|
U.S. DEPT. OF HEALTH AND HUMAN SERVICES
PUBLIC HEALTH SERVICE
FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
OFFICE OF BLOOD RESEARCH AND REVIEW
NUCLEIC ACID TESTING FOR HCV AND OTHER VIRUSES IN BLOOD DONORS WORKSHOP
WEDNESDAY, SEPTEMBER 16, 1998
The workshop took place in Conference Rooms D and E, Parklawn Building, 5600 Fishers Lane, Rockville, MD 20857 at 3:30 a.m., Edward Tabor, M.D., Chairperson, presiding.
PRESENT:
EDWARD TABOR, MD Chairperson
JAY S. EPSTEIN, MD Director OBRR
HARVEY J. ALTER, MD Speaker
MIRIAM ALTER, PhD Speaker
CELSO BIANCO, MD Speaker
JENS BUKH, MD Speaker
ROGER DODD, PhD Speaker
INDIRA HEWLETT, PhD Speaker
STEVEN KLEINMAN, MD Speaker
NICO LELIE, PhD Speaker
MICHA NUBLING, PhD Speaker
BETTY ROBERTSON, PhD Speaker
BUY RAUTMANN, PhD Speaker
JOHN SALDANHA, PhD Speaker
LEONARD SEEFF, MD Speaker
PETER SIMMONDS, MD, PhD Speaker
MEI-YING W. YU, PhD Speaker
INDUSTRY PRESENTERS:
ANDREW CONRAD, PhD National Genetics Institute
CHUCK FRISBIE Alpha Therapeutic Corporation
SUSAN CUSHING, MD Baxter Healthcare Corporation
ELIZABETH DRAGON, PhD Roche Molecular Systems, Inc.
BARBARA MASECAR Bayer Corporation
CHARLES WATSON, PhD Centeon L.L.C.
CHRISTINA GIACHETTI, PhD Gen-Probe
SUSAN STRAMER, PhD Amer. Red Cross
GERALD ZERLAUTH, PhD Immuno AG
RICHARD FLANDERS, PhD Abbott Labs.
MILTON PSALLIDOPOULOS, PhD North American Biologicals, Inc.
CONTENTS
Opening Remarks
Overview and Epidemiology of HCV
Prevalence of HCV and Other Viruses in Blood Donors
Nucleic Acid Detection of HCV and Other Viruses in Blood Donors
Genetic Diversity of HCV
Application of PCR to Molecular Epidemiology of HCV
Discussion
Industry Presentations: Test Methodologies and Experience with Nucleic Acid Testing of Plasma Pools
National Genetics Institute
Alpha
Baxter
Roche
Bayer
Centeon
Gen-Probe
ARC
Immuno AG
Abbott Laboratories
Nabi
Regulatory Issues in Assay Validation and Standardization Assay Validation: Regulatory Perspective
U.S. FDA
PEI
EU
Standardization of HCV RNA Assays
NIBSC
CLB
CBER
PEI
Discussions
Issues in Implementation of Nucleic Acid Testing for HCV and Other Viruses
Panel Discussion
Closing Remarks
PROCEEDINGS
(8:35 a.m.)
DR. EPSTEIN: Good morning. I know we still have people queued up in the hall trying to get registered and get their packet, but in the interest of having an on-time agenda I'd like to open the meeting.
I'm Jay Epstein, Director of the Office of Blood Research and Review in CBER, and it's my pleasure to welcome you to this FDA-sponsored workshop on Nucleic Acid Testing for HCV and Other Viruses in Blood Donors.
First I'd like to congratulate all of you for getting through the building security and for finding the conference room. I suspect that the rest of the day's activities will seem easy by comparison.
As you know from the published agenda, this workshop has been convened to explore the current state of the technology and implementation of nucleic acid testing for screening of blood donors, particularly for hepatitis C, and to assist FDA in making determinations that will influence U.S. regulatory policy on such testing.
FDA has been focused on promoting the development of gene-based testing of blood donors ever since the discovery of gene amplification methods.
Although blood products have become quite safe with the progressive introduction of multiple donor screening and testing methods over the last two-and-a-half decades, the effort to introduce gene-based testing has taken on a sense of urgency since the recognition several years ago that such testing could improve blood safety by substantially shortening the window period of infectivity prior to serologic detection in donors acutely infected by a number of transfusion transmissible agents.
This point was emphasized at an FDA scientific workshop four years ago in September 1994, which was on the feasibility of genetic technology, to close the HIV window in donor screening. Proceedings of that workshop were published in the March 1997 issue of Transfusion.
Since then FDA has approved investigational studies of gene amplification of donor samples using mini-pool protocols for several agents. And let me remark parenthetically that if any of you has not heard what a mini-pool is you're probably at the wrong meeting -- although you'll certainly hear it by the end of the day.
FDA also recently published a guidance in the Federal Register on July 10th of this year on the manufacture and clinical evaluation of in vitro tests to detect nucleic acid sequences of HIV-1. And this is a paradigm then, for us putting forward the review and approval criteria which we then think will provide the framework for product approvals in this area.
Gene-based testing of individual donations remains an ultimate goal pending further technology development, and we are aware that the NHLBI has sponsored some of the technology development in that area.
Now at this workshop this morning we will explore the current state-of-the-art regarding the epidemiology of hepatitis C and the sensitivity, specificity, and reproducibility of nucleic acid tests for hepatitis C and other viruses of concern in donor screening.
We will also hear about the clinical experience with nucleic acid base testing, including the prevalence of detection in donors and the issue of genetic variation for HCV. In the afternoon attention will shift to implementation issues and questions pertinent to standardization and regulatory control.
We anticipate and we encourage discussion of a large number of practical issues, such as the requirements for test sensitivity, possible limitations to the size of many pools, the definition of a reactive and a confirmed screening test result, strategies for inventory control, and look-back, and the prospect for replacing some existing tests with nucleic acid-based tests.
We know that these questions are on many people's minds. I would ask however, that our meeting participants accord the FDA representatives the opportunity, both to express their personal opinions as well as when appropriate, to summarize FDA's current thinking, while recognizing that the official policy statements will be developed and published in good time based on thoughtful reflection.
At this point I would like to thank the members of the organizing committee who all have worked very hard to bring about this meeting. I would particularly like to acknowledge both the scientific and the regulatory contributions of Dr. Indira Hewlett who has been FDA's leader in this area and who not surprisingly, is also a chairperson of this meeting.
Additionally, I would like to note the high level of international interest and cooperation that has characterized the development of methods and standards in this field, and to thank our international participants for making the effort to join us today.
Now, I hope that I've succeeded in conveying to you a sense of FDA's purpose in convening this workshop. I'm confident that the sharing of information and opinion will accelerate our effort to establish nucleic acid-based donor screening for hepatitis C and other agents.
So now let me turn over the meeting to Dr. Edward Tabor, who's Director of our Division of Transfusion Transmitted Diseases, who will introduce our first, distinguished speaker.
CHAIRPERSON TABOR: Good morning. We'd like to ask that each of the speakers, when they've completed their talks, please sit at the central tables in front of the podium.
The first speaker is Dr. Harvey Alter from the Department of Transfusion Medicine at the National Institutes of Health.
Return to Table of Contents
DR. HARVEY ALTER: Decisions about nucleic acid testing are really very easy -- of HCV -- are really very easy, and I think we're on that course and there will be no question about it. Hard decisions are really going to come up, for instance, the BPAC meetings in the next few days -- things about CJ disease.
For instance, the FDA is proposing that we add this new question to the donor screening process. This question will ask, have you ever, even once in your life, had intimate contact with the brain or spinal cord of a cow? This is going to be difficult to decide on.
My job today is to give you background on hepatitis C. These are data from the CDC, from Miriam Alter's work. In acute hepatitis in the U.S. only about 15 percent of cases are hepatitis C-related. So it's not a big player; doesn't cause a lot of acute, severe hepatitis.
The problem with hepatitis C is that it is the primary cause of chronic hepatitis in the United States. And that's because about 85 percent of the people who develop hepatitis C infection become persistently infected; although 15 percent do resolve, as we'll discuss later on. And that's an important distinction from other persistent infections such as HIV.
But the real question is, why does such a large number have persistent infection? And we don't really know that answer, but one of the possibilities is the so-called, quasi-species nature of the hepatitis C virus that Dr. Bukh is going to talk about in detail later.
And that is that the virus can exist simultaneously as a series of closely-related but immunologically-distinct variants. And this is a real case -- our famous Hutchinson patient -- in which Patricia Farci cloned in sequence, 105 clones from this patient during acute phase of hepatitis.
At that time he had a predominant strain, which is 57 percent of the clones with the same, but in addition there were all these other clones. A total of 20 different variants present in his serum at a single point in time.
So that even if this major strain were contained by an appropriate, neutralizing antibody response, any one of these other strains would be ready to emerge as the predominant strain and escape the immune attack.
This is some more interesting work from Dr. Farci. We've looked -- from our post-transfusion hepatitis studies where we have serial, acute phase samples -- we've looked at the changes, the viral variation during the acute phase of hepatitis C infection.
I think this is the only study that has looked at it early on -- infection -- and samples were taken here just after the first PCR of positivity; here at the first ALT elevation; here just at the rising ALT; and here at the peak of ALT. So all within 16 weeks.
There were ten clones at each point that were sequenced. And you can see here at week 3, early in the infection, there were three variants which I've labeled A, B, and C. By week-8 none of those were still detectable, but now D was the predominant strain but there was also E, F, and G.
By week-13 one could detect A again, so it was still around, but now there was H, I, J, K, L, M, N, O, P, and by week-16 G, L -- all the way through U. So an entire -- in 16 weeks an entire alphabet of hepatitis C viral strains. And any one of these would be capable of becoming the predominant strain during an immune attack.
So I think this is one of the major reasons for this being a persistent infection, and HIV has a similar occurrence.
Now, I'm going to pre-empt my talk by showing this overview to say that there are really three major forms of hepatitis C infection. Shown here in blue are the ALT elevations; in purplish color the RNA levels; and in yellow, the antibody levels.
And what we know is that now, at least from our studies, that 15 percent of the people recover; that is, they are persistently antibody-positive and persistently PCR-negative over many years. And their ALT is normalized.
And another 15 percent go on to rather severe disease within the first six to 10 years of their infection and die from hepatitis C within that timeframe. Those people have chronic ALT elevations, persistent HCV RNA, and persistent antibody.
And then there's the majority of people which I think is about 70 percent, who have this very benign, longstanding infection that we'll talk about throughout this talk, where again they're persistently PCR-positive, persistently antibody positive, and have these very low-level ALT elevations that are intermittent.
And you can see in all these scenarios that HCV RNA appears very early in the infection in weeks 1, 2, or 3 post-exposure, and antibody doesn't appear until week-12 or so. So this is the window that this whole meeting is about; the period from here to here. I don't have my other window slide but I'm sure somebody else will show it today.
But it's this nine weeks or so -- 9 to 12 weeks -- that HCV RNA testing will possibly interdict and virtually prevent all transfusion-associated HCV.
Now, the clinical cause of hepatitis C I've alluded to. This is the famous single case from Kendo Kiyosawa in Japan, but it shows what can happen in the worst-case scenario. And that is, one gets a blood transfusion, here's acute hepatitis that was biopsy-proven so there was no underlying chronic liver disease -- this was acute hepatitis.
A year later this patient has chronic, persistent hepatitis, a mild form of chronic liver disease; four years later chronic active hepatitis; three years later briding necrosis, which is a pre-cirrhotic region; two years later cirrhosis; and three years after that, hepatocellular carcinoma leading to death.
So within 18 years this patient went from blood transfusion to death. There's no doubt that this sequence can occur. I'm going to paint the other side of the picture as well but I don't dispute this occurrence; it happens. And hepatitis C is a leading indicator for liver transplantation here in the U.S. and throughout the world.
In our studies in fact, although most of the people on initial biopsy are very mild to moderate disease and even on follow-up biopsy either improved or had stable histology, about a quarter of the patients progressed over time. And we wound up with 20 percent of our patients having cirrhosis, and that 20 percent figure, histologic cirrhosis in about 20 percent of HCV patients comes up from study to study.
And in that group who had cirrhosis, three of them died of liver failure, three of them had very severe liver disease but died of their underlying heart disease, but one had cirrhosis for ten years and was still very well compensated and died of something else. And we have another patient now has cirrhosis for 22 years and is very well compensated. So you have this spectrum even among those who have cirrhosis.
This is another picture of the bad side of this infection. And these are data from Myron Tong who works in a tertiary care center so he's getting the worst cases referred to him. But in his population, of those who are referred for hepatitis C, 44 percent had chronic hepatitis but 46 percent of his population had cirrhosis and 11 percent had hepatocellular carcinoma.
Part of that was because he had a large Asian population who seemed to be more prone to hepatocellular carcinoma. So they had a disproportionate amount of carcinoma among the Asians. But nonetheless, this overall 11 percent carcinoma and 46 percent cirrhosis is about the worst numbers that I have seen, and I think it reflects the severity of the cases when they're referred to him and the Asian population.
But now the other side of the coin. Perhaps the -- and I think almost certainly, the bigger number. These are the summary data from the multi-center study headed by Leonard Seeff. You've all seen these before and Leonard is here in the audience. But I think it's a critical study, and I say it with some bias because we're participating in it.
But it was very well controlled. And shown here in white are the cases, and in orange and green the two control groups -- very carefully matched control groups.
And now after 22 years, when one compares the overall mortality among the patients who had hepatitis which turned into -- the most would be hepatitis C in the 1970s, compared to the controls who were transfused but did not develop hepatitis, the overall mortality is identical. After about now, I think 22 years in the latest analysis.
There was a slight increase in liver-related mortality among the hepatitis C case, but it was very slight; about one-and-a-half percent increase over the controls. So this is a fairly benign picture over at least the first two decades of this infection.
And we now have studied -- not a very good slide and I don't know if it can be cleared up -- but these are data from the Irish study of women who received contaminated lots of Rhogam. They were followed up for 17 years at this time. It's a little bit longer now.
But just the bottom line of this is when you look for fibrosis in these patients 17 years later, 57 percent had no fibrosis and only two percent had cirrhosis -- I'm sorry, two percent had cirrhosis and two percent had severe hepatitis. So maybe wind up with four percent with cirrhosis in the first two decades.
These were young women when they were infected, but I think again, this says the first two decades for most people are relatively benign.
And here's our own data from our hepatitis C positive donors that we now have in long-term follow-up. These studies were dealing with Jay Hoofnagel's group and one can see that these are liver biopsies.
And the vast majority of these patients who are being biopsied on an average of 18 years after the onset of their infection -- and we know that from when they took drugs or had a transfusion -- have very mild, histologic lesions. And only this 13 percent have either cirrhosis or severe, chronic hepatitis that might lead to cirrhosis. So we're getting the same picture from all of these studies.
Poynard has done a very nice study to look at the progression of cirrhosis to -- the progression of fibrosis to cirrhosis in a large, multi-center study conducted in Europe. And basically he developed a fibrosis unit that I won't go into, but essentially one unit equals one stage.
And so it takes about seven-and-a-half years to go from stage zero, which is no fibrosis, to slight fibrosis, 15 years to go to more advanced fibrosis, another seven-and-a-half years -- or 22 years in total -- to get to a lot of fibrosis, and 30 years to get to cirrhosis.
So this is a slowly evolving disease in those who do develop cirrhosis and maybe a lot of these patients who are doing well in 20 years are going to suddenly do badly, but I don't think so.
And here's another interesting study by Fattovich which shows that even when you get cirrhosis with hepatitis C, you can stay compensated for a very long time. And here's the 5-year survival in people who have cirrhosis -- is 91 percent. And the 10-year survival probability is 79 percent, after you have documented cirrhosis. So this is a very, very slowly evolving process.
Now what makes some people go on to cirrhosis and other people not, or what makes some people have very severe cirrhosis and others lesser? Well, we don't know all the co-factors but one of them clearly, is alcohol.
And this is an interesting study by Correo in which they compared teetotalers and alcohol abusers -- defined by 175 grams of alcohol per day. And two major findings from this study.
One, if you are HCV-negative, even if you were an alcohol abuser, the relative risk of developing cirrhosis was 15 compared to the reference of a person who was HCV-negative and a teetotaler.
But if you were HCV-positive -- if you had hepatitis C plus alcohol -- the odds ration went up to 147. You greatly increased the risk of developing cirrhosis. And that's been a consistent finding from study to study.
Well, a long time ago I developed this guess as to -- I developed this sort of an estimate of what would happen if you took 100 people who had hepatitis C, virus positive, and what I've already shown you is that 15 percent resolve the infection so you're left with 85 percent, 85 people who develop chronic infection.
Of those, 20 percent will develop cirrhosis and 80 percent will have a stable disease for a very long time, or will die of a non-liver-related death. So of the 17 patients now who have cirrhosis, let's say if even 75 percent of them have a mortal cirrhosis before they die of something else, the total number of people who die of their hepatitis C would be only 13 percent within the first two decades of their infection.
Now ultimately, some of these stable people will move over here into the mortal group, but most of them will live out their normal lifespan or will die of whatever they were going to die of in the first case. So less than 15 percent mortality was my guess based on our own studies and some other preliminary studies.
Well, the Seeff study has now advanced to looking at morbidity, and Leonard has given me permission to show this slide in advance of publication -- of what we hope will be publication. And it's a complicated slide so I'm just going to focus on a very brief area.
