing, as well as other potential interventions, and triggers largely obtain from the surveillance data available through CDC, that might serve as start and stop triggers for other potential interventions.
I think it is important to mention that this is a discussion simply to put some of the options and some of the pros and cons on the table for discussion. While you can expect FDA to focus its thinking quite soon in order to give time for ramp-up for any potential interventions, this talk shouldn't be interpreted as FDA current thinking. I am simply putting items on the table for discussion.
First, I actually want to compliment CDC for their effort on addressing this epidemic. Through the ArboNet system, which involves both human and animal surveillance, as well as the intense post-transfusion study presented by Dr. Petersen earlier today, not to mention Lyle's uncanny timing in publishing a paper Modeling Post-Transfusion Transmission of West Nile that published a matter of months before the first case actually occurred, I think they have done a tremendous job.
In fact, we can expect that in the coming year, mosquito and animal data will be available, as well as human data regarding West Nile epidemic. However, as described earlier, the epizootic data is probably going to be of somewhat limited use for various reasons which Lyle detailed, and it is more likely that the human epidemic data will serve as a more useful trigger for some of the potential options.
Now, the West Nile virus meningoencephalitis data reported to CDC, our understanding is that is quite comprehensive and complete nationwide. West Nile fever data, on the other hand, does vary from community to community. Some areas of the country work up fever cases, get the testing done, and then they become categorized as West Nile fever cases, other areas of the country do not.
One consideration for defining a human trigger would be to use the combination of these two levels of reports and allow the fact that surveillance might be a little tighter in one area of the country versus another.
Now, when the voluntary market withdrawal was instituted for fresh frozen plasma and other frozen products last year, it utilized human West Nile case data at the state level. In fact, data is available at the county level, and potentially allows a little more fine tuning of use of some of the surveillance data. One could use individual counties or county plus a surrounding county to define a region.
As reported earlier, the transfusion transmission studies, the actual cases were recognized by CDC last year, and a close investigation of the known 14 cases has led to, albeit it limited data, on the viremia and the symptoms of known implicated donors.
Just an observation that West Nile is the first mosquito-borne TTD which is indigenous to the U.S., arguably, the first vector-borne if one doesn't consider babesiosis.
It has epidemic clustering both in time and geography, and I think in considering interventions, one needs to keep this in mind, it differs from some of the other types of post-transfusion infections that we have been dealing with, and it raises new issues regarding donor screening and blood testing.
NAT technology clearly paved the way for direct rapid testing of donors. The flexibility involved in NAT raised the potential for developing a test within a nine-month period, which is really quite incredible.
What are some of the unique challenges and questions related to West Nile? First, is related to testing, the biggie, geographic area and timing of the 2003 epizootic and epidemic is totally unknown at this point.
CDC feels I think that some of the early data will help to be predictive of the likely extent of the epidemic, but even the early data are not yet available.
Secondly, some aspects of West Nile virus NAT assay performance are unproven at this early stage. These involve not only sensitivity and specificity of the assays, but also scale-up capabilities, you know, will, in fact, all the equipment or software, et cetera, be available when it is needed even if the actual reagents are available.
Location of test sites, INDs are currently under development, will they be adequate to cover any sites having epidemic West Nile virus if that site happens to be covered by a different IND and there is problem with test availability or something.
So, all these things need to be considered to be reactive to cases as they occur.
I think it is probably a reasonable approach that implementation plans should be compatible with a worst case 2003 epidemic, and then we can really readjust as data tend to prove that otherwise.
There are considerations of minipool NAT sensitivity, which leads to the question are other additional protections needed to prevent transfusion transmission or minimize it.
With respect to donor screening, we heard this morning some limited data on the pre/post-donation donor symptoms, 3 of 14 implicated donors reported pre-donation fever plus or minus other symptoms.
The background prevalence of fever independently is 3.7 percent, but combined with some of the other symptoms, comes down a little bit to a little less than 2 percent, and potentially, some combination of symptoms might have good predictive value without undue loss of the normal donor population.
Rash and eye pain were mentioned this morning. In fact, I think we need to look at the data a little harder to see whether those additional symptoms were before the donation event or after the donation event, and exactly how they would fit in.