In 1974 there were 103 people who were hepatitis C-positive. Most of these developed their hepatitis C as a result of blood transfusion but there were a segment of these who were already HCV-positive when they entered the study.
But what I want to focus on is what happened to these 103 people -- almost the same number as in my schematic. Well, what happened was, there was just like the 15 percent I predicted that were anti-HCV-positive, HCV RNA-negative that appeared to have recovered.
But here's a group that was unanticipated: ten percent of these patients not only were now RNA-negative but they were also antibody-negative. They had no evidence of their prior HCV infection. So the natural recovery rate could be as high as in the recovery range of 25 percent.
But for this purpose I want to focus on this 73 percent who are persistently antibody-positive and persistently RNA-positive. Only half of them -- or actually 40 percent of them -- had biochemical evidence -- 40 percent of the total population had biochemical evidence of chronic hepatitis. So 60 percent had normal ALT levels on at least two samples.
Some of these in this group who had chronic hepatitis and were HCV RNA-positive were biopsied, and based on those biopsies we made some extrapolations: that 30 percent of the people in this group would develop cirrhosis; that less than five percent in the group who had no biochemical evidence of hepatitis but were RNA-positive would develop cirrhosis; and less than one percent of these who recovered would develop cirrhosis -- and it really would probably be zero.
If you add all those numbers together you come up with a cirrhosis figure again, of less than 15 percent in the first 20 years.
So the question then is what's going to happen after the first 20 years? If this is where we are now, at around 15 to 17 percent, will the curve continue such that even after 60 years no more than 50 percent of people will develop cirrhosis? Or will the curve for some reason accelerate so that by 60 years everybody will have cirrhosis? Or even by 40 years everybody will have cirrhosis.
But there's no reason to suspect the sudden acceleration. And in fact, the curve may actually asymptote here. I didn't show that but it's possible that people who haven't developed cirrhosis within 20 years may never develop cirrhosis. They may be selected out as being those who are not going to progress. So I think the curve will be here or lower rather than anything higher.
So when you look at hepatitis C it's like the blind man looking at the elephant. It depends on where you see it; from what angle you see it as what you think it is.
If you're a blood banker or a primary care physician you're seeing people who are totally asymptomatic, who are anti-HCV-positive, who would have no clue that they had hepatitis C if you hadn't picked it up on a blood test. So we're seeing the very mild spectrum of this rear -- the rear end of the elephant. And I think as even shown on the slide, this is the biggest end of the HCV story.
However, if you're a gastroenterologist or hepatologist you're having some of these patients referred to you and you're seeing some of the more severe cases, but you also realize that most of the cases you see are really quite benign. But you have a more balanced perspective.
On the other hand, if you're Myron Tong in a tertiary care center, or you're a transplant surgeon, you're seeing people with terrible, end-stage hepatitis C and you think this is a horrendous disease.
Well, the truth is, it is. It is a horrendous disease and it is a mild disease, and we're just talking about proportions. And it's my estimate that the big proportion is the big rump of this elephant and that most people -- 70 percent of people -- will have a lifetime of relatively benign hepatitis C.
But that's my guess and we don't have the data yet because we haven't lived long enough, and I clearly will not. In any event, you've been a very attentive audience and I thank you for that attention. Thank you very much.
CHAIRPERSON TABOR: We'll save questions for the discussion period after the break. The next speaker is Dr. Steven Kleinman from UCLA.
Return to Table of Contents
DR. KLEINMAN: Thanks, Ed. Formerly from UCLA, I should say.
My task for today is to talk about the prevalence and incidence of hepatitis C in U.S. blood donors. I'm going to confine my remarks to volunteer blood donors and I will mention at the end of the talk a little bit about some of the other viruses that we're concerned about.
So again, the objectives are to discuss HCV prevalence and incidence in allogeneic donors based on these data to project HCV genome amplification testing yield, and then to contrast the prevalence and incidence of HCV to other transfusion transmitted viruses.
I'll be using two data sources: one is the REDS database which encompasses the years 1991 to 1996, and the other is the database from the American Red Cross. These data are for 1996 and '97, and these data were supplied to me courtesy of Dr. Susan Stramer of American Red Cross.
Now the database for REDS retrovirus epidemiology donor study that most of you have heard of, sponsored by NHLBI, consists of approximately one million annual, allogeneic donations which have been collected from five U.S. blood centers, and the viral markers that we analyzed -- this slide's taken from a previous presentation -- were for: HCV, which we'll talk about today; antibody HBsAg; HIV antibody -- which we did some modification by excluding certain false positives; and HTLV antibody, which I won't say too much about today.
And what we did was determine incidence rate for four, overlapping 2-year intervals, and we defined incidence as the number of incident cases or sero-conversions. So negative anti-HCV tests progressing to a positive anti-HCV test per 100,000 person-years, and all donors included in these calculations gave at least two donations in the specified interval.
And then we also calculated annual prevalence in first-time donors, and that's pretty straightforward. It's the number of positive, anti-HCV positive donations in first-time donors per 100,000 first-time donations, and then we looked at these over time.
Now, the tests that were used to generate this data, we started generating the data at the beginning of multi-antigen EIA testing or EIA-2. Most of the centers use the Ortho EIA.
And the confirmatory tests for the first year of data were first generation RIBA, and then from July '93 through the end of '95 were second generation RIBA. And the first half of '96 is the same but the second half of '96 which I'll show you, is based on version 3 testing, or EIA-3 and RIBA-3 testing.
Now the Red Cross database which is '96 and I think the first half of '97, are from American Red Cross regions that comprise 38 percent of American Red Cross collections. The data were compiled for previous BPAC and so -- which was looking at plasma production issues, and so it's for data less than -- for donors less than 60 years of age; which, since that's the large majority of donors I don't think puts much of a skew in the data.
Three of the regions in the Red Cross database are also in the REDS database, and the HCV assays from June '96 onward all use the EIA-3.0 and RIBA-3.0.
So these are the data that I'll show. Now, this first slide of data is basically, if you just take a rate per donation. So all allogeneic donations collected, how many are HCV antibody-confirmed positive. So this is not really an incidence or a prevalence; it's a frequency that you'd see in everyday screening.
And as you can see here, if we take the aggregate REDS data we get 152 per 100,000 -- or, that's 1.5 per 1,000. If we take 1996 and look at the REDS data you can see we get 105 per 100,000. For the first half of the year we're going to use 2.0 and 125, so a slight increase when we use 3.0.
And if you take the Red Cross data which is partially 2.0 and 3.0, we get 112. So for 1996 and 1997 we seem to have consistent rates of slightly more than one per 1,000 donations testing positive.
Now, this is prevalence in first-time donors, so if we only look at not all donations but donations from first-time donors, again taking the comparable time periods you can see in aggregate we had about -- these are 95 percent confidence intervals here -- we had 541 per 100,000 during the 5-year period. If we stick to 1996 we have close to 400 per 100,000 using 2.0; and using 3.0 it increases.
And as you can see here, this is statistically significant. And I think it has several potential explanations which are involved in switching from 2.0 screening tests to 3.0, as well as switching from 2.0 RIBA which generates some indeterminant test results from infected people, and those people are now positive by RIBA 3.0.
So I don't think this represents a dramatic or even real increase in HCV in donors. I think some of it may be part of changing tests. And again, the Red Cross data is consistent, falling within these same ranges. So we're talking about, in first-time donors, 400 to 500 per 100,000.
And in the REDS database we took a look at the temporal trends in HCV prevalence and you can see here the reason the aggregate data is larger than the '96 data is because of the first -- primarily the first year of HCV testing where we were in first-time donors, finding rates between 600 and 700.
Now, I'm not -- in re-analyzing this data I'm not convinced that this is really a decrease in prevalence. It may be the fact that we were confirming data with RIBA 1.0 at that point, and so a positive on 1.0 may not always be a positive on 2.0. But I think the major point here is a fairly consistent picture of HCV in the first-time donor population.
Now, here's trends in HCV incidents. So this is in repeat donors, and you can see here that hasn't changed over time. And we're looking at levels of about four per 100,000 per person-years.
If we take a look at the incidence data in 1996 we can see that again, we had four per 100,000 person-years. But again, when we were converting from -- when some donors at one donation tested by 2.0 and another one tested by 3.0, we're not clear if that's a new infection or if it could be an infection that was missed by 2.0 screening.
And so when we take a look at the aggregate data we get an incidence of apparent seroconversion of about 6.2 per 100,000 person-years. And when the Red Cross did the same thing they got an incidence of 11.6 per 100,000 person-years.
And the point I want to make here is I hope, illustrated on the next slide -- although I have a feeling these are slightly out of order. But let me go back. The point that I want to make here is that we really need to talk about ranges of incidence as we build models, and so I'll show you some subsequent calculations.
We take a point estimate. Most people have been quoting the number that REDS published several years ago about the risk of HCV being one in 103,000 -- risk of a window period HCV unit being one in 103,000 antibody-negative units.
And I just want to emphasize that that's an estimate and that that estimate has a range around it, and when we try to predict what will happen in the future we're better off taking the whole range and not thinking that the point estimate represents absolute truth.
Anyway, if we now take a look at our frequency of positive donations and we break it down into whether the donor was a first-time donor or a repeat donor, you can see here -- this recaps some of the earlier data; this is in the Red Cross database -- 4.25 per 1,000 donations in first-time donors; only at 0.29 per 1,000 donations in repeat donors.
Since first-time donors make up 21 percent of the database we can project that we'll have in 12 million units collected yearly in the U.S., we'll have a little more than 10,000 HCV antibody-positive results coming from first-time donors.
Repeat donors, 79 percent of the database, we'll have about 2300 results coming from repeat donors. Now the ratio of positive findings in donations from first-time donors versus repeat donors is 14.7 but the actual numbers in our yield, the ratio goes down to 4.6 because it's a weighted average.
But however you look at it most of our antibody-positives come from first-time donors. This should be no great revelation to anybody in the room since first-time donors represent prevalence and again, repeat donors represent new infections.
Now, in a manuscript that was published a couple of years ago we looked at -- within the first-time donors we looked at this slightly differently. And this is '92/'93 data. And we looked at people who make one donation to REDS, said it was their first-time, only donation, and never came back -- their rate was eight per 1,000 donors.
And so obviously, they didn't come -- one of the reasons they didn't come back was because they were positive and couldn't come back -- for the positives -- and then other people elected not to come back.
But we had a group of first-time donors who did come back and make subsequent donations in those two years, so they started out as first-time donors and became repeat donors, and their rate was actually the lowest: it was 0.4 per 1,000 donors HCV-positive. So the point is, when you look at a first-time donor some of those people will go on to be repeat donors. They seem to be as safe as -- or have at least as low HCV infection rates as any of our first-time -- as any of our repeat donors.
Now, to summarize the HCV temporal trends here, the incidence has not varied over time since we began collecting data in '92, through the period of EIA-2 and RIBA-2 tests. The prevalence in first-time donors has decreased. Most of that decrease came subsequent to 1993 and has remained fairly stable. And the first and only-time donors show the highest prevalence.
Now turning to incidence for a minute, I want to go through relative risk. Everybody's seen this: the per unit risk is the incidence rate times the length of the window period. And so when we get to the window period for HCV, this is the window period as determined by Correo's data of transfusion-associated HCV cases, and I think it's duplicated by several other people.
And you can see here from the time of transfusion it takes about 53 days to elevated ALT, about 70 days to EIA-3.0 positivity, and about 82 days -- these are average figures -- to 2.0 positivity. So it's only a donor in those fist 70 days of infection who would be HCV antibody-negative and potentially capable of infecting recipients and potentially capable of detection by HCV genome amplification testing.
Now here's a slide taken from Mike Busch's look at the older transfusion-transmitted virus study repositories. And I show this because of a couple of reasons. What it illustrates is -- and they determined time from transfusion to ALT elevation of 90 in a series of -- I'm not sure how many of it -- about 30 or so patients; and also time to PCR detection.
And I want to emphasize the ALT thing here because - and remember, ALT takes a couple of weeks before -- it comes up a couple of weeks before antibody. And this basically shows a distribution of times to ALT development with about half of the patients developing an elevated ALT at about 45 days or so.
But there is a tail on this distribution, and while it's not directly relevant to this presentation it is relevant to issues of how long the window will be, and certainly I think, came into discussion when determining how long donor retested plasma needed to be held in quarantine before you can assure that 95 percent of the units were beyond the HCV window phase.
So this is just to estimate -- this is really in summary, to show you a range of window periods for HCV antibody development, but all fairly tight within a couple of months.
Now, if we then do our risk calculations, again as I emphasized, we estimated one in 103,000 donations might be infectious -- antibody-negative but infectious. But the 95 percent confidence interval was relatively wide.
Now all the newer incidence calculations using more recent data are all within this 95 percent confidence interval; however, the original point estimate shows a slightly lower risk than point estimates using the newer data and that's why I wanted to use the range in the next slide here.
And what I've done in this slide is try to, based on both prevalence and incidence, estimate how many positive PCR -- HCV-positive PCR tests we would expect to generate when screening the U.S. volunteer blood donor population -- assuming for a moment we were doing single unit HCV GAT testing and then making the extrapolation that our pool testing will pick up everything that our single unit testing would if it's sensitive enough -- as I'm sure other speakers will talk about -- you get these numbers.
We have 13,440 HCV antibody-positive donations per year, out of which we estimate about 90 percent will be PCR-positive -- both based on data that Harvey showed about the ten percent who resolve infection, and on studies that have actually been done, both in the U.S. and Europe on HCV antibody-positive donors.
So that would yield about 12,096 PCR-positive units. We then go into the window period units, those people who are antibody-negative but infectious, taking the range of the lower 95 percent confidence interval up to the point estimate of 103,000.
We'd get anywhere from 116 to 428 donations that were donated within this window period of which, since it takes about 12 days on average for the PCR to become positive, we can estimate that we'd pick up 56, 58 out of the 70-day window. So we'd pick up about 80 percent of these as PCR-positive. And so we'd get an estimate of somewhere between 93 and 342 PCR-positive units from these potential, infectious, window period donors.
Now the third category that we've talked about in the past are persons who are persistently antibody-negative, who have viral nucleic acid, and who are therefore chronic carriers who are never picked up by antibody testing.
A number of years ago we thought that these people are quite common, but I think most of us now believe, looking at some of the data that's come out of nucleic acid screening programs, both in Europe and in the plasma industry here, that these people are in fact, rare, and I don't think will have much of effect on the overall calculations.
So you can see here that if we summarize here PCR-positive donations, 97.3 to 99.2 percent will be from antibody -- HCV antibody-positive donors. And the ratio of PCR yield in antibody-positive donors to that in antibody-negative donors, will be anywhere from 35 PCR-positive antibody-positive units per PCR-positive antibody-negative units, all the way up to 130. And it's probably more like the 130 range because that probably is more in line with the point estimate for HCV risk.
So the take-home message is, if we screen antibody-positives most of those -- that will be the largest contribution to our PCR yield.
I want to turn now to the other markers very briefly, and summarize the REDS 1996 data on viral marker prevalence in first-time donors for the other markers. Here's our figures for HCV 2.0: 399. Here are our other figures for the other virus: HIV at 15.3. So the lowest prevalence in first-time donors, HTLV at 45.8 -- and these are 95 percent confidence intervals.
Surface antigen, somewhat comparable to HCV; about half of HCV at 212. But again, we believe that surface antigen positivity doesn't tell us the whole story about HBsHBV infection since people who acquire infection obviously don't keep their surface antigen for life. They lost it; it's a transient marker.
Here's the HIV prevalence numbers over the years to show that we did see a drop in prevalence between '92 and subsequent years. And this has held true. We don't really have an explanation for why this drop.
We know there were revised criteria put into place in blood donor screening in 1992, but in fact, most of these revised criteria were a bit more liberal and did not require a lifetime deferral for certain behaviors but had changed those to 12-month deferrals. So I don't know why that would explain it other than the fact that maybe blood centers looked at their questions more carefully and maybe the screening process was improved as a byproduct of FDA coming out with its guidance document.
Now here's our viral marker incidence, so new infections in '95/'96 in REDS per 100,000 person-years, once again we get the same sort of relationships. HIV is low, HTLV is relatively low.
Surface antigen is higher but we did an adjustment on surface antigen -- I don't have time to go into it today -- but we do believe that certain neutralized surface antigen positives, when these donors don't have anti-core, actually represent false positive surface antigen tests.
They don't affect the prevalence of surface antigen very much but they do affect the incidence because the numbers are smaller and that effect is amplified. And so I'm not sure which of these numbers are correct; whether it's five per 100,000 or 2.5 when we do our adjustments, which are based on some assumptions.
And then for HCV we're at 4.0. So these data are all summarized -- oh, I'm sorry, here's one more slide on HIV incidence to show that while prevalence decreased, incidence, while it looks like it's varying a bit, these confidence intervals all overlap, and 1996 is back up here at 1.5.
And so in fact, the incidence of HIV -- so those donors who we're not picking up by antibody testing who might be transmitting infections to other people, really has not changed since 1991. It's been constant and it's been at a very low level.
And here I've summarized all the relative frequency data in this last slide and I'll take you through the columns. First-time relative prevalence using the '96 REDS data. If we take HIV as a relative prevalence of one, then HTLV is three times more common in first-time donors, surface antigen is detected 14 times as commonly as HIV, and HCV is detected 26 times as more commonly.
So it's that marker that has the biggest yield in our screening of blood donors today; at least the direct infectivity marker. Obviously, anti-HBC would have the highest prevalence.
Now, this is the -- all donation data from the Red Cross database and you see quite similar relationships, although surface antigen seems to be a bit lower in proportion; HCV again, is very high.
And if you then take the incidence data, the actual donors that we're worried about that we're missing with current tests who might transmit infection, again, if HIV is taken as the reference at 1.0, HTLV is quite comparable.
Surface antigen, if we adjust and remove the false positive cases, is probably only slightly -- so hepatitis B transmission is probably only slightly more common. However, if we don't remove those cases we might project that it's much more common. And HCV again, two-and-a-half times the risk of HIV.
So that's the last slide. I think the points I want to leave you with are that, when we compare the viruses, HCV really is a good virus to target as our first -- in our aim to improve the safety of the blood supply it's probably the one that carries the greatest risk currently, given our screening procedures.
And at least if one takes into account the potential for some percentage of those infected recipients to develop chronic disease. Thank you.
CHAIRPERSON TABOR: Thank you, Dr. Kleinman, for that interesting talk. The next speaker is Dr. Peter Simmonds from the University of Edinburgh.
Return to Table of Contents
DR. SIMMONDS: Good morning and thank you very much for the invitation to this meeting. I've been asked to talk about the protection of hepatitis C in blood donors and detection of other viruses, and so I'm going to start with hepatitis C virus and just describe very briefly our experience of introducing PCR-based screening for blood donors in Scotland and Northern Ireland.
The background really, to our screening has been the regulatory requirement that's imposed on us to actually screen final pools for hepatitis C RNA. And clearly, the costs of actually getting a positive pool are substantial, and clearly, pre-screening donations in mini-pools is fairly cost effective in preventing that occurring.
So in fact, they're not really providing the framework for our group to actually get the mini-pool testing set up. And of course that has knock-on benefits in terms of establishing of framework for the screening for other viruses as well.
Just to summarize our final pool testing, we've been doing it for about two years and have screened the equivalent of roughly 600,000 donations. And about a month before we started mini-pool screening we actually got a final pool positive, okay, which is roughly made out of 6,000 component donations. And obviously, that was quite expensive.
Well, it would have been had we been able to use U.K. plasma. Anyway, so we got a positive unit and through the archive we managed to identify the donor. And it was actually a donor who was very close to the window period and was actually only positive in one of the four, commercial screening tests available, and that was mono-reactive in RIBA-3. We don't know any more about the donor. He has not donated again.
So with this framework established we've got mini-pool testing set up and obviously, in the future we acknowledge the need to screen for red cell release, and possibly in the future, platelet release, and obviously to introduce screening for other viruses; and obviously HIV-1 is clearly a candidate virus.
Okay, so mini-pool screening has been going since April 1998, and we've got a mean turnaround time of three to four days, which isn't really good enough for platelets and it will be difficult for red cell release as well at this stage. So clearly that needs to be reduced.
Our initial testing is done in pools 96. These include antibody-positive samples. When we get the serology results these are excluded and we retest. So far we've screen 150,000 donations. We got about seven positive pools and these all contained antibody-positive units. These in fact, were the only antibody-positive units that were PCR positive so far. So we've picked them all up in the mini-pool system.
So far we have no antibody-negative PCR-positive units, and that's roughly in line with expectations. We think, are basing our strategies on around about one in 200,000 -- perhaps one in 400,000 -- will actually be picked up this way. So clearly we've not tested enough to find a positive at this stage.
So what we are a bit anxious about are these further development issues. Turnaround time is clearly necessary. We would like to go towards closed tube detection of PCR products to avoid contamination. And at the moment we're just about at the stage of validating parallel, HIV screening for one at group O, and the new virus, N.
I'm going to turn on to other viruses, and the second part of the talk is really just to try and briefly summarize what the clinical issues are involved in screening for some of the other, more recently discovered viruses. And clearly there are issues that are relevant that surround hepatitis G or GBV-C.
And what I'm going to describe is just a small study where we try and assess what the significance of this virus will be. That's where we wanted to work out the prevalence of viremia in the donor population, and then to try and make some sort of clinical assessment to the donors and to establish whether there are any particular disease associations in individuals who don't have hepatitis C co-infection.
And then to do some incidence calculations to see whether in fact, we could actually screen for it anyway. And we screen 1,020 in pools of ten, split them down into component donations. We identified 23 positive donors out of 1,020. These are all regular donors and it's possible that the prevalence of viremia may be higher in first-time donors.
And 19 came back and they had a mean age of 32. We could establish no parental risk factors for infection from interview. The ALT level was 20 and if anything, was lower than the ALT levels in controlled donors. And they were quite carefully clinically assessed, and in fact we could find really nothing remarkable about them at all. All the other liver function tests were normal, there were no other disease associations that we could see.
Unfortunately, we had no clinical indications to do a liver biopsy and so we don't really know what that would show. But obviously, the suspicion would be that in fact, that might be normal.
I think the relevance for any attempt at PCR screening is the incidence calculation we were able to make. These were regular donors; in fact, in the archive we were able to retrieve samples from previous donations collected right back to about 1984.
And what we did was to take a series of prevalence calculations from '84, '86, and '89. And in that way we could actually retrospectively work out what the incidence. And when we did this we were actually quite surprised to see that in fact, there had been 17 seroconversions for PCR in a mean of 9.7 years.
And so the incidence in this population is about 200 per 100,000 donations. So it's much, much higher than any other viral markers we've just heard about, by possibly as much as a factor of 100 greater than hepatitis B and 50 times greater than hepatitis C.
And the problem with that then is that if you were going to screen for it, in mini-pools at least, you'd have a real problem, because obviously pools of 100 for example, would contain about -- ten percent of these would actually contain a positive unit if this was the incidence now.
And if you made pools of 500 you'd find half of them were positive for hepatitis G, and that's after excluding your positive donations already. It's just the high incidence actually stops you doing it in mini-pools.
Okay, so provisionally at least, and obviously you know, we need to keep this under review, there is no clinical indication to screen for hepatitis G. Technically we can't do it with the current setup because it's not feasible to do in mini-pools.
And of course the final thing is, is that even if we did it we'd actually have a major reduction in the donor panel. We tried this and obviously that has knock-on effects in terms of blood supplies.
So the provisional decision at this stage is not to introduce any hepatitis G screening, and obviously we're keeping that position under review.
I'm now going to talk very briefly about the newly-discovered transfusion-transmitted virus. And again, we're going to try and attempt to address some of the issues that surround possible transmission of TTV by blood transfusion.
This is really a summary of data published by Nishizawa at the end of last year. There's a new DNA virus. It's been partially characterized. And it was actually initially discovered by subtractive PCR from patients who had non -- post-transfusion, non-A to non-C hepatitis.
They identified a 500 base pair clone after three rounds of subtractive PCR. And the sequence seemed to be coding but didn't correspond to anything on the databases. There's been some characterization work of this virus.
It seems to have a single-stranded DNA genome which is quite unusual, and the only real virus family which is comparable would be the parvaviruses. The clone's been extended out to about 3000-some hundred bases, and it's arranged with a couple of open reading frames. We don't know whether this is a full genome or not; probably not.
There is a large open reading frame which is largely enriched and it's been suggested this might be a structural protein that may form a nucleic capsid.
So we embarked on a study looking at relevance of this blood transfusion. It's quite easy to pick up by PCR and so we wanted to really look at the prevalence of active TTV infection of blood donors.
And we've done a series of subsidiary investigations looking at its role in fulminant hepatitis, and also to see whether in fact, it's transmissible by blood products. And you can get an idea of that by screening hemophiliacs who have been treated with non-viral, inactivated concentrates.
Just to give you the basic data, we screened 1,000 donors. They're screened initially in pools of ten and then split. And splitting them we identified a frequency of nearly two percent per 1,000 donors, so that's quite high.
I'd like to just qualify this prevalence figure. The virus itself is extremely heterogenous in sequence, and certainly none of the primers that are published to-date including our own, will necessarily pick up all genetic variance of the virus.
And in fact, it could well be that in fact, the more we study the virus and gauge its variability, the higher this prevalence figure is actually going to get. Okay, certainly we know that in Scottish blood donors there a much more divergent variant than we've been able to pick up so far.
There's no association at this level with hepatitis G, so there are no cases of co-infection with TTV. There were an over-representation of males in the blood donor population. And in fact, the mean age was 53 which is very unrepresentative of the blood donor population and gives you a clue about its epidemiology.
And at the moment we're getting the archive samples out to look at the incidence of the virus in the same way as hepatitis G.
We've since done some survey work in other countries, and it's quite remarkable how the prevalence varies in different parts of the world. In Japan it's been described to be present in about 12 percent of donors, but if you actually look at some of the data generated from tropical countries you see a quite different pattern of epidemiology.
So for example, in African countries you can see prevalences ranging from seven percent in Sudan which we don't understand, right up to 83 percent of the adult population in Gambia, with Zaire being roughly 50 percent.
We find the same very high prevalence in Papua, New Guinea, and also some of the indigenous peoples in Central and South America. And we've done some work looking at the acquisition of TTV in Zaire and we've been able to show that in a population where the adult prevalence is around about 50 percent it seems to be that you acquire TTV in the first year of life.
So for example, in children born of a cohort of about 150 women, we've found an instance of 54 percent over the first year of life. Interestingly, they're all negative at three months but they became positive at 12 months, and we think that in fact, there's an environmental source for the infection.
Part of the reason for saying that is that if a child acquires TTV it's not related to the TTV status of the mother. So a child from a TTV-negative mother is as equally likely to acquire TTV as a child of a TTV-positive mother.
And so, you know, broadly I'd say the epidemiology is more comparable to hepatitis A, especially with this differential prevalence of viremia in different countries.
Okay, there's widespread contamination of blood products with TTV. These are figures from pre-virus inactivation that we published before. These are volunteer unpaid donors. We can see quite frequent detection of TTV in at least Factor VIII and Factor IX concentrates.
We also detected after virus inactivation. So for example, amongst these volunteer donors' products which were largely solvent detergent treated, we still detect TTV quite commonly with no evidence that in fact, this has actually removed the virus from the concentrates.
On the other hand I think on these very small figures, the concentrates manufactured commercially seem to be lower than these. The difference may be because these are largely heat-treated whereas those are solvent detergent. And this sort of makes sense in terms of the virus possibly being non-enveloped and more similar to parvavirus than to hepatitis C.
And the evidence at this stage will be that it is transmissible by blood products, because we do see high frequencies of infection in the hemophiliacs. And I just want to qualify these prevalence figures. These are based on primers that will not pick up all genetic variants of TTV, so in fact these figures may be higher than shown here.
But clearly, if you look at patients with hemophilia A who were treated before 1986, so they received non-inactivated concentrates, you can see an increasing prevalence of viremia with the disease severity and Factor VIII usage. Similarly for Factor IX where the moderates to severe have a greater than 50 percent prevalence of viremia.
This figure is the important one. These are hemophilia A patients who have received only virally-inactivated concentrates. And so far we've only found one positive out of 19, but clearly we need to increase that number to establish whether any of the inactivation procedures are effective for TTV.
Because clearly we know they're not effective for B19, and we get regular transmission of B19 in this particular hemophilia group.
So in summary, the regulatory position as regards hepatitis C actually does give us this sort of possibility of setting up a framework to get PCR screening blood donors established. And if you can do that then obviously there are a series of interesting issues concerning what viruses you'd actually want to screen for.
Obviously the greatest clinical benefit will be for recipients of non-inactivated components, so these are the red cell and platelet components. So clearly that's going to be a major strategy or priority in terms of any further development and work.
And key in that is actually getting the turnaround time down to about a day. But having said that, B19 still transmits through blood products and clearly non-enveloped virus may also survive some of the inactivation procedures. So again, PCR screen may be relevant for blood products as well.
And obviously, the criteria that we need to consider, obviously what the frequency actually is, what the incidence is, whether the other methods for screening are adequate. And obviously, there's a contrast between hepatitis B and hepatitis C.
And then finally the disease associations of the virus, and I think we've more or less drawn a blank with hepatitis G. We may draw a blank with TTV in the future.
And finally, I'd like to just acknowledge the groups in the University of Edinburgh and the Blood Transfusion Service who have been responsible for this work. Thanks so much.
CHAIRPERSON TABOR: Thank you, Dr. Simmonds, and we'll also look forward to hearing more about your TTV work tomorrow at the Advisory Committee Meeting.
The next speaker is Dr. Jens Bukh from the National Institute of Allergy and Infectious Diseases.
Return to Table of Contents
DR. BUKH: I'd like to thank FDA and the organizers of this workshop for inviting me.
Hepatitis C virus is the sole member of the genus hapacivirus within the flavivirus family. Among the flavivirus the hepatitis C virus is most closely related to a group of unclassified -- the so-called GB agents -- in particular, to a virus called GB virus B; a virus that causes acute hepatitis in the experimentally infected Tamarins.
The single-stranded, positive sense, INA genome contains a single, long, open reading frame that encodes structural and non-structural proteins. Characteristic is the extensive, genetic heterogeneity at the nucleotide level and the deduced amino acid level throughout this open reading frame, but in particular in the two envelope genes and also hypervariable region as Harvey also mentioned, has been defined in the amino-terminal end of E2.
Now, this genetic heterogeneity in the open reading frame makes it next to impossible to design primers and probes for sensitive diagnostic assays throughout the open reading frame. Fortunately however, the five prime non-coding reading or untranslated region of about 341 nucleotides contains highly conserved domains.
This slide shows the genetic heterogeneity of the 5 prime UTR of hepatitis C among 90-plus HCV isolates from around the world. The red bars indicate the percent of these isolates which differs from the consensus sequence. And green triangles indicate precisions with insertions in certain isolates.
And the important message of this study that we published back in 1992 is the existence of these universally conserved sequences. For example, this region of 63 nucleotides that are variant among all these isolates that represented all of the six major genotypes of hepatitis C virus. So these are of course, unique for diagnostic assays.
I should mention also that in the 3 prime untranslated region that consists of the short, variable sequence followed by a poly U-UC region of variable composition and length, but at the very 3 prime end there's a highly conserved sequence that was actually first identified in 1995.
This sequence of about 98 nucleotides is highly conserved among the genetic variance of HCV. However, the secondary structure, especially at the very 3 prime end, is very strong so this might cause some problems for standardizing nucleic acid test using this sequence.
Just to illustrate the importance of selecting conserved primers or probes for diagnostic assays, I'd like to show the data from this study where we selected five different primer sets from at the time, what we thought was conserved sequences based on available sequence data.
One such primer pair was from within the 5 prime UTR; two other different primer sets had the sense primers in the 5 prime UTR, and the anti-sense primers in the capsid gene. And finally, a fourth primer set had the primers within the conserved helicase sequence in NS3.
We showed that these primers had equivalence sensitivity on a matched control, and when we tested equivalent amount of HCV RNA from 114 first-generation anti-HCV positive patients in a nested RT PCR assay, we found that the primers from NS3 only detected about one-third of the patients that were detected with the primers from within the 5 prime UTR.
And we found that all of these 27 patients detected with the NS3 primers were of genotype 1, whereas patients detected with the 5 prime UTR primers represented all of the six major genotypes of HCV; again, illustrating the importance of selecting conserved primers for diagnostic assays.
Now, we originally characterized the existence of six major genotypes by analysis of the E1 gene from 51 HCV isolates from around the world: genotype 1, 2, 3, 4, 5, and 6. Some of these major genotypes had well-defined subtypes such as genotype 1A and 1B.
The same six major genotypes were characterized by Dr. Simmonds by analysis of partial NS5B sequences. And analysis of full-length sequences have indeed, verified that HCV can be classified into six major genetic groups.
This slides shows a phylogentic analysis of the polyprotein in sequences that we recently performed of representative isolates with genotype 1, 2, 3, 4, 5, and 6. It has been suggested that the isolates that were published as genotypes 7, 8, 9, and 11 actually cluster with the genotype 6 isolates. And here is shown the full-length sequence of genotype 11. The full-length sequence of genotypes 7, 8, and 9 were only published last month.
And also it's been suggested that the isolates published as genotype 10 should cluster together with the genotype 3 sequences, and an analysis of full-length sequences certainly support these suggestions.
So in essence, HCV can be classified into six major genetic groups with a huge number of sub-types that vary to varying degrees.
This slide shows the geographical distribution of the six major genotypes of HCV. Clearly, genotype 1 is the predominant genotype in North and South America, in Europe, and also in most areas of Asia.
Genotypes 2 and 3 are found throughout these regions at a somewhat lower prevalence. Genotype 4 is the predominant genotype in Egypt and in Central Africa; although other genotypes have been described in this region of the world.
And genotype 5 is the predominant genotype in South Africa. Both of these genotypes are only found sporadically outside of Africa. And genotype 6 constitutes a significant proportion of isolates in Southeast Asia and are only found sporadically outside of this area.
A number of indirect methods have been developed to determine the genotype of HCV. Dr. Komoda and co-workers developed a type-specific PCR for HCV genotyping in which a portion of the core gene is amplified in a first round of PCR with conserved primers. Then a nested round of PCR is performed with conserved sense primers and type-specific, anti-sense primers, and a genotype is deduced from the size of the PCR amplicon.
This method is quite specific but the shortcoming of course, is that this method can only discriminate between a few of the described genotypes.
Another method that was developed by Dr. Simmonds and also by others is the restriction fragment length polymorphism for HCV genotypes in which typically a portion of the 5 prime UTR is amplified with universal primers in a nested PCR and then the amplicons are exposed to digestion with specific restriction enzymes. And the genotype is deduced from the specific pattern.