We need more precise donor loss estimates. This was from a controlled sample built into the design of the post-transfusion case investigations, but it would be very helpful if industry could, in fact, conduct some surveys of these symptoms for various time periods, one week plus or minus the donation event, two weeks or three weeks, et cetera, because the current data relate to three weeks, and are at a limited number of samples.
A large point is that there are persistent blood supply problems, and these increasingly point toward a quite fragile donor base, so whether we are talking about specificity of an assay or a screening of a donor, all of these will impact the current blood supply.
We do need to keep in mind the donor loss in this circumstance would be "temporary," probably related to 28 days post-initial recognition, but even temporary loss of a donor has its own inherent problems because once a donor is deferred, sometimes they tend not to come back as readily, so that has its own implications.
Post-donation information reported in 5 of 14 of the implicated donors you heard this morning are reported post-donation. Again, the background prevalence would be about the same, but we know that post-donation reporting tends to be inefficient and the products may well have been transfused by that point.
Another potential intervention--and this is one that I think needs to be considered very cautiously--would be real-time suspension of blood collection or blackout in effect of blood collection in a geographic area experiencing human West Nile cases.
Now, this was retrospectively sort of the approach taken in 2002, when there was voluntary market withdrawal of fresh frozen plasma and other frozen products in parallel with human West Nile virus cases defined at the state level plus or minus a week, I believe.
The potential for transmission would arguably be reduced by this approach because the epidemic tends to have microfoci, so that if one simply stopped collecting blood in an area that appeared to define the epidemic, one could quite possibly intervene in potential transmission cases.
However, a couple big areas of concern with the blackout of collection, first, is that if you establish that as a precedent, it may, in fact, not be sustainable if the epidemic worsens, and you may, in fact, have to change standards midstream to maintain an adequate blood supply.
Even if done at a local level, there is a potential for severe strain on supply.
So, those are some of the potential interventions. What are some of the potential triggers? Some of the obvious ones, West Nile cases in animals, the epizootic, West Nile cases in humans, the epidemic, are these going to be defined by geographic area, and available would be state, county, and some combination of county data.
One could also perhaps envision a situation where a large county, for instance, Cook County in Illinois, shouldn't be treated the same as a very small county in another state. One could potentially tie a case rate per 100,000 population to a trigger.
The actual scope of the regional and national epidemic also would be a factor in defining the state of the epidemic.
Another factor, potential trigger, would be a defined end of human cases and a defined end of the epizootic.
Additional triggers. Test availability and effectiveness. Is the test available in very limited supply or is it available and able to be implemented nationwide, and what is the effectiveness, the sensitivity of the test.
Is there evidence of local or regional blood shortage? Clearly, it is always a balance between implementing measures to protect from a safety threat versus introducing potential compromised safety by having inadequate blood supply, so that needs to be a trigger, as well.
In fact, changes in the NAT testing prevalence itself could be of value particularly in signaling the end of epidemic in a certain area.
So, what I have done in the balance of the talk is simply combine some of these factors and again just for discussion purposes, put out some potential interventions and their triggers.
To begin with, considering NAT testing, what would be the start triggers for NAT testing? First of all, say that NAT testing, high-quality NAT tests are available under IND in unlimited availability. One would consider that the most appropriate response would be U.S. testing, universal testing as soon as possible.
Now, say the test is available only in limited availability. The potential trigger would be the epizootic. I know Lyle mentioned this is of limited usefulness, but I think one thing it does is it signals the potential for human exposure and is available for a period of time, roughly a month, as I recall, prior to the first human cases being developed. So, it gives some sign to put something in place before a human case actually occurs.
Say a test has very limited availability, only a single lab has a test, and one needs to institute some form of protection, then, potential of the human cases could serve as a trigger, but by then you already have transmission in place and a little late in the game.
What would be the triggers for stopping NAT testing? In an unlimited availability situation, I think this is something that needs further discussion. One can always make the argument that if you have a geographically or temporally focused infection going on, that even if it is not going on in a certain area of the country, a donor can always get on the plane from somewhere else in the country and attempt to donate in that area.
As I said, this needs discussion, but perhaps the West Nile NAT prevalence itself could be a driver as to whether NAT testing should be continued once the season is defined as being over.