The shortcoming of this method is due to the high degree of conservation in the 5 prime UTR. Because of this conservation a number of the recognized genotypes cannot be distinguished.
The LIPA assay which was developed by Dr. Styver and Maertens at Inno Genetics -- this assay is based -- consists of a strip with a genotype-specific probes down to it. And then this strip is reverse hybridized with amplicons of the 5 prime UTR.
The shortcoming of this method is the same as those for IFLP and is due to the high degree of conservation of the 5 prime UTR. So for example, in Italy many genotypes that were classified as genotype 2A with this method actually turned out to be 2C. And there's other examples like that.
Finally, I should mention that there's also serotyping methods on NS4 developed by Dr. Simmonds and co-workers. This method can also determine type-specific antibodies to the six major genotypes of HCV.
Again, the most specific method of cross sequence analysis, and it's generally recommended to perform sequence analysis of core E1 or NS5B since the most comprehensive reference data is available for these regions. But in fact, genotype-specific differences exist throughout the open reading frame.
And I just want to mention that of course with the highly conserved sequences available in the 5 prime UTR, it should be possible to design diagnostic tests, both qualitative and quantitative, that have equivalent sensitivity against the various genotypes of HCV.
However, the commercial tests had had problems in this regard. We have used both of these tests in chimpanzees infected with various genotypes of HCV and a number of other studies have looked at this issue. The monitor tests which is -- the first-generation monitor test which is still at the market clearly underestimates the titers in patients infected with genotypes other than genotype 1.
And I think this is being addressed in a second-generation test that I believe will soon be marketed. Also, this test underestimates the titers in patients with high viral titers.
The first-generation bDNA test which is now off the market, clearly also underestimated the titers of certain genotypes. In the second generation test this problem has presumably been eliminated. It certainly seems like that.
However, the problem with this second-generation test of course, is the low sensitivity. I believe there's a third-generation test underway to address this issue.
These genotype-specific differences of course relate to all in-house RT-PCR assays as well: qualitative or semi-quantitative. And this points to the importance of using standardized controls as the ones developed by Dr. Lelie in Europe where he -- for certain genotypes.
And I also wanted to mention that we have generated plasma pools of prototype strains of HCV of genotype 1B, 2A, 2B, 3A, 4A, and 5A in chimpanzees. This was study done in collaboration with a number of scientists and also among those, Steve Feinstone and the FDA.
And we've determined the genome titers of these pools with available commercial testing, also, and by in-house tests -- the tiers of which are shown here. And we've determined the infectivity titers of those pools by reverse titration in chimpanzees, and they all had infectivity titers ranging from 103 to 105 infectious titers -- doses per ml.
We also have a plasma pool, the 8s pool that Dr. Harvey Alter developed, with known infectious titer. And this is a genotype 1A pool. It is our hope that these standardized pools could be useful in standardizing diagnostic assays.
But also of course, these pools would be important -- could be important for challenge in future vaccine studies in chimpanzees.
I will also mention another very important aspect of the genetic heterogeneity of HCV is the fact that HCV circulates cross-species in a single, infected individual. Whereas the differences between isolates of some of the major genotypes can be as high as 35 percent, and isolates within a subtype vary typically by five to ten percent. The differences among cross-species is generally less than two percent.
We recently determined the complete open reading frame sequence from 18 clones deduced from the acute phase sample that Harvey Alter also mentioned earlier. And we found that cross-species -- that some of these sequences varied by about two percent over the entire genome.
We found the changes were found too, after genome, although there was most changes in the hypervariable -- in the envelope genes, in particular in the hypervariable region. But in fact, the cross-species could be observed throughout the genome.
We also went on and studied a sample taken 20 years later in this patient. We determined the consensus sequence of the complete open reading frame. And this slide shows the mutation rate in this patient during 20 years in the individual gene regions, expressed the substitutions times 10-3 per site per year.
Nuclear-type mutation rates are shown in blue and amino acid mutation rates are shown in red. The highest note that I've divided the E2 gene into the hypervariable region, which is the amino terminal 30 amino acids of E2 and the remainder of E2 are 333 amino acids.
The highest mutation rates were found in E2, P7, and NS4A, but there was a particularly high mutation rate in the hypervariable region and most of the changes in this region resulted in amino acid changes.
And of course, this region is the region that has been found to undergo sequential changes in infected individuals. And also this region has been found to contain potential neutralization epitopes. So as Harvey Alter indicated, this region is thought to be of great importance for the persistence of HCV.
Now, one of the problems with studying the cross species in humans or in experimentally-infected chimpanzees, is that from the outset these patients are infected with cross-species. So it's different to see which mutations are new mutations or which mutations are just selection of pre-existing variants.
Last year we developed an infectious clones of genotype 1A and genotype 1B, and the infectivity of these clones was tested by transfection of chimpanzee liver with INA transcripts. And the advantage of studying HCV evolution in such infected chimpanzees is that these chimpanzees presumably are infected with the monoclonal virus population.
We transfected two chimpanzees with the infectious clone of genotype 1A. And this slide shows the course of the infection of one of these chimpanzees. This chimpanzee had an acute resulting infection with viremia from week-1 to 23 post-inoculation. We could not detect the virus in serum samples from week 24 through week 52 post-inoculation. This site shows data to week-40.
The viral titer detected with the monitor test and also with in-house shown in red dots, and with the in-house RT-PCR shown in blue columns -- that viral titer increased from about 102 of week-1 post-inoculation to peak viral titers of 105 to 106.
The chimpanzee developed acute hepatitis with elevated liver enzyme values and also, although not shown in this slide, necroinflammatory changes in liver biopsies. The second-generation ELISA tests became positive around week-15 post-inoculation, and again the virus was cleared in this chimpanzee.
The other chimpanzee however, although this chimp was infected with a monoclonal virus, developed chronic, persistent infection. This chimp had viremia from week-1 post-inoculation through week-70 post-inoculation. This slide shows data through week-52 post-inoculation.
Again, the viral titer increased from about 102 at week-1 up to peak viral titers of 105 to 106. The chimpanzee developed acute hepatitis with elevated liver enzyme values and inflammatory changes in liver biopsies.
This chimpanzee became positive for anti-HCV in the second-generation ELISA test at week-13, and also we detected two antibodies in this chimpanzee with an experimental ELISA test performed at Abbott. And the E2 antibody titers increased over time.
The viral titers following appearance of antibodies and acute viral hepatitis decreased by one to two logs and reached its lowest level at about week-24 post-inoculation with titers between 1000 and 10,000 copies. Again, the viral titers rebounded somewhat and was about 105 after one year's follow-up.
Now, we've studied the sequence in this chimpanzee over time. We determined the complete open reading frame sequence at seven different timeframes during the first year of follow-up. At week-2 post-inoculation the complete -- the nucleotide and the deduced amino acid sequence of the complete open reading frame was identical to that of the infectious clone.
However, at week-8, 12, and 20 post-inoculation there was a single nucleotide change that resulted in an amino acid change, and this slide only shows amino acid changes in the NS3 protease domain. This mutation could actually be detected already at week-5 post-inoculation.
At week-31 and week-40 there was six additional amino acid changes: two in P7, one additional change in P in NS3, two changes in NS5A, and a single change in NS5B. And at week-52 there was a single, additional change in NS3.
Note that there was absolutely no changes in E1 or E2 proteins during the first year of follow-up. And clonal analysis at week-31 of multiple clones showed no evidence of cross-species in this chimpanzee. And in particular, all of the clones that we analyzed had identical, hypervariable region sequence both at the nucleotide and at the deduced amino acid level.
So a main difference from the mutation rates that we saw in this chimp after one year of follow-up and the patient H that we followed for 20 years, was a total lack of amino acid changes in this envelop proteins, suggesting that clearly this virus was able to persist in this chimpanzee without evolution in this region.
Maybe also suggesting that a number of factors could influence the -- could result in the high persistence rates of HCV. Of course we have to remember that these studies were performed in chimpanzees and there could potentially be important differences in humans and in chimpanzees.
And the data that I've presented here today, the actual data, was performed in the Hepatitis Virus Section at NIH. Thank you.
CHAIRPERSON TABOR: Thank you very much, Dr. Bukh. The next speaker is Dr. Betty Robertson from the Centers for Disease Control and Prevention.
Return to Table of Contents
DR. ROBERTSON: Thank you. First of all, I'd like to thank the organizers for inviting me to give this talk. If I could have the first slide please; the remote.
The title of my talk today is the Application of PCR to the Molecular Epidemiology of HCV. And as mentioned earlier, Harvey used the elephant analogy. I'd like to use the iceberg analogy to look at HCV infections in which about five to maybe 20 percent of infections are symptomatic -- either acute or chronic -- but 80 to 95 percent of them are asymptomatic, whether they're acute or chronic.
And as in previous studies which have looked at HCV infections within the United States -- and I'm focusing on HCV infections within the U.S. -- most of the studies have looked at these cases here. And these are summarized on this next slide.
There have been three or four studies which have looked at chronic liver disease patients. Most of these individuals are from tertiary care units. And what we see is in most of the studies there's been approximate ratio of subtypes 1A and 1B being equivalent in these studies.
The one study in which there was a different ratio of sub-type 1A to 1B was from the Zein Study in 1996. These three studies here used either LIPA or a mixture of LIPA, RFLP, and sequence to assign a sub-type. The Zein Study here used sequencing of the NS5B region.
We have focused on a population of patients from the NHANES III collection. This is the National Health and Nutrition Examination Survey, and the study population was selected by a stratified, multi-stage, probability cluster design, and the participants were representative of various ethnic groups within the U.S. population. They were selected to represent the total civilian, non-institutionalized population within the United States.
The methods that were used to do the analysis for this particular collection were the avid HCV assay, the HCV-2 assay, followed by matrix confirmation. RNA detection was an in-house, nested, 5 prime UTR amplification. Sub-type determination was done by NS5B RT-PCR followed by sequencing.
And the overall sero-prevalence using antibody assays was 1.8 percent within the United States. And within the various ethnic groups that were represented in this study, whites had a sero-prevalence of 1.5 percent, blacks, 3.2, and Mexican-Americans 2.1 percent.
If we look at the sero-prevalence by age we see that there is increasing sero-prevalence with age with the majority of the individuals -- majority of the anti-HCV positive being between the ages of 30 and 49, as shown here on this slide. Also shown are the different ratios of anti-HCV positives in the different ethnic groups.
If we look by race and ethnic group and gender in this particular population for sero-prevalence, what we find is that males overall have a higher sero-prevalence than females, and in particular, black males had the highest sero-prevalence with over four percent serum anti-HCV positive within that particular group.
We then looked at the PCR-positive samples within this population. The total number of samples that were antibody positive were 378. Of these, 364 were available for PCR testing. We then used the 5 prime UTR region for detection and found 283 of these positive. This calculated out to a 73.9 percent of the antibody-positive samples were RNA-positive.
And then when we looked at the NS5B region we were able to successfully amplify 97 percent of the 5 prime UTR samples. And these will be the ones that I'm going to talk about now.
As we look at the proportion of antibody-positive that were RNA-positive, in general 60 to 70 percent of the antibody-positive samples were RNA positive. The one difference was found in black males in which they had about a 95 percent of the antibody-positive samples were also RNA-positive.
This is the overall data that we found from this population of anti-HCV positive individuals. The ratio of 1A to 1B is slightly different with about 52 percent being 1A and 27 percent being 1B. The remaining sub-types that were present in this population include 2A, 2B, 3A, and a few number of 4A and 6A.
This looks at the sub-type distribution in the various ethnic groups. And what is striking when we looked at this analysis is that whites and Mexican-Americans had roughly the same ratio of 1A to 1B as we found in the overall population. However, if we looked at the blacks the proportion of sub-type 1B that is represented in this particular ethnic group was much higher than in whites and Mexican-Americans.
When we looked at sub-types as distributed by age we saw no specific sub-type at any particular age group, and each sub-type was generally spread over the general curve of anti-HCV positives -- with the exception perhaps, in the greater than 70 age group in which there were a higher number of 1B infections.
The conclusions from these studies were that the overall anti-HCV sero-prevalence in the United States is 1.8 percent, with blacks having the highest sero-prevalence. Secondly, the majority of the infected population is between 30 and 49 years of age. In general, 75 percent of anti-HCV-positive individuals are RNA positives; however, 90 percent of the antibody-positive black males -- that should be 95; typographical error -- of antibody-positive black males are RNA positive.
Within the general U.S. population sub-type 1A predominates over 1B. And lastly, the proportion of sub-type 1B infections is higher among blacks compared to other ethnic groups.
And lastly, I want to acknowledge the individuals who are responsible for the data I presented in this talk. Omana Nainan, FengXiang Gao, Emory Meeks performed a lot of the laboratory analysis; Pat Coleman, Linda Moyer, Helen Margolis, Miriam Alter involved in analysis in the epidemiology of it; and Deanna Kruzon-Moran and Jerry McQuillan who were involved in the analysis and conception of the studies. Thank you.
CHAIRPERSON TABOR: Thank you very much, Dr. Robertson. We'll now have a 15-minute break. After the break we have time for discussion and questions followed by industry presentations, and I think the industry presentations will provide some of the most interesting technological information of the meeting.
I understand that some of the speakers in the industry portion are interested in using a multimedia format, and we do have equipment available. So if any of you are planning to do that please see me now.
The coffee and other food is available in the cafeteria which is on this floor. The floor is arranged in a somewhat circular format so you can get to the cafeteria by going either left or right when you leave the conference room. We'll reconvene in 15 minutes. Thank you.
(Whereupon, the foregoing matter went off the record at 10:20 a.m. and went back on the record at 10:40 a.m.)
Return to Table of Contents
CHAIRPERSON TABOR: It's time to reconvene the workshop. Could I ask all of the speakers from this morning's session to sit at the tables and we'll have a period of questions from the audience. Would anyone asking questions please use the microphones that are -- there's one at the front of the audience.
Perhaps I can start off the questions by asking Dr. Robertson why she thinks the percentage of black males with HCV infection, with anti-HCV and HCV RNA detectable is as high as 95 percent; which is quite a bit higher than the general population.
DR. ROBERTSON: I don't have an immediate answer for that, unless Miriam Alter might have an idea. It's an interesting observation at this point.
CHAIRPERSON TABOR: Could we ask Dr. Miriam Alter? Do you have any explanation for that? It's just an observation then, at present?
DR. MIRIAM ALTER: It's also interesting that they happen to have -- they're more likely to have 1B than 1A, but we haven't done any cross-analysis to look at RNA by genotype within that sub-group to see if that's the reason or if that -- well, it wouldn't be a reason, but if there's some correlation between that as well in that sub-group. But I don't know.
CHAIRPERSON TABOR: As a long-term, public health issue this might mean that several decades from now there could be a more serious problem with severe liver disease in the black males with anti-HCV than in others? Is that a possibility?
DR. MIRIAM ALTER: If in fact, this is related to progression, which we don't know.
CHAIRPERSON TABOR: Thank you. Are there any questions or discussion from the audience? Please identify yourself when you begin.
MR. BRASS: I'm Cliff Brass, Schering Plough. I just wanted to make one additional comment which relates to this and I think it will be an important public health issue. There are several small studies now soon to be presented which suggest that the response rate to typical therapy in African-American patients is much lower than the caucasian population.
I wonder if there's any other epidemiologic evidence that may give us a reason for that, and just suggest again that this is going to be an important issue.
DR. MIRIAM ALTER: If it's related to genotype then it would be consistent with what we see with other patients with genotype 1 and 1B, particularly. So that -- do you know whether or not they were more likely to have 1B as well?
MR. BRASS: Well, in the studies that I'm familiar with, both populations predominantly had 1, and I'm not sure of the ratio 1A to 1B. I'm not sure if the studies are big enough to really ferret that part of it out.
DR. MIRIAM ALTER: Right, but that could be the reason behind their much lower response as well; in that they are more likely to have 1B.
MR. BRASS: Although I think most of the data I'm aware of right now suggests that, probably certainly for combination therapy and probably for monotherapy, the 1As and the 1Bs have a similar response rate.
MR. JACKSON: Brooks Jackson, Johns Hopkins. A question for Harvey Alter or Dr. Robinson. Of the persistent HCV antibody-positive donors, about ten to 25 percent are PCR-negative, do you think that's -- really reflects containment of the virus or just failure of the PCR assays to detect point mutations in the primer binding sites and that sort of thing?
CHAIRPERSON TABOR: I don't see Harvey Alter. Dr. Robertson, would you like to answer that?
DR. ROBERTSON: I guess -- I don't know whether it's an inability for detection based upon point mutations in the primer binding site, especially when you're using the 5 prime UTR. The sites are fairly conserved.
You know, it might be possibly a reflection of the viremia level. That's always a possibility.
CHAIRPERSON TABOR: Dr. Miriam Alter. We're going to have to get rid of the one dead microphone that's there.
DR. MIRIAM ALTER: There are some supporting evidence that would suggest these could in fact, represent resolved infections in that you're much more likely to find RNA negativity in people who also have persistently normal enzymes. Whereas those with elevated enzymes obviously largely have abnormal ALTs and so it would suggest that in fact, it could be real.
And I know at least in Harvey's data this was a repeat -- these individuals were repeatedly sampled over at least a year and in some cases, more. So he -- even though he's not in the room, we speak for each other all the time -- I think he feels fairly confident that these truly represent individuals who have resolved their infection.
And if I'm not mistaken there are biopsy data showing that these individuals have normal livers. And I think that data is probably available from Europe as well.
MR. TICEHURST: John Ticehurst from FDA and Johns Hopkins. Back again, dealing with this observation about the ratio of RNA to antibody positivity in certain groups, I wonder if you could look at it from a sort of flip side; that maybe there's some differences -- that's pertinent to the subject of the meeting -- if there are differences in antibody response or persistence in certain groups that would be a reason for that ratio to be different.
Have you -- you look puzzled, Betty Jo. Do you understand what I'm -- what I'm saying is, is it possible that there are, for example in the African-American males, that they would have a -- there would be a reason for fewer of them who are truly HCV-infected to be antibody-negative, either relating to chronic infection or a window period?
The only way to look at that, you'd have to look at a subset of your antibody-negative people for PCR positivity.
DR. ROBERTSON: We have not looked at the antibody-negative population; that's true. Based upon what I showed we've preselected for antibody-positive.
DR. SEEFF: Leonard Seeff from -- I'm not sure where I'm from.
CHAIRPERSON TABOR: Now from NIH.
DR. SEEFF: NIH and the VA. With respect to the racial difference, we have some samples I think, that you know about Ed, going back to 1948. A very large sample that we happened to screen for hepatitis C and there was a fantastically, strikingly different frequency of HCV sero-reactivity between whites and blacks.
It's about 25 times higher in blacks than it was among whites. And actually, it's interesting. These were stored sera; they had not been collected in the way we would like to have it collected to make sure that nothing is lost. But a large proportion of the entire groups -- some 65 percent -- were still PCR-positive in saved sera over a 50-year period.
And when breaking according to the racial grouping, about 80 percent of the ones who were black were positive by PCR and a much lower percentage of the whites. And follow-up to see what's happened to them, we have a lower frequency of mortality than we would have expected, but the mortality was higher among the black group than the white group.
So there's a real issue I think, about the difference in the frequency of sera positivity and outcome, and now the issue of response to treatment, which seems to be less -- the results seem to be less good among blacks than among whites -- is an area that really deserves a lot of attention.
CHAIRPERSON TABOR: Thank you.
DR. LELIE: Nico Lelie from CLB in Amsterdam. I've a question to Steve Kleinman about the use of the RIBA-3 -- the introduction of the RIBA-3 and the reliability for confirmation. We did sort of validation study a couple of years ago with the RIBA-3 test and we found that especially double bands could very well be false-positive. And so you have a chance for about a ten percent to find an indeterminate result.
It's also a chance of about one percent to have a double reactivity that in fact, is non-specific. And I have a feeling that this contribution of non-specific RIBA-positive, double reactivities also are included in the numbers or the slight increase in prevalence of HCV infections in your risk analysis.
So maybe could you comment on what happens with the counseling of donors that have those double band reactivities in the RIBA? Or maybe Susan Stramer, what is the experience in the United States?
DR. KLEINMAN: Yes, that's a good comment, Nico. We haven't had a chance to look at the data carefully enough to break out the two bands from the three bands within the REDS experience.
And actually, this increase in prevalence in the 3.0 tested is something that we just really analyzed quite recently for this conference. So I think a lot more work needs to be done. I think there are several, potential explanations.
One other possibility that's revealed by the studies is that we presented last year through the REDS group at ABB that there is an additional -- you can find some samples that are 3.0 positive, RIBA-3 positive that are missed by EIA-2, but they were all PCR-negative.
So whether we're picking up false positivity or resolved infection is unclear. Now I know that -- Sue can maybe comment on this -- but the Red Cross has done studies on 2-band positives and taken them to PCR, and I don't recall the exact figures but was it 45 percent or --
CHAIRPERSON TABOR: Sue, could you --
DR. KLEINMAN: Yes. Maybe I'll let Sue address that and maybe she can address -- as far as I know the counseling messages are not different depending upon whether people have two bands or three bands at this point.
DR. STRAMER: No, the counseling messages, Steve, you're correct; they're the same. But I support what Nico said and probably the increase of -- it's about less than ten percent of confirmed positives converting from RIBA-2 to unlicensed RIBA-3. And under our research protocol for using the unlicensed test part of our research goal is to look at the 2-band positive and determine what percentage of those are true positives.
I don't have any follow-up data for those 2-band positives but what we do know is that 44 percent of the 2-band positives are in fact, RNA-positive; whereas 90 percent or greater of the 3-band positives are RNA-positive. So there's definitely a shift in the numbers of viremic donations that are associated with both of those.
CHAIRPERSON TABOR: Thank you. I think we should move on to the industry presentations. We've had the luxury of extending some of the talks of this morning a few minutes because Dr. Gretch was prevented from attending today at the last minute.
However, I think during the industry presentations we should try as much as possible to stay within the timeframe. We'll begin with a presentation by Dr. Conrad from National Genetics Institute.
Return to Table of Contents
DR. CONRAD: I guess it's still good morning. I'm going to present the methodology that we've used to several million donations in the plasma industry. My colleagues from Alpha and Baxter will present the clinical data derived from this methodology.
Essentially what we're doing is a sample pooling regime using a matrix, and in consideration of what matrix -- what size, what shape, what color of matrix to use -- the issues are sensitivity, time for identification of a donor where we eventually hope to extend this into the whole blood industry, and if you can identify donors reliably, and the quality control exists to make sure that that happens reliably.
So under IND we began investigating hepatitis C HIV on these donors using a pooling algorithm dependent upon automatic pipetting devices computers that was very data management intense. I'm going to describe the one that we decided to use.
Basically you could use a variety of different matrices: 2-dimensional pyramids and 3-dimensionals. I think some of the other speakers from industry will talk about the 2-dimensional and pyramid methodologies. We opted not to use it for reasons that we had trouble with it, and arrived at a 3-dimensional matrix.
And essentially what a 3-dimensional matrix is, it's -- the pipetting device makes a cube. It's really not a cube but to understand it for humans it's basically a cube. What the pipetting device does is it takes 512 samples -- and it doesn't have to be 512 samples; we validated the system for up to 512 samples it can be any symmetrical cube smaller than an 8 X 8 X 8 is what we validated, which is -- 8 X 8 X 8 happens to equal 512.
It puts those samples into row, layers, and columns. What happens then simply, is you test the entire cube. If the entire cube is negative then the components within it are negative. If the cube is positive you test 8 rows, 8 layers -- the yellow -- and 8 columns. The intersection of those row, layers, and columns is the positive donor.
It's fairly straightforward when you say it like this. It was complicated to get it to work.
Basically, the device that we chose is called the TECAN Genesis machine. It's been two years of hard work getting these machines up to snuff where we've modified the software I think probably 60 times to have it working as well as it is now.
As of the 10th of September we have tested 3,500,000 donations with this system. In the clinical trial components of this discussed by Alpha and Baxter you'll get some ideas of the accuracy, the functionality, and it may be even a bit of the expense involved.
Some very important terms that you'll hear from my colleagues from those two companies. Basically, the master pool is the term we use for the 512 cube. It's called a master pool. It's automatically made in duplicates so we have some resolution if there's any issues with positivity.
It's tested four times in our system in a multiplex mode. In order to ensure that we could detect all different strains of hepatitis C equally we have to use two primers. We do everything in duplicate. We think redundancy is one of the most important quality control steps that you can employ.
The row, layers, and columns we call them primary pools. There's 64 samples in a primary pool. And then the individual sample we call the individual sample. It's very important to note, by the time we call an individual "positive", we've actually done 12 tests. So there's some concern about PCR and false positivities.
At National Genetics we separate everything into different buildings, we have all this fancy contamination control. But there's nothing like 12 repeats to help, and I'm sure people like Dr. Stramer who are very experienced in, you know, what repeat testing can bring, can tell you that 12 iterations of a test brings an awful lot of security, and I think that's what you'll see from the Baxter and Alpha data that's presented.
The system is not tremendously prone for false positives, and we learned an awful lot about people who were truly positive though some of the other tests didn't convince us of in these processes.
Basically what any one given run looks like is, it's all the samples. There's 60 samples per membrane; they're transferred onto a membrane. These are positive controls, these are positive samples with the dark bands. They're done with one primer set and then with an exactly identical second primary set. Then they're hybridized with an internal control. Any negative can only be called a negative if the internal control shows and the sample is not showing.
Basically -- so you would see is, in a multiplex run you would see that there would be two different repeats of the sample -- this is how we test a master pool -- two with a second primer set. You can see for this particular sample the second primer set didn't hybridize as well and that has to do with the heterogeneity of the hepatitis C virus, even in the 5 prime, non-coating region.
We wanted to show that the internal control -- we also evaluated what we called the split mode which was a non-multiplex mode -- this was at the urging of some of the fine folks from the FDA -- to compare the internal controls and the non-internal controls.
You can actually see that there is an inhibition of the internal control competing with the native HCV in cases where they are certain heterogeneities in the sequence.
So a positive looks like this, a negative looks like this, but note the negative has to have all four of the internal controls showing. And this is essentially what the computer would read and then give us the positive or negatives, feed that back into the TECAN file, implicate the positive or negative donor. We would then pull out the positive donor sample and test it.
Just briefly, it's important to know that in order for the HIV and HCV portions of this test we performed ICH-3 guideline limit tests to determine the sensitivity. Sadly, there really are no standards for HIV or HCV; this is what we did for both of them. However, we did use the standards from the EuroHep from Nico Lelie's group, from CDC from Ming-ying Yu of FDA.
We used as many of the quality panelists that we could get our hands on, and from Dr. Saldanha -- sorry to leave you out there -- at NIBSC, to ensure that some of the numbers that we're giving are in agreement, at least closely in agreement with the numbers that you'll see from those panels.
For hepatitis C we have a sensitivity, a mean sensitivity around 13 copies per ml using these sensitive methods. For HIV where we needed to compete rigorously with p24 -- and you'll see some data that the pooled PCR system is actually better and certainly equal to p24 testing. And Chuck Frisbie from Alpha will show that to you.
We had to increase the sensitivity by centrifuging more of the original master pool so we brought it down to nine copies. And that was in order for it to compete effectively with p24.
Genotypes, there's been a lot of talk and some of the founders of the information of genotype, Drs. Bukh and Simmonds, can tell you that this is a big issue.
What we did is, we cloned all the major sub-types, put them into transcription vectors, transcribed them, quantitated the RNA and high copy level, diluted it, and then detected it with the multi-primer system that we're talking about and found that the p-values or the Rho values were highly significant, indicating that the system that we're employing now detects all the major genotypes and their sub-types evenly.
That was a bear. But we used those highly conserved regions in the 5 prime region that Dr. Bukh showed earlier. We called them our CON primers, or consensus primers, as well as some of the other primers in order to do this.
Results from the clinical studies performed in collaboration with Alpha and Baxter will be presented today; represents roughly 700,000 donations that we did during this finickal window when we followed the donors up, from about 100,000 donors. So it's 700,000 donations from approximately 100,000 donors. They'll talk to you about that.
What I've tried to do here is describe the methodology, make you familiar with the terms. Now you're going to hear from the results of the studies. I just wanted to mention, we've done pilot studies in hepatitis B and we are finding higher incidences of hepatitis B than we originally suspected.
And it is similar to what Dr. Kleinman said; that there is a tremendous amount of hepatitis B that is not being detected with the surface antigen or core antibody tests. And that data will come forward later. We'll stick with HCV and HIV in those clinical results.
Also, prepare yourself to hear comparisons with the p24 antigen tests, and this is going to be the comparison of pooled PCR in the 512 pool as compared to HIV testing with p24 in the individual. And that's an important distinction. This isn't PCR in the individual sample against PCR in the -- p24 in the individual. It's pooled PCR versus the individual.
You'll also hear some of the implications and the importance of ALT. Does ALT precede antibody or not precede it; some of those issues. We'll really expand on window periods and the length of window periods. This is the first time we've really been able to prospectively look at window periods by using nucleic acid testing.
I think that you'll see a little interesting data on the different antibody kits because now we have a standard to compare them, and that's the gold standard issue. Is nucleic acid testing really the gold standard in that it's the hallmark of the virus; without the nucleic acids there there's not an infectious virus.
There's been questions whether or not the detection of nucleic acids is really -- you know, the fact that some people don't have nucleic acids but have antibodies; what's the origin of that? And we really do believe that that's someone who's resolved the infection. Because we've tried with many different primers, many different nucleic acid tests, and those patients seem not to have HCV.
So that's what we'll see in the subsequent two talks. And I will then turn it over to Chuck Frisbie from Alpha for the first part of that presentation.
Return to Table of Contents
MR. FRISBIE: I'm Chuck Frisbie and for the last two-and-a-half years I've been helping Alpha to implement PCR testing of plasma pools in collaboration with National Genetics Institute. And what I'm going to present today is some of the data we've obtained in our clinical trial and also post-clinical trial HCV testing.
This is a brief description of our clinical design. Eligible subjects had to be for HCV, antibody-negative, PCR-positive. And once enrolled these subjects would be followed for six months or until sero conversion.
For HIV it was similar except for the fact that to qualify for enrollment you could either be p24 antigen-positive and/or PCR-positive. And you'll see later where there are -- actually HIV donor who was enrolled was enrolled based only an the initial PCR-positive result and the absence of p24 antigen.
As far as participation we had 33 study sites. From those sites we collected and tested 342,729 donations and again, over a 4-month period. And in that number we found 22 individuals who were qualified to be enrolled in the HCV study. Of those, 13 were enrolled and nine were not.
For the HIV study we had found four individuals who were qualified and two of those were enrolled and two were not. However, all the data obtained was considered in our evaluation.
This slide gives some numbers regarding the number of donations that were interdicted by PCR that would not have been found to be a reject based on any other testing. For example, with HCV they were negative for antibody and they were not elevated for ALT. We had 75 donations that fell into that category.
In the post -- that's over a 4-month period, again. In the post-clinical trial we added more sites to the study and it lasted -- so far the data includes 12 months, approximately, and we have 373 additional donations that have been interdicted based in that same category, for a total of 448 since we've implemented PCR testing. And again, this is through testing of the 3-dimensional plasma pool of 512.
HIV, reflecting the smaller incidence, we found six donations that were not detected either by p24 antigen testing or antibody testing. And those were also interdicted that would not have been otherwise.
This overhead shows data collected during the clinical trial. For HCV donors it demonstrates sero conversion periods. The chart on the left here shows negativity. The green bars are PCR and antibody testing; the gray are only antibody testing.
And on the right side, the blue bars depict PCR-only positivity and the red caps show sero conversion. And in some of these instances donors who were not enrolled were not followed to sero conversion. In one instance we had an individual who, when sero converting, HCV was not picked up by -- or, HCV RNA was not picked up by PCR in the pool.
The two donors best demonstrate the sero conversion range. This donor here, number 3, shows approximately 93 days of PCR-positive donations in the absence of antibody, and also shows a negative PCR result and antibody result of course, immediately prior to the PCR detection.
This individual here is similar; however that period of time is only 28 days before that individual sero-converted.
These are the four HIV subjects; again, similar. The blue bar again, is PCR positivity in the absence of any other maker and the yellow depicts PCR and p24 antigen positivity. And the red cap again shows sero conversion. Here we had a range of eight to 20 days.
And I want to emphasize, that period represents the period from the last sero-negative result. So they could have sero-converted anywhere here, or this individual could have sero-converted anywhere within this range. So we are counting from the point of the last test that was antibody-negative from the first PCR.
We observed an interesting phenomena through the course of HCV testing. This slide represents HCV antibody reactivity in percent ranging -- and the red line depicts when PCR pool testing was implemented.
As you can see, the green band shows the actual reactive rate; the blue band is a rolling average before and after PCR implementation. And you can see a significant drop in the reactivity rate of HCV antibody, suggesting perhaps the donor population is being cleared by the PCR tests before the sero conversion.
Now, as Dr. Conrad had mentioned, we have done some studies of -- comparison head-to-head studies with p24 antigen testing against HIV PCR in a pool of 512 donations. When we tested the samples with p24 we used both the culture and the added license test, and if either of them produced a positive result then they were counted as positive.
This contingency table shows that 61 of these samples were found positive by both PCR and the antigen tests. None of the samples that were positive by p24 were negative by PCR. And 32 of the samples that were negative by p24 were positive by PCR.
So we show at least equivalency to the p24 tests when testing PCR in a pool of 512 against p24 in single donation.
So to sum up, we are concluding that PCR pool testing decreases the viral load in the manufacturing pool; that PCR pool testing also provides an opportunity for the infected donor to seek early treatment; and HIV PCR is at least as effective as p24 antigen testing in screening out window period source plasma donors.
Thank you.
CHAIRPERSON TABOR: The next speaker is Dr. Cushing from Baxter.
Return to Table of Contents
DR. CUSHING: I'd like to thank everyone for inviting us here today to talk about the data from the clinical trial.
This study was conducted in two phases. The first phase was the screening phase and the study duration was approximately three months. We began it on December of '97 and ended in March. The plasma donor qualifications for this study to be PCR tested were to be EIA-negative and also, unlike Alpha's study, to be p24-negative.
We had 46 plasma centers participating in this study, and the number of plasma donations PCR tested were approximately 345,000 donations. And the number of plasma donors tested was approximately 42,800. And the number of master pools tested was approximately 675, and we had a mean donation per donor in this study of eight.
Phase II of the study was called the sero conversion follow-up phase. And the duration of this phase was six months as a maximum or sero conversion. And out of the 46 centers we had ten centers that had eligible donors.
And eligible donors for this phase were found to be HCV or HIV PCR-positive and antibody negative. In the case of HIV it would also be p24-negative. Donors that were enrolled in this phase were tested by PCR and antibody testing on a weekly basis.