In a situation where the NAT test has only limited availability, a potential trigger for that would be a defined end to the epizootic, and this could be defined at the county or similar level, and under very limited availability, similarly, defined end to the human cases.
What are some other potential intervention triggers? Donor screening for West Nile symptoms and product quarantine and retrieval based on post-donation information.
There would be a couple way to potentially implement this. One would be the typical direct screening of donors, which in many centers now is done by self-administration of the donor form, would involve modifications to the form, which takes some time.
It could also be done by adequate upfront information to potential donors and inclusion in an information sheet. In fact, most of the donor deferral occurs before the donor ever comes to the blood center, so I think there should be consideration of that, as well, if this is planned.
What would be a trigger to start donor screening? I think here would be another situation where a start of a defined epizootic in an area would be appropriate, but in the case of supply concerns, potentially, one would consider the human cases. In either situation, one would need to carefully monitor supply.
Donor screening, when would be best for it to stop? Similarly, West Nile epidemic or potentially human cases, probably the epizootic would be somewhat more protective.
Cessation of blood collection in West Nile endemic areas, the blackout concept. One consideration would be that such a strategy might only be appropriate to start if an epidemic is highly localized and there is no test available. That would be one situation where it might be arguably an appropriate response.
One would obviously need to carefully monitor supply and the start of such an intervention could be defined by human cases at a carefully defined geographic level. Stop of that approach would be end of the human cases at again the same geographic level or availability of a West Nile sensitive test.
In summary, the uncertainty of the timing and the scope of the 2003 West Nile epidemic, if there is one, requires planning for several simultaneous interventions to maintain a safe blood supply.
I am actually hoping that there will be adequate mosquito control, so that we will see a lot less than we had last year, but again that is another variable which is unproven at this point.
Pre-definition of intervention triggers will assist the planning process, and we recognize that blood collection facilities need time to put in place whatever is recommended.
Planning will develop more focus as the 2003 surveillance information becomes available, and FDA very much seeks comments from the committee regarding the interventions and the trigger concepts presented.
DR. NELSON: Thanks, Alan.
DR. SCHMIDT: We heard before that the highest levels of HHS had interest in maintaining the blood supply. It would seem that before the media starts in, that there really should be proactive action.
The public believes you, they believe the CDC, but this might perhaps have to be kicked upstairs to the Assistant Secretary of Health, if there is one, or the Surgeon General, and to lay it on the line and say what we know, what we don't know, what might happen, say it in advance.
It is sort of like--I hope this doesn't kick down the idea--but it is sort of like Code Orange, if it doesn't happen, then, you take credit for it, but the point is you lay it out in advance, before the panic starts.
DR. KLEIN: I really think that this is really very creative thinking and it's the right time to do it, and I really applaud your doing that. I do want to make a comment, though, about donor history.
It is a comment that relates to some of the past donor questions that have been imposed upon the blood collectors. We spend a lot of time, a lot of money, and a lot of science validating assays, as we heard today.
We spend no time, no money, and no science validating donor history questions, many of which have almost no sensitivity and even less specificity. They do a great deal of damage.
So, while I think the idea is actually a good one, before anyone comes up with a question that is then published and imposed upon the blood collectors, I think we ought to do some kind of validation of the donor history, so that we can be sure it does something to enhance blood safety, and not something to detract from blood supply.
DR. NELSON: Part of the problem, of course, is that we don't currently have a West Nile epidemic, so we could tell how many false positives there were, but we couldn't measure the true positives.
DR. KLEIN: I am well aware of that. We don't have a variant CJD epidemic either. One could argue, I suppose, that having imposed a series of questions, we haven't seen a single case, so it must be working.
I don't mean to be facetious, I really would caution us, because we have made mistakes in the past that I think have been costly, not so much in terms of dollars, but in terms of donors, and I think we need to think very carefully so that we don't make such errors again.
DR. ALLEN: I agree with Dr. Klein. This is a very intriguing presentation, and I thank you and the FDA for it. It raises, in my mind, some real serious considerations that have to be looked at, and those aren't our questions to be considered this afternoon, so I will be brief, but let me just mention a few.
This issue of clinical illness and symptoms, I think that does need to be looked at very carefully. My own personal standard before I donate is I am totally healthy for at least a week. Even if I have had a respiratory infection or something, and I get called, you know, two or three times in a row, please come in and donate, I don't do it.
One of the questions I routinely get asked is do you feel well today, and I think that is an appropriate question. It might be extended somewhat. I think something could be done here without changing things too dramatically, that might help increase deferral for relatively minor symptoms either in the pre- or post-donation period. I think that definitely needs some consideration.
The issue of geographic suspension of collection, I am really concerned about that one. I think it is going to be difficult to explain to the donor population. I think that potentially that is one that when you try to reinstitute it in an area, you are going to find that donors may not be coming back in, having been turned away for something that they ill understand.
I think it is going to play havoc with the local collection systems, you know, who may have to shut down perhaps, depending on the criteria, for weeks at a time. Given what we heard, that even in the high incidence areas, what was it, 2 percent of the population may have been infected in a given season, I think this is something that we have to look at very, very carefully before taking any recommendation on.
Also, the issue of starting and stopping testing. I, because of cost and a whole host of other considerations, don't want to see any more testing done than absolutely essential. It increases the probability of errors, it increases costs, time, the necessity to purchase and maintain expensive equipment and all the rest of it.
Nonetheless, we have long established in our country that blood is collected under a single standard, and that was obviously extremely important during the HIV epidemic. We didn't establish local standards here and there. We didn't establish a minimum set of standards. We established a national standard.
I think that there is a lot of reasons why that concept needs to be maintained, or if it is going to be changed, needs to be looked at very carefully.
I would like to hear from Roger Dodd, I don't know if Sally Caglioti is in the audience, but what is the impact in a laboratory of starting and stopping testing at certain periods of time in the year rather than having here is a new standard and we are doing it on every unit that goes through and is going to be verified as acceptable for release.
What about maintenance of equipment that isn't used for a period of time, you know, is left sitting idle and then may need to be cranked up on short notice if something happens?
I think these are issues that are worthy of exploration, but I think we have to look very carefully at all of the implications before making any decisions.
DR. STRONG: I would just like to amplify those comments since I have some operational responsibilities at a blood center. Starting and stopping anything, whether it be a questionnaire or donor testing, is a major undertaking. We have to write SOPs, we have to go through validation exercises, we have to train staff.
All of those things take time, and it could easily be that it would take you the period of time that you were stopping to stop, and as long to start again, that you wouldn't have accomplished anything.
So, from a very practical point of view, the starting and stopping really doesn't seem even feasible.
In terms of stopping collections, if we are actually able to hit this timeline and starting in July, of course, we are starting in the summertime. The summertime is notoriously one of our most difficult time with keeping up with collections.
I just can't even imagine that anyplace in the country could stop collections and make up the deficits no matter how small. We struggle every summer to get through that. I also commend you for coming up with some optional scenarios. These don't seem very practical and it is going to be very difficult for us to implement either of those two options.
DR. FITZPATRICK: It's great that we brought these up and I echo Harvey's comments, but one thing is unfortunate that on the agenda, you have asked for discussion, but there isn't time for discussion.
Second, there won't be another BPAC until the end of June, and by then it will be too late to have your mechanisms in place, so you are going to have to do this without much discussion.
You have already established a trigger and that is 1 July. You have informed the industry and the testing, the manufacturers, that by 1 July we are supposed to be testing the entire blood supply. So, I think you have already established a trigger to start testing and set a date. So, I am a little confused by that.
You have said that it is 82 days until the first reported case from last year, and we don't have a strategy, and I am concerned about that. I am concerned about the fact of there is two impacts to stopping collection in a geographic region. One is the obvious impact of the blood supply in that region, how do you maintain it, how do you get donors, what is the impact on the donors.
The other is a personal impact, what do you do with that staff that has been idled, and how long are they to be idled, and what is that collection center to do with their idled staff? Do they pay them? With what? Is there any mechanism available to them to keep that staff on for an unknown period of time? I think that is a very unworkable idea.