One of the things that we did in this was to add blinded samples to our master pools. We had 50 control samples that were both positive and negative samples for HIV and HCV. And of these 50 samples, all 50 of them were correctly identified by PCR.
We had zero HIV PCR donors who were eligible for enrollment in our Phase II study. We had 22 HCV PCR donors eligible for enrollment in the Phase II study. And of these eligible donors, five of them declined to enroll. And other four sero converted prior to enrollment in the study.
We had six donors who did enroll in the study and sero converted within six months. And we had two eligible donors who enrolled in the study and did not sero convert in the 6-month period but they were found consistently to be PCR-positive. And I'll speak more about those donors later.
Also, I told you that we had 22 donors who were eligible to enroll. Five of these we found to be false positives or eligible donors whose initial positive HCV PCR test result could not be confirmed upon repeat testing.
And we feel that the probably root cause for this was contamination during pooling on the TECAN -- and we think that three of the cases were due to that. And we think the other two were due to contamination at the donor center during sampling.
The corrective action that was taken was to upgrade the TECAN pooling software -- and Dr. Conrad talked about that. We've also made -- and it's not conformation, it should be confirmation -- procedural changes. Initially we sent in the sample for confirmation that had been on the TECAN and now we sent in a backup sample, so we think that this will control for the contamination due to pooling on the TECAN.
And we've also implemented a new sampling methodology. We were using the cut-and-drip method; now all of the centers are using a closed method with vacuum tubes.
From our study we did have ten HCV PCR sero converting donors. And we found that the range of time from the first positive HCV test -- PCR test -- to sero conversion -- and again, this isn't infection but this is the first PCR-positive test -- was anywhere between 30 and 115 days.
And we found that the range of time from the first HCV PCR test to deferable ALT levels defined as two times the high normal, or an ALT of greater than 76, was anywhere between 30 and 101 days.
And HCV-positive donors who exhibited deferable ALT levels prior to sero conversion was three out of our ten. We had two out of the ten donors who sero converted -- I don't know, actually we had two of the ten donors who never did exhibit elevated ALT levels but they did sero convert.
And this is a graph that demonstrates the relationship between sero conversion and elevated ALT levels. And you can see that there is a rough correlation but it isn't -- you can't use it exclusively to predict. And of course the two individuals who never did have elevated ALT levels are not even represented on this graph.
This is a graph that -- it's a little difficult to tell but it looks at what the PCR or the viral load was of five of our donors where we knew when their first PCR-positive test was and time to sero convert.
We did a quantitative PCR analysis on these donors and what we found is that all of them started with a fairly low level. Now this is a logarithmic scale so even this fairly low level could be 10,000 or more. And fairly rapidly the levels increased, and within three days you can see some of them going way up.
Now, some of these donors' levels, viral load, stayed elevated the entire time to sero conversion; whereas others -- this pink one was kind of all over the place. And the donor represented by green, his levels dropped way down. So it does seem to be extremely variable.
Of our ten sero converting donors we found a male/female ratio of ten males and zero females. And our mean female -- or rather, our mean male to female donor ratio was 68 to 32. And we found no correlation with the number of prior donations and HCV infection as defined by the PCR analysis.
You can see out of our ten donors that we only had one donor who was a first-time donor. We had two donors who had donated between one and five times; one donor who had donated between six and ten times; one donated between 11 and 20 times; one between 21 and 40 times; one donated between 41 and 60 times; and we had three donors who had donated between 61 and 80 times.
Okay, now these are the findings that we have for the HCV PCR-positive, non-sero converting donors that I talked about before. Here the male/female ratio was zero males and there were two females.
The very first time they came to these centers their PCR tests result were positive. So we have no idea when they first became infected. The duration of HCV PCR-positive tests was greater than 176 days -- which was the length of our trial.
Their viral loads by quantitative PCR during this time ranged anywhere between 140,000 and 4.5 million copies per ml. The duration of HCV branch chain DNA-positive tests correlated with the PCR-positive test and was also positive for greater than 176 days.
Now the duration of normal ALT levels was greater than 176 days, so at no time did they ever exhibit a deferable ALT level. And the duration of negative antibody tests result using the Abbott HCV EIA-2.0 test was negative the entire time.
We then went back and performed the ortho HCV EIA-3.0 test, and here we found that the results of this antibody test was positive, and it was positive from the very beginning of their PCR-positive tests.
I mean, I don't think that we can say anything really, from this about the validity of the two tests because we didn't set out to check that, and so we certainly have not looked at the reverse to see if that's true. But certainly, given this system I think what we can say is that there was nothing that we routinely would have done that would have picked up this infection in these two individuals.
Also, in the course of performing the HCV PCR test on our sero converting donors we found that there was a transient negative PCR result following an HCV PCR-positive result of three out of 78 PCR tests. And we think that possibly the cause for this is the formation of an EIA undetected antibody that mediates viral clearance.
We did do quantitative PCR analysis in this case, and it did come out as undetectable. And the quantitative PCR values preceding or immediately following the negative ones were extremely low.
So in summary, what we've done is demonstrate the ability of PCR testing of pooled plasma samples to detect HCV-infected donors; we validated the process of pooling donor samples, testing the pooled samples by PCR and linking donations positive by PCR testing back to the donor; and we have collected data on the sero conversion of donors who are positive by PCR testing and non-reactive by HCV antibody testing.
And I'd just like to thank all of my colleagues at Baxter who were involved in this, and the colleagues at NGI and ICRC, our contract research organization. Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker is Dr. Dragon from Roche.
Return to Table of Contents
DR. DRAGON: Good morning. I'm going to present today and describe, the Roche Molecular Systems products for nucleic acid testing for HCV and plasma donations. I'm going to be speaking about current reagents that are either about to be on the market or are in some countries, perhaps available already. I will not be talking about future, single unit screening at this time.
To give you a brief history of our collaboration with Bayer Corporation, we each filed a joint IND in July of 1997 to begin the work looking at pooled specimens. We then had Bayer begin testing in September of 1997, and just this month we have actually filed amendments for our IND to upgrade the tests from the Version 1 reagents to the Version 2 reagents.
We are calling our blood screening products AmpliScreen as to contrast them to the diagnostic products -- the Amplicor products -- and I'll describe some of these differences.
We have two platforms that we are supporting. You will see data from Bayer on the microwell plate format where we are working with pools of 96 with reflex testing and pools of 8 X 12 followed by single resolution.
In addition, we are beginning collaborations where we will be working on our automated COBAS Amplicor platform where we will be working with pools of 24 with subsequent, secondary resolution of pools of 4 X 6 and single resolution after that.
The most significant improvements between the Version 1 and the Version 2 tests are twofold. One is the ability to equivalently amplify all the known genotypes of HCV.
We have kept our primers constant; however we have put co-solvents into the master mix. There's been a slight master mix formulation change which has resulted in this higher efficiency of amplification for all the genotypes.
Secondarily, we have an improved sensitivity due to procedural changes in our specimen preparation procedure, including a larger input volume.
Briefly, to just look at the summation here, the most important things to see are that in the first-generation kit, the microwell plate kit, we did have an internal control but it only controlled amplification and detection. It was not included in the specimen preparation.
In the new methodology we now actually have extracted internal controls, either with the microwell plate or the COBAS Amplicor system. In addition, we used to start with 100 microliters of specimen. We now have gone to a one ml input sample volume. We used to resuspend in one ml; we are now resuspending in 200 microliters.
The most important difference as I said, is a procedural difference. We're using what we now call the multiprep specimen processing where we'll take a one ml specimen -- in this case from the pools -- and we do a one hour spin at 23,600 times g. Following that spin we will draw off 900 microliters of the supernatant, leaving 100 microliters still in the tube.
We will then add the quanidinium lysis buffer spiked with the QS or IC control; mix, incubate so we have viral lysis going on at this point in the presence of the internal control. We'll then follow with a isopropanol precipitation, pellet the virus, do a wash in 70 percent ethanol, resuspend in the specimen diluent, and then proceed with amplification and detection.
I'd like to share with you some of the results that we have from in-house on the non-clinical performance characteristics of this test. It will compare both microwell plate and the COBAS Amplicor work we've done.
The first test was working with the WHO International Standard. And we prepared using the standard which started out at about 50,000 international units per ml. We diluted this into a negative HCV plasma and we brought it down to either 200, 150, 25, 15, 12, or 10 IUs per ml of HCV.
We then did 24 independent specimen preparation amplification and detections to look at sensitivity levels and reproducibility.
And what you will see is that with the microwell plate tests as we've described, we had 100 percent hit rate at 25 IUs per ml. Interestingly you'll see that we have only 23 of 23 indicated here. In that particular reaction we had one invalid result which was an IC-negative result. So again, the internal control acts as a flag if things are working or not in the mixture.
For the COBAS we've actually been able to achieve again 100 percent at 25 IUs per ml and greater than 95 percent sensitivity at 15 IUs per ml.
A second international standard that we have looked at is the NIBSC working reagent, and in this case we're looking at -- we've been able to get down to 100 percent to 70 genome equivalents per ml, and at 60 we're still at greater than 95 percent.
With initial results that we have looked we feel that one Roche copy is equivalent to approximately 1.4 IUs per ml. We're looking at about a 4-fold difference between what the NIBSC standard is and what the IU is, which I think is similar to what Dr. Saldanha has seen and reported in the past.
Genotypes are very important. We've looked at genotypes from two perspectives. We've looked at them from actually real clinical specimens that have been documented to be a variety of genotypes and then do serial dilutions with them looking for analytical sensitivity.
We too, also have cloned copies of the different genotype sequences at the 5 prime untranslated region and have done limit detection for various input copies numbers of genomes. And you can see that we are five copies per PCR where it's still the 100 percent hit rate.
In conclusion, I've just presented to you some of the non-clinical performance data that we are starting to generate within Roche for the Version 2.0 kit. My colleague from Bayer will present some of the clinical studies that have gone on.
The important things to note is that we are looking at now with the Version 2.0 kit, a sensitivity of less than 50 international units per ml which is well below the 100 IUs that was recommended by the Agency to be able to use for plasma screening.
We have picked up the remaining of the genotypes that we did miss in the first-generation test. We also still have our own sterilization capabilities working in there.
Finally, we have performed interfering studies -- interfering substance studies with a variety, and we have no known inhibitors. We have also validated both EDTA and ACD as the collection tubes of method.
Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker is Barbara Masecar from Bayer.
Return to Table of Contents
MS. MASECAR: Good morning and I appreciate this opportunity to provide you with an update of the current status of nucleic acid testing for HCV in plasma donors being performed by Bayer Corporation at our facilities in Raleigh and Clayton, North Carolina.
Briefly a timeline. This replicates very closely the slide that Betty showed. We jointly filed an IND in July of 1997. Testing was initiated at our PCR laboratory in Raleigh in December of 19 -- I'm sorry, September of 1997.
We began testing with the version 1.1B test kit using the Amplicor Version 1 reagent with a modified specimen preparation to give us added sensitivity. And that has been in-place in use since September 1997 to the present.
And the initial validation work done to support the IND filing in July was based primarily on the NIBSC working run control.
As far as the test system and work flow that we use at our laboratory -- this has been presented at other meetings -- but our samples are received, batched and decapped. Our batches are in multiples of 96 and an 8 X 12 array. We use the Hamilton AT plus to create the mini-pools of 96.
If that is positive we go to the reflex testing of the rows and columns, and then the individual positive is identified by the intersect. And this is a very basic graphic representation of the rack if we test the column samples and the row samples.
In this example column 7 is positive, row F is positive. The only sample shared in those two mini-pools is number 67 so we've identified our positive.
We are anticipating very soon the transition to the Roche Version 2.0 test. We have filed our amendment this week, both Roche and Bayer, and we will transition to Version 2.0 next month. There will be essentially no change in the test system or the work flow. We still maintain our mini-pools of 96.
The version 2.0 does incorporate an ultracentrification step. This is the multiprep method that Dr. Dragon referred to. We don't have this in here really to affect the hepatitis C sensitivity necessarily, but it really is in anticipation of adding HIV testing next year.
And now that we have the WHO International Standard available the validation work for Version 2 was based on that standard. And that's what I'd like to present.
The validation of Version 2.0 that was filed in our amendment was performed according to ICH guidelines. We concentrated in two major areas: the limit of detection by limiting dilution using the WHO standards, and then after that was performed standardizing our own in-house control in international units -- which is a positive plasma unit that we have, high titered.
And then after we determined what the limit of detection was for the assay and we had our in-house controls calibrated, we used a less dilute, in-house control to input into the total system to include the creation of the mini-pools to verify that operationally our assay would meet the validation requirements.
The limit of detection was performed simply by the creating of dilution panel to bracket the detection limit of the assay. And we performed this with three operators with not less than 16 tests per dilution.
And the graph on the next slide shows the data represented. At 50 international units we had 96 percent positive -- which is represented by this datapoint. At 25 international units we had 90 percent positive, and so we just did a simple linear interpolation. So we determined that at 95 percent test positivity, Version 2 using the multiprep, our limited detection is 46 international units per ml.
Then we did the same set of experiments with the in-house controls. This in-house material had been previously quantitated several times with the Chiron bDNA assay. And we diluted the in-house control to get the 95 percent test positivity rate and then assigned that the 46 international units per ml level.
The operational validation for our total test system, we took our in-house control and from the stock of that we diluted that to approximately 4400 to 4600 international units per ml. And that would represent the detection limit times the 96 dilution factor that we see in creation of our mini-pools.
The total test system was employed -- real-life situation, real techs, real pipetters, real lab -- and we found that in 144 tests we were positive 137 times, which gave us a 95.1 percent positive, which agrees quite well with our initial validation.
A limited number of these that were positive were carried through to the row and column test and they were correctly identified in a blinded fashion in every case and they were all positive.
Briefly, as Dr. Dragon mentioned, Roche determined the limited detection for Version 2.0 in a similar manner. They were positive 100 percent of the time at 25 international units, and their 95 percent test positivity rate was between 20 and 25 international units.
Regarding the correlation of unitage, in our Bayer validation studies using our in-house control that itself was directly quantitated with Chiron bDNA, we found that one international unit was equivalent to approximately two genome equivalents.
And as mentioned by Dr. Dragon, in Roche validation studies, one international unit was equivalent to approximately 4.5 genome equivalence. However, the Roche Version 2 Amplicor monitor assay was used as an intermediate step here, so that could affect the difference. And also we're talking a single lot of Version 2 that was used.
Regarding the overall performance of the Version 2.0 assay, to-date there has been excellent agreement between independent validation studies performed at the Roche and Bayer laboratories.
Our in-house control has been calibrated. We will run a 100 IU control with each test. And the performance of this control taken with the validated detection limit of Version 2 meets and exceeds CPMP recommendation for test sensitivity, and also what FDA has stated.
Brief clinical study update: we are ongoing with our clinical study; it has been ongoing since testing began last September. It supported the amendment. All initially positive clinical study samples that were discovered with Version 1.1B were polled from ultra-low, freezer storage, diluted 1:96 and retested, and all were found positive with version 2.0.
We anticipate continuation of our clinical study for one year or less from the time of transition. We have waited to do the repository samples until we are fully transitioned to Version 2.0, so that will be a major goal.
And we anticipate that data accumulated to-date with the Version 1.1B will be submitted in support of Version 2.0 kit licensure.
Future directions: we expect to implement HIV PCR testing in the first quarter of 1999. We already have the specimen prep in place so this will be a shared sample prep with HCV. And then implementation of a third target as yet not yet named, in the fourth quarter of 1999.
In conclusion I'd like to acknowledge some people that provided data for this talk: Rita Sun and her group at Roche Molecular Systems who performed the validation work; Todd Gierman and Michael Gray at Bayer Corporation for the validations that were performed at Bayer. Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker will be Dr. Watson from Centeon.
Return to Table of Contents
DR. WATSON: Good morning. Thank you for the invitation to present our results, however preliminary they are. When contacted about speaking today I was asked to concentrate on our results and not our methodology. If anybody wants to know about the methodology that was presented by Dr. Weimer that Blood Safety Meeting in February. I will go into a little bit of it just so you can see the differences in the systems.
So this is our initial report. We concentrated on the results that we have and the very preliminary clinical results from our sub-study.
Our IND is basically the same as just about everybody else. We want to see if pool testing can identify potentially more infectious units and remove them from production.
Our IND started -- not started -- we received authorization to proceed in February this year. We started collecting samples under the IND in April and we started our testing in May. So what I'm going to present in May is a report on the first 90 days' results.
We have a deferral policy. Any donor who is PCR-positive will be deferred, both in the corporate registry and the national registry. And we have a look-back/look-forward policy which I'll explain. I'll go into some of these numbers that we actually have units that we've interdicted under that policy.
Originally we started with a 3-month look-back. We've had to expand that to six months. And we also look forward because it takes time to do the PCR test. So from the time the PCR test, that unit result is available, all the other units that were given subsequent to that are also removed.
Our system tests for three viruses. All samples have to be -- all units have to be serology-negative. They have to be p24 negative. They have to be ALT negative. If a unit is ALT-positive we will not test that unit but we will test all the other units from that donor that were ALT-negative.
The virus load in this is not test sensitivity. This is the concentration in genome equivalence in a unit that needs to be present for us to find a positive PCR result 95 percent of the time. For HBV we used EuroHep; for HCV we used Pelispy; and for HIV we used the preliminary working standard.
Our pool size is 1,200 to begin with; our maxi-pool. We then have an intermediate pool of 120, then we work back to 12, and then we go to the individual donor. Twelve works well for us because that's our logistics system.
Okay, here's our results: 600,000 samples; five HBV-positive donors; 36 HCV; and one HIV. The HIV was interesting. It was his 13th donation since November, but unfortunately we haven't been able to find him.
By the way, these numbers include both donor applicant and qualified donor because we test everything. Now, we have not had an opportunity -- what you're getting here is raw data. We have not been able to sit down and analyze anything yet. Our IND is for one year. There's no 300,000 cutoff or particular timeframe less than a year; we decided to go with one year so we're going to have a lot more data to analyze.
Average number of units -- actually, in reviewing this this morning I realized that this slide is wrong. That's the average number of units identified during the period of time for look-forward and look-back.
If we look at a 60-day timeframe and we say that that 60-day timeframe where we would have the units within the company, that would be -- at a minimum, that would be 40 percent. So you're looking at 260 units in a 3-month timeframe that we've stopped from going to production from donors that were PCR-positive. Our donors do come back multiple times.
We have clinical testing -- we call it a sub-study -- eligibility, PCR-positive. What we're going to do is, we're going to have bring them back once a month, we're going to do a PCR test, we're going to do all the relevant serology testing, and we're going to continue that for a maximum of six months, or until sero conversion -- or, if they're rather interesting, maybe we'll extend it on them.
We're just in the beginning of that, since we just started our testing. By the time you notify the donor and try and locate him there is some time that goes along.
And the last slide shows what we've done as far as enrollment and what we've been able to test. So far we've only enrolled one HBV donor, and hot off the result press yesterday morning came another PCR-positive. We do not have the serology result yet. My guess is, we'll get that later this week.
Three HCV donors have enrolled. We've tested two of them. For their initial sample both were PCR-positive. They both continue to be antibody-negative. We have a third one that recently enrolled. We just received the sample earlier this week, and we're now beginning to get the second samples on the other two donors.
And the HIV, we're looking but we can't find him. And that's the results that we have. Thank you very much.
CHAIRPERSON TABOR: Thank you. Our next speaker is Dr. Giachetti from Gen-Probe.
Return to Table of Contents
DR. GIACHETTI: Okay, first I'd like to thank the organizers for inviting me to give this presentation. I will talk today about Gen-Probe's high throughput assay for the single strand of detection of HIV and HCV RNA. My presentation will focus mostly on the methodology we'll use as well as results with HCV (unintelligible) as well as analytical sensitivity.
These objectives are to have an analytical sensitivity for both targets, HIV and HCV, of 100 copies per ml. Here to demonstrate detection of infection before sero conversion, detection of all non-subtypes of HCV and HIV, able to incorporate an internal control and in the future other targets, analytical specificity higher than 99.5 percent, show non-cross reactivity with other blood borne pathogens, and have throughput and amenable to automation.
Assay protocol is simplified here. Our technology uses an integrated approach where sample processing, amplification, and detection are all performed in the same tube. This simplified test steps reduce the chance for contamination, and one operator is able to process 200 samples in six hours.
First step of the assay sample preparation, we use target capture and magnetic tactical separation. First step is to lyse the virus. We use 500 microliters of a specimen. During the lysis the plasma or serum is mixed with the lysis buffer which contains captured probes as well as micro particles and internal control.
During this step the viral RNA hybridized to a captured probe that is a chimeric sequence that will be complimentary to the viral RNA as well as contained in polidity tail, and these polidity tails will attached to a polidity tail that is attached to the macro particle.
We use magnetic racks to be able then to wash away the plasma and all potential inhibitory substances. These render the samples ready to amplify in the same tube.
Next step is the amplification. We use transcription mediated amplifications. This is an isothermal reaction, utilizes two enzymes -- reverse transcriptase as well as T7 polymerase -- can be used for RNA or DNA targets, and produces an RNA amplicon. The reaction is exponential for using more than one billion-fold amplification in less than one hour.
For detection you use the hydrization protection assay. This assay utilizes acridinium ester probes. The reaction consists in three steps. First step is a hydrolization in solution where the level probe hydrolyzed to the viral RNA.
Second step is a selection where the level on the unhydrolyzed probe is hydrolyzed as shown here, where the label on the probe that has been hydrolyzed to the amplicon is protected. The third step is detection where the level on the protected hydrolyzed probe is detected by chemiluminescence.
In order for us to be able to incorporate an internal control in each sample and to be able to detect an internal control without the need to separate a reaction into different vessels, we use the two kinetic analysis.
This is an application of the hydrolyzation protection assay that utilizes acridinium ester level probes with the (unintelligible) kinetics of light off. We have different type of probes: probes with very high kinetics of light off that we call pleasure probes, that hydrolyze to our internal control; and probe with the slow kinetics of light off, which we call lower probes, that hydrolyze to our targets.
So each reaction would have two results: one is for the internal control that validates a reaction in each specific tube; and the second result for each sample is the presence or absence of a target.
Next I will show data about analytical sensitivity and clinical sensitivity for HCV detection. We determined our analytical sensitivity preparing dilutional panels which were calibrated using the international standard provided by Dr. Saldanha using a conversion factor of 2.3 to transfer genomic equivalence per ml.
With the different operators, different days, several lots of reagents, it all totaled about 120 replicas of each level. And our sensitivity is 100 percent at 100 genomic equivalence per ml, 92 percent at 30 genomic equivalence per ml, and we go down to 12 percent at one genomic equivalence per ml.
We also participated in several studies. This is a study organized by John Saldanha this year. When we compared our detection limit using different reference standards, here if we use the WHO reference standard we have detection of 100 percent at 10 international units per ml, and 93 percent on three international units per ml.
Part of this study compared also the CBER FDA standard, the Paul Ehrlich standard, and the Pelispy standard. The CBER standard as well as the Paul Ehrlich Institute, the standard runs very close. We have 100 percent detection at 100 genomic equivalence per ml, 70 percent detection at ten, 40 percent at three, and we go 12 percent at one.
These correspond very well with the previous data that I showed with our in-house panels. Pelispy also would have good sensitivity showing 100 percent at 360 genomic equivalence per ml, 95 percent at 76 genomic equivalence per ml.
To explore the application of these assays not only to single dot units but also to pools -- and this is part of the collaboration (unintelligible) -- we tested several sero conversion panels that were obtained from BBI, Nabi, as well as Bioclinical Partners.
And we tested the different panel members at undiluted samples or in a pool diluted fashion. What we found here is that the days of detection before sero conversion is very similar in all the cases.
I should mention, and here shown by (unintelligible), that unfortunately most of these panels we were RNA-positive in the first bleed of the panels. So if there were differences between the undiluted/full diluted samples we are not able to detect it.
Here is more panels, the (unintelligible) panels. And this is the only panel out of 20 or 21 that we tested that show a difference between the undiluted sample and the full diluted sample. In this case was 35 days before sero conversion and here was 28 days.
Here are data from the Nabi panels. We have a broad difference in detection limits going from 60 days down to zero.
Conclusions. I will not show specificity data which show the specificity of the reaction. We haven't found any cross reactivity with other infectious agents or autoimmune conditions of heart or liver diseases. We haven't found any interfering substances so far. It works very well with different kinds of anticoagulants or serum or a percent of problematic samples.
The sample processing method removes potential interfering substances. It's very efficient. And we found internal control very useful for confirmation of amplification performance in each sample.
Sensitivity conclusions are here. We have 100 percent detection at 100 genomic equivalence per ml, and 50 percent detection at four genomic equivalence per ml. We were able to demonstrate HCV RNA detection on average, 39 days before sero conversion. I haven't shown, but we have demonstrated also very sensitive detection of all genotypes from 1 to 5. We haven't tested 6; that's why it's not there.
Finally, I'd like to acknowledge that this project has been funded in part with Federal funds from the National Institute of Heart, Lung, and Blood, and also in the future, Karem Corporation will share in our efforts. Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker will be Dr. Stramer from the American Red Cross.
Return to Table of Contents
DR. STRAMER: Thank you very much. Initially I was told that I could have Dave Gretch's 15 minutes so I had intended to use it, but I'll just speak faster now that everyone else has seemed to have used it.
I was asked to present data on information we've collected at the American Red Cross thus far using pooled GAT testing, and also talk about our future implementation plans. I will first go through the data that we have collected initially in collaborative studies with Andy Conrad and Rich Smith at National Genetics.
These are some of the parameters that one must consider when looking at pooled GAT testing, and I will go highlight some of these studies that we have done, again, in collaboration with NGI.
We will be collecting samples -- these are whole blood samples now, in contrast to the previous speakers. We talked about plasma only for further manufacture. In the whole blood sector we obviously have to deal with the issues of red cells and platelets as well as plasma.
We will be collecting samples in a closed tube container to minimize transfers -- minimize contamination from transfers. This is a plastic tube that may be frozen. It is a gel separator. It's called a plasma preparation tube, manufactured by Becton Dickinson.
It is spray-coated with EDTA and as you've heard references by speakers in the past, both EDTA and ACD are probably acceptable anticoagulants. Serum is not an acceptable sample for GAT testing; at least in our hand.
Looking at stability in two phases -- because again, we collect whole blood -- we've had to do a whole blood component to our stability studies and a plasma component to our stability studies. Since most of the samples are antibody-positive, one must distinguish between stability in antibody-negative and antibody-positives because frequently you will see differences.
These studies were carried out at two different temperatures to understand what temperature we have to do shipping at following collection. This is a whole blood, pre-sero conversion series, and you can see over the entire course of 72 hours we didn't really see any major decreases in data.
The blue color here represents the lower temperatures: 6 to 10 degrees which is the upper end of our shipping validation versus elevated temperatures, ambient temperatures, 25 to 30 degrees. We will ship at 6 to 10 degrees because it does afford us higher recovery and better stability as has been reported in the literature for both HIV and HCV RNA.
These studies again, are HCV. We have not yet completed our HIV studies.
Looking at the same series post-sero conversion -- or in the case of antibody-positives -- you basically see the same phenomenon. At the lower temperature, 6 to 10 degrees, you see higher recovery but you really don't see a difference in rates of decline.
You do have lower recovery in perhaps a more significant -- well, they're both significantly different in this case, but less recovery at least, in the higher temperatures. So again, we will be going with 6 to 10 degrees or shipments on ice.
Following separation in the PPT, which the gel separator separates the cellular components from the plasma above the plug following centrifugation, these are the results of our combined plasma stability studies looking at combinations of whole blood at 6 to 10 degrees, whole blood at room temperature, plasma at low temperatures, ambient temperatures, low temperatures and high temperatures.
But again, as has been reproduced in all of our studies, we see the highest recovery, best stability long-term when we use the combinations of the low temperatures for both whole blood and plasma.
And from the results of our study we will be recommending a 3-day whole blood ice shipment followed by a 7-day plasma time on ice with no greater than 24 hours at room temperature for the entire process of collection, shipment, pooling and testing -- including resolution testing.
Converting now to sensitivity data, again, these are data that we collected with NGI looking at window period reductions, if you look at antibody positives collected from this serial plasma donor. So this again, is the antibody-positive curve and this is the RNA-positive curve.
If you look at the time difference between the first antibody-positive -- here's the antibody cutoff, here's the first antibody-positive. So the first PCR-positive here, you can see that there's a window period closure of 26 days.
In our studies with NGI we use pools of 500, so using a cutoff of their pooled PCR tests here you can see even with a pooled PCR test we do get significant window period reduction even using pools as large as 50 donations.
Here's another series in which you can see the same effect. Here there's a 40-day window period reduction between the first antibody-positive and the first GAT-positive. And you can see here high viral load. So the important features to take home from HCV RNA sensitivity studies is long window period and high viral load.
If you look at a combination of all the plasma series we looked at in our studies with NGI, which were 19, this shows you the different phases during sero conversion. Time here indicated on the X-axis, the cutoff for the PCR and the viral load during different phases of sero conversion.
So here's the first phase, pre-antibody positivity. So this is the RNA-only samples. And the mean copy number here was greater than five million copies per ml. So really pool size is really not a factor of important consideration when you're doing HCV because again, viral titers are so high.
Looking at some specificity studies we did using crossover matrices, we did a study of 20,000 unlinked donations in which we removed the sero reactive samples. From our studies looking at these agents -- HIV was not included -- what I want to highlight here are really three findings.
We did have, for HBV DNA and HCV RNA, two unresolved pools of this study. And in combination, if you look at these unresolved pool results it would have held up release of 8.1 percent of blood products. So that was an unacceptable outcome of the study.
But interestingly enough we did find one HCV-positive which was a unique genotype to the spiked samples that we included in the study. The study did include spiked samples of genotype 1A and our finding was a genotype 3A.
It was sero negative by all tests; both the EIAs and the HCV RIBA-3, our strip immuno assay. I mentioned it was genotype 1A -- where's our spiking sample -- was 1A. And because this study was done in serum we believed we had a low viral recovery because the samples were handled in serum with long, ambient temperature storage times.
I mentioned that all the spikes were correctly identified. I just wanted to acknowledge that even though we used a very complicated pooling system which I'll briefly discuss, NGI did an excellent job of resolving, identifying to the correct location, all positive spikes.
Interestingly enough, relative to our HBV DNA viral input, we've recovered all DNA. As we know DNA is a lot more stable than RNA, but in the cases of HCV RNA we only could recover 23 percent of our viral inputs; again, because we did the study in serum.
We used a 2-dimensional matrix which has been discussed at least twice -- one for each dimension. And as we talked about, by pooling in an X-direction and a Y-direction testing the results in pools, if you have a positive pool in this dimension and this dimension, the only unique donation represented by those two reactive pools would be in the inner section.
We used this approach in a 2-tiered fashion such that four reactive pools were required to identify two reactive sub-pools to identify one reactive donation. Be that as it may, the two yellow boxes here represent our spiked sample.
Interestingly enough -- I just said you would need four pools to identify donation. Here we had these two columns as positive -- that's A5 and A6; we had two rows reactive, B1 and B6, which pointed to the two sub-pools which implicated the one spiked sample.
However, we also had an additional reactive pool at B8. The inner section did not contain a reactive donation so by the definitions that I and others have told you, you would assume that this would have been an aberrant result. But because of the way we designed the study we were required then, to test all of the sub-pools.
We found another sub-pool reactive that corresponded to this one-fifth reactive row I showed you. But his column was falsely negative. I also mentioned to you that you need two primary pools to implicate a single donation. So we identified finally -- this was the sister positive to this sample to identify the unique donation that was at genotype 3A.
So here, this positive was masked because it was in the same column with another positive, and here we had another false negative result. The reason that these two false negative results could have occurred is because again, the sample study was done in serum and perhaps we did have RNA degradation.
Moving forward, what we will be doing is, we not only would like to do the pooling but we will be doing the testing. So we really need to implement a kit that we can test on-site. Again, because we're dealing with plasma red cells and platelets, turnaround time for us is extremely critical.
So what we are now endeavoring in is a joint IND with Gen-Probe as the technology was just described by Christina Giachetti. We will be doing pipetting and testing, we will control CGMP which gives us the flexibility of using backup laboratories. We will be evolving to decrease pool sizes.
I mentioned that we started at pool sizes of about 500. Now we will be using pool sizes of 128 and in order to decrease turnaround time and be able to release red cells and platelets at the same time we release plasma, we probably will have to evolve to smaller pool sizes.
We are using a multiplex test, the Gen-Probe test. It's semi-automated; the features have been described to you. One advantage of doing this type of testing -- if we can have the turnaround time impact the release of all products -- that is, cellular and plasma products -- you've seen the data that perhaps testing that we do today such as p24 antigen and ALT, could be removed since they would not have efficacy in those cases.
So in our joint IND with Gen-Probe, what our goal is as others have described, is to evaluate the efficacy, feasibility, and performance characteristic of the multiplex, transcription, mediated amplification assay, and to meet the European requirements for GAT-tested plasma for further manufacture by the July 1, 1999 CPMP deadline.
But what's critical to us in the whole blood industry is that we must initiate testing in a way that does not compromise the availability of blood, but generates information and support of the eventual GAT-based control of labile products as well as plasma-derived products.
I mentioned that we're doing a 2-part IND. This requires IRB approval; our clinical data to be collected and provided to FDA validating our specific intended use. We will evaluate and plan to meet all the international standards that are required; that is, the CPMP 100 international unit per ml standard.
We plan to also meet the PEI standard and the FDA standard that has been tentatively set at 100 copies per ml.
When we're dealing with whole blood there are factors here that are much more complex than dealing in the plasma arena, and it's complicated from the point of donor all the way through the patient. And if you don't think each of these areas are affected one should think again because they certainly are. And I will highlight some of those.
Due to the complexity of our system we're managing our entire project very centrally with a centralized team that makes all management decisions. We will operate off a single set of procedures; one laboratory. We will have a centralized information flow because we're using code 128 whole blood numbers.
We will plan to implement at all locations, all allogenic donors. We will use as a specific sample type the plasma preparation tube. This will go to the specific lab for a Gen-Probe test. Other tubes will go for serology. We will be removing the HIV1 and HCV serologic samples so that they don't contaminate pools.
I will end with this slide, our first phase will be under IND as an evaluative phase. We will use conservative policies to assess logistics; that is, false positive rates, turnaround time. It will include one million donations.
We will then hopefully move to a second phase in which we can make some better management decisions involving red cells, and instead of reacting to a single donation as we plan to do now after the resolution of pools, we then would be able to respond to a single, reactive pool. And again, then we would proceed to including the entire system.
Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker is Dr. Zerlauth from Immuno.
Return to Table of Contents
DR. ZERLAUTH: Thank you for the opportunity to share with you a few of our results that we have obtained by using PCR in a routine screening scenario. I do not have any IND-related data yet because we have not yet filed an IND but we are in the process of compiling one and should be submitted very soon; probably this month.
I would like to go briefly through our IQ-PCR system, the PCR assay system that we have in place. We have the classical extraction on guanidinium isothiocyanate and phenol chloroform/ethanol precipitation.
We have a single round PCR; it's not nested. It's single round. We are using RTTH. We are using hot start and we are using fluorescent-labeled primers, and we are using an internal control which I am going to show you in a second.
We have size-specific separation of our amplicons and polyacludamide gels. And we are detecting those size-separated amplicons based on their fluorescence, by a laser-induced fluorescent equipment.
Our internal standards which we have called internal quality markers, are almost identical to the target we are interested in. We just have modified the lengths of the amplicons to be created. We have added a few bases -- a few mean 5 to 11 bases -- that we have added to the original sequence.
We then have transcribed it into RNA for HIV and HCV, and we used plasmids for the HBV. So we have not altered the primer binding sites, only the lengths of the target sequence. We now add those constructs into each vial so the internal control goes along with each vial to be treated.
We have co-processing with the virus: co-construction, co-reverse transcription required, co-amplification, and co-detection. Which leads to the must that one band can be seen as a result if the whole process worked correctly. In case of a positive sample with virus present, we have to see two bands.
So in case we don't see any bands we know that process had some failure and that means go back to square one.
In reality that looks like this. In lane A corresponding to this scan graft here, you see one peak only. This is our amplified material, fluorescent for that one fluorescence primer that we have added. We have five -- only three here shown -- but we have five bands of lane markers of known size, so we can exactly calculate the lengths of the amplicon which is given us by the computer.
And this is 153.59 and the expected length is 154 for HIV in our case. This would be a positive PCR but a negative result.
On the contrary in lane C, we do have two peaks. This is the positive, the wild type peak for HIV, and we can easily distinguish, as you can see here, the lengths, the different lengths of these two amplicons.
In the middle, corresponding to lane B, we have no amplified material indicating that something either went wrong or as it is in this case, we have control -- completely negative control which goes along of course, with every test.
Using this method we, as many other speakers this morning have shown, went along to show our sensitivity. This is a scoring graph. The percentage of samples being found correctly positive in a serial dilution of an HCV preparation.
These are nominal values, genomic equivalence per ml, and you can see that the 95 percent cutoff level is in the range of 150 genomic equivalence per ml. In our hands the calculations to the WHO standard is 1:4.
So this is about 40, 45 international units per milliliter sensitivity, case sensitivity; which corresponds fairly nicely to the Centeon data, leading to something 104 viruses per milliliter needed in the original sample to be detected in our system.
Just to show you that the nominal virus load here is not an invention of ourselves, we have calibrated our scoring graph with the NIBSC working standardizing from John Saldanha.
We have made dilution corresponding to 1,110 genomic equivalence per milliliter, and repeating that 26 times under different conditions over the length of about five months we succeeded in showing that we have to expect scoring with that material that has been quantitated by a number of colleagues around the world.
Now I would like to show you some of our data that we have obtained by using this PCR in the frame of quality control. We had in mind to use this PCR a little bit different from what the outline now shows that has been given by the FDA.
And here I would like to show you first of all the data we have obtained in the last, about three years, testing more than six million donations. And this is just a summary of one recent observation period.
But first of all I would like to go a little bit more into detail just to give you an idea of how this procedure is done. Looking at qualified donors that undergo all the conventional testing we bring in the donor that qualifies in this particular donation, into an inventory hold of three months. Each sample is sitting there for three months.
After three months, prior to being used, we are making a pilot sample pool which is an additional vial that goes along with each donation. We do not cut pigtails or cut off the corner of a bag, but rather have samples. It's a closed system. We have an additional sample to each bag.
We are doing PCR and if the PCR of this pilot sample pool is negative these samples in this pool are released for pooling. A real pool is formed; pool out of all these donations. And we are doing another PCR, which is the test of record required by the European authorities.
If this is negative we release that for production; if not we would destroy that pool. If, in the pre-screening period, the sample pilot pool turns out to be positive in that period I'm showing to you, we had no possibility to go down to the single donor.
This was done by intention because we are quality control, not a medical diagnosis laboratory. So we made all -- well, we didn't create the possibility to go down to the single donor. That was kind of blinded for us, just not to come in touch with clinical laboratories.
So we cannot identify the donor by itself, but we have several steps as we made the pools to verify the different samples show positivity. And we have not identified the individual donor, and of course not at sero conversion, but we know that the testing was all right.
In a period of about one year covering most of 1997 we have tested roughly 1.8 million donations and we have 36 -- I'm saying here, donations because I don't have the donor on-hand, so to speak. But it was 36 cases that turned out to be PCR-positive.
And please note this is even after a 3-month inventory hold; a period which was used to clean out all post-donation reports, look-back positives, and such.
We have also found two donations positive for HBV and we have found none to be positive for HIV. And contrary to production pools made from the real donations we found all negatives so far. We never, ever had any positive production pool that we have submitted to the authorities.
Now let me turn to the new donor screening that goes concurrent with classical market testing. We have set up this system and it's about to be installed -- not yet running but very soon -- in order to comply with both worlds: with the European requirements that we have covered in the first phase quite well, and also to obey the rules given out by the FDA, just to be more closely to the donor when we are going to test.
Now, this I have explained to you. The difference is now that the qualifying donor will undergo again, the conventional testing, and those donors that qualify according to standard rules, we will also make a pilot sample pool. But we do not put the samples into an inventory hold for three months but rather do the testing immediately.
If this pool is negative all these samples go into an inventory hold of a length that I'm going to talk about in a second. And then we follow the standard procedure: after inventory hold we would release it for pooling, do the test of record to cover European requirements, and release negative production pools, and we would destroy positive production pools.
If such a pilot sample pool turns out to be positive we now can detect or identify the donor and the donation. Of course, the donation will be rejected, the donor identified, and now we can start to go onto a look-back. Which means having those samples on hand we can now call the samples from the look-back storage and do individual PCR.
And the lengths of this inventory hold period of time is one of the targets that we are addressing. We would like to know what is the optimal inventory hold to cover this -- or most of the window donations, and not to have the plasma sitting around for much too long, which is a costly process as you can imagine.
Now, consider at the time point A you have a cohort of donors and some of them or one of them might be in the diagnostic window. And PCR does have a diagnostic window as well. Just as a serological diagnostic window, PCR will not detect an infected person at day-1 of infection. It takes a couple of days as we have seen, until enough virus has been produced to be seen.
So if a person comes in right after infection we would not see it even on type PCR. So we have to make sure that this person comes back a second time after a given period of time, which of course is dependent on the window period of the virus.
Let's say this is 30 days. After 30 days each window donation, or each window period donor should have sero converted -- or PCR converted, not sero converted -- and we will be able to detect. And now you can imagine it will take a couple of days more because if it's today and we need 30 days, the donor won't come back in 30 days.
We are adding so-called observation period for the time to start of 30 additional days, and we calculated from the return rate of our repeat donors that we will have 90 percent of donors being here at the -- will come up within two months for a second donation.
And by this approach we expect that the efficiency of a 60-day inventory hold is helping to reduce the likelihood of having overlooked a window donation by 90 percent. That's what we have to build into our -- or, what we want to build into our IND.
Thank you for your attention.
CHAIRPERSON TABOR: Thank you. The next speaker is Dr. Flanders from Abbott.
Return to Table of Contents
DR. FLANDERS: Thank you. I want to present some information, internal development information on our HCV assay that runs on the LCx instrument.
First of all just to describe the overall process, this is using already existing platform -- the LCx Analyzer. We do a sample preparation which I'll describe very briefly, for extraction of the RNA, amplified normally and then automated -- detected in an automated way on the LCx Analyzer.
The sample prep procedure that was used for the HCV assay as well as HIV and HBV assays that we've developed assays for internally, and uses the Chiagen column procedure modified to include a vacuum manifold to provide for ease-of-use and also to provide for multiple sample volume capability.
Every assay, each assay includes or each run includes of course, a standard negative control and positive controls which do go through the sample prep procedure as well as each individual sample includes, as we've heard on several of the assays, an internal control which has the same primers as the target.
And then there's a non-sense sequence in the amplicon that's inserted, and we can detect both the target and the internal control simultaneously in the single sample on the LCx -- primarily as an indication that the PCR reaction performed as expected. Again, the internal control does go through the sample prep with the sample.
Now, I will describe just some internal data for the performance of the assay. This assay is not currently involved in IND studies. To understand the sensitivity and what has been lacking over the course of development of these various assays but now we do have some standards that we can compare to and to keep everything apples and apples, I want to describe the sensitivity of the assay with the WHO standard.
We prepared a sample from the original stock to about 5,000 IUs per ml, and then serially diluted that standard over 1,000-fold. We processed each of the diluted samples multiple reps of each of the samples through sample prep using either a one ml input sample of the diluted sample, or .2 mls of the diluted sample.
The one ml was originally designed such that from one ml of input sample the final extracted RNA could be split into three separate assays: one for HCV, one for HIV, and one for HBV. So we don't use the entire content; we use about a third of that. And then test each of the process samples with the LCx system.
Using the standard with a one ml sample prep which is a little more sensitive than the .2, we can detect with a cutoff signal of about 50, we can detect about ten copies or 10 IUs per ml of the standard; with a .2 ml it's about 40 IUs per ml.
And this is consistent with the difference in the sensitivity between a one ml sample prep and a .2 ml sample prep. So I think similar to some of the data that we've seen for some of the other systems.
In order to understand whether or not this sensitivity which is in fact, in a serial dilution and is a single donor, understand whether or not pooling of the same standard would have an impact on sensitivity, we tested and created 20 separate pools. We used a pool size of 64 specimens, so about 1200 individual, negative plasmas.
We used a Hamilton Microlab pipetting station for creation of the pools. Each pool was spiked with the WHO standard to reflect if one of the specimens was at 5,000 IUs per ml consistent with the Paul Ehrlich recommendations.
This results in a final pool that should have about 78 IUs per ml in the final pool. We processed each of the pool samples through the sample preparation; again, using either a one ml or a .2 ml input sample volume and then tested each in the LCx system.
Before I go to that data, this is just to show that prior to spiking the pooled samples with the standard, they were also tested without the spike and they were all negative for HCV. This is the internal control signal which is positive on all the samples, and this is the HCV signal which is of course, negative.
In the case of the one ml sample, the internal control again as I said, was positive for all the pools -- positive after spiking as well so they were all positive for the internal control -- and the HCV signal was positive with an average signal of 80.6.
And at the .2 ml the internal control is positive. Out of the 20 pools, 17 were positive. There were three that were slightly below the cutoff of 50. The average signal here was 62.5. This is very consistent then with the dilution curve that I showed previously. It's indicating that the pooling isn't impacted -- does not impact the sensitivity of the assay in this study.
We've also tested, as many have described, the various -- that looks out of focus to me, but anyway, these are all the different sub-types of HCV. We've tested it with true isolates of the various sub-types as well as transcripts that have been quantitated independently of all the various sub-types to the same level.
We do detect all the sub-types, A1 through 6, equivalently, and if we do this in a quantitative way we've shown with a quantitative HCV assay that we quantitate the various subtypes across the dynamic range from about 1,000 copies up to 107 copies equivalently as well.
Of course, there's a wide separation between the negative population and a normal, positive population, partly reflected by the fact that the amount of virus in most positive samples is quite high. Which has also been reflected and we've seen before, in the sero conversion panels.
I just have two here; we've tested a lot more than this. But just representative of what we see in most of the panels that are generally available; that is, that the RNA is already positive on the very first bleed that's available, and prior to antibody sero conversion.
The assay on all the panels that are commercially available have been tested has been positive on the first bleed available except one -- and again, this is from Bioclinical Partners. This one actually does go through both a sero conversion or sero conversion ALT conversion, as well as an RNA PCR GAT conversion.
But again, very rapid increase. This is day-8 I believe, and day-11; within a 3-day period it's already a very high viral load and very high signal for HCV.
And actually I think that's it. We are not in an IND study currently with this assay, but that reflects the HCV assay and its performance internally. Thank you.
CHAIRPERSON TABOR: The last speaker in this session is Dr. Psallidopoulos from Nabi.
Return to Table of Contents
DR. PSALLIDOPOULOS: Thank you, the organizers, for inviting me here. Thank you for attending this very short, brief introduction to Nabi's PCR test system.
Our primary goal is to devise a plasma test system to screen out units that are going to the Nabi immunotherapeutics in order to limit the virus load, to define maximum increase this safety of Nabi products, and these test should be of a high level of screening.
Very quickly, introduce you to the design, our methodology, and give you some of our results and our future directions.
We have a physical separation of the PCR processes through two different buildings using different laboratories, to get the PCR and then to get at work areas within the PCR laboratories. We have a unidirectional work flow, we have dedicated equipment and consumer boats and non-consumer boats in each work area or laboratory. And we generally follow the good PCR laboratory practices.
Our plasma pooling is what has already been described by Dr. Conrad from NGI. It's a 3-dimension matrix using the 512 pooling sample and using robotics to do the pooling. Our testing algorithm is already described.
Basically if the master pool is negative then all the 512 units are released. Otherwise we test the primary pools to identify individual unit which is then tested and verified in an independent laboratory like NGI.
The nucleic acid extraction again, is the standard ultracentrifugation, with the addition of an external PCR control to give us a feeling of the sense of the quality of the RNA extraction. We're using the RNA study guanidinium method to extract the RNA.
Basically, one-fifth of the RNA sample is duplicated tested for the HCV and the same thing for the PAW109. The amplification process is the single test that combines RT-PCR for the first round of amplification, followed by nested amplification using rTth, and the primers that are described by the FDA PCR-2 protocol.
The detection method is the ethidiumbromide gel, standard procedure, using the followed recommendation system and keep records of both pre- and electronic forms.
In the past we have participated in the well viral quality control proficiency study which were given half ml samples and we actually used an old protocol of extraction, diluted down into two-and-a-half mls and then extract the entire two-and-a-half mls.
And this is the performance of our assay on the replicates. We have achieved somewhere around 40 copies of genomic equivalence, the 50 percent hit rate.
We also use Pelispy, which is obtained from the Dutch Red Cross, Nico Lelie -- it's a run control. And this is the results of the last, about 36 runs, and you can see that the performance of our assay is about seven genomic equivalence, a 50 percent hit rate.
So having established this assay, current screening studies is that we screen something like 5,000 units so far; that they went into clinical lots, and we haven't found anything positive yet. We are preparing to submit an IND and we are performing the assay validation.
Our future plans is to transfer this assay to the clinical labs and introduce new testing -- a prime test for other viruses as needed, and also currently we're trying to use the genetic analyzer to detect the amplicons.
That would give us enhanced sensitivity, which is an issue for the next point; to use this in other Nabi products. As I said in my introduction one of the primary causes of using the PCR test is to eliminate the virus load to define maximum.
Nabi uses other products like civacir which is made up from pools of HIV-positive plasma units. This is quite the reverse. In the first time we want to eliminate any positive units; here we actually get all the positive units.
This have to be fractionated so we use various virus elimination and virus removal protocols and the PCR is important to make sure that these pools are negative before they enter the plant so the sensitivity and specificity for the assay has to be improved.
This is where we stand at the present. Thank you.
CHAIRPERSON TABOR: Thank you. We'll now break for lunch. I'm afraid the only nearby place to eat is the cafeteria in the building. Most of the other sandwich places on adjacent streets are too far to go. We'll reconvene in about an hour, depending on the efficiency of the cafeteria lines. Thank you.
(Whereupon, a brief luncheon recess was taken at 12:47 p.m.)
AFTERNOON SESSION
(1:51 p.m.)
CHAIRPERSON TABOR: We're ready to begin the session on Regulatory Perspectives on Assay Validation. The first speaker is Dr. Indira Hewlett from the Food and Drug Administration.
Return to Table of Contents
DR. HEWLETT: Thank you, Ed, and good afternoon everyone. This afternoon I'll be discussing some of the major regulatory issues in regard to assay validation from an FDA perspective.
I think we can all agree at this point that nucleic acid testing may be the most sensitive method currently available for early viral detection. And implementation of nucleic acid testing would result in reduced viral burden in blood and plasma.
The blood and plasma industry has proposed testing plasma pools for nucleic acid, and this is partly because pool testing may be most practical at the present time.
Plasma pool testing is currently occurring in at least three different scenarios. The first involves the use of an in-house test developed by a blood product manufacturer. The second approach is to use a commercial tested. And a third option is to contract out the testing to a testing service that is competent in performing this type of testing.
Regardless of the testing scenario, FDA's view is that assays used to test plasma pools rather than single donations would still be considered to be donor screening assays and therefore subject to validation and license requirements.
This would ensure manufacturing consistency and define the performance characteristics of the assay by clinical studies. As with other blood screening assays, donor notification and product management strategies will have to be in place at the time of implementation.
I'll now get into the issues, specific issues in regard to validation. Assay validation to be stated briefly, consists of development of the appropriate quality control procedures to assure manufacturing consistency under GMP of the components of the test method or the test kit, and the accumulation of relevant laboratory and clinical data to support the intended use of the product and the manufacturer's claims.
I should mention that the points I'm going to discuss rather briefly today are actually discussed in great detail in a draft guidance document for industry in the manufacturer and clinical