Starting and stopping testing, Dr. Strong I think addressed that very eloquently. One thing that I don't see that was discussed previously is that when this started, we identified a set of patients who were at risk. There were a couple of comments on others about labeling the blood as tested or untested, or as West Nile positive or West Nile negative, and using clinical judgment to provide the blood to the appropriate patient. I think that is a possible strategy that could be used in those areas where there is a geographic solution or problem.
If there is a trigger point of human cases being reported and you have a geographic region where you think the blood supply is at risk, and we don't have enough tests, then, yes, I think you should probably divert tests to that region to test, but we have those problems that Dr. Strong mentioned, if they aren't testing already, then, they have to be trained, they have to be validated, they have to be assessed, and that all takes time, and quite frankly, between now and July 1st is probably not enough time to even do what we are being asked to do by July 1st.
So, unfortunately, there aren't a lot of solutions, we have some conundrums here. The impact is unknown, the risk is known, and the interventions are few. So, I think it might be best to maybe have some special session to look at what can be done, because the public is aware from last year that there is a problem, and we are aware that patients will be at risk, so I think some interventions need to be come up with, and to me, besides those you have mentioned, the one of identifying the at-risk population and trying to protect them may be the most quickly and easily put in place even though it may not be the most effective, it will at least be partially effective.
DR. SCHMIDT: Unlike you all, Florida has had a very warm, wet spring, at least south of Gainesville, and the talk is already on that things are going to start early this year.
DR. NELSON: On that optimistic note, I would like to maybe move on and we will have some more discussion in a little bit.
The next topic is testing of source plasma and the clearance of West Nile virus in plasma-derived products. First, is Mahmood Farshid from FDA.
G. Testing Source Plasma Donations and Clearance
of West Nile Virus in Plasma-Derived Products
Mahmood Farshid, Ph.D.
DR. FARSHID: My brief presentation is with two questions or two issues with regard to West Nile and plasma-derived products, which they are noted here.
The first one is whether viral validation data on model flaviviruses are sufficient to demonstrate West Nile virus clearance.
Basically, the question is the current validation data, which is done on model viruses, is sufficient or the specific validation needs to be done.
The second is whether the clearance steps in the manufacture of plasma derivatives obviate the need to screen donations of source plasma.
I will address the first question first and try to present some pros and cons.
The question which is frequently asked, are current viral clearance steps effective against West Nile virus, and the clearance level predicted for West Nile is comparable to those obtained with model flaviviruses based upon taxonomical and physicochemical similarities between the viruses, as we heard in presentation today, it is an envelope virus, and it is a flavivirus.
There are some data presented in the last workshop, and we will hear more today which indicate that these predictions are confirmed more or less based on this data which presented publicly.
Also, it needs to be noted that no transmission of West Nile by plasma-derived product has been reported.
However, to provide further assurance on the robustness and reproducibility of West Nile clearance in various plasma-derived products, specific viral validation studies should be conducted using West Nile virus.
The rationale for further validation studies are indicated here. The clearance results vary depending upon the virus and the product being tested. Basically, for plasma derivative, viral validation studies must be product and process-specific.
That is to say that generic validation is not applicable to human-derived products or products that may be contaminated with human pathogens.
So, the key word here is the specificity of the validation and its relevance of a particular process and product.
This is simply as example to underscore the need for the specificity in terms of validation studies, which shows the different manufacturing conditions could substantially influence the clearance capacity of a given inactivation step.
Here is 6 different manufacturers and activated hepatitis A and albumin by Pasteurization. As we can see, there are huge differences in the level of the clearance as low as 2.9 for the concentration of 4 or 5, and as high as 6.6, the same thing for the concentration of 20 to 25 percent, as low as 2.5 and 6.6 log being reported.
This slide basically here, the intention is not to outline what is required for validation, but again to underscore the importance of having a specific validation for a given process or product. We put some of the parameter which is required for validating the small or laboratory scale model, and to establish the relevance to the actual manufacturing process.
One need to include all the critical process parameters within the clearance steps performed at the fullest scale, and adopt actual production methods as far as possible. Assess the impact of unavoidable differences, and maintain relative values, such as volume and geometries, and also absolute values, such as time and temperature.
Accurately transfer the process from laboratory scale to full scale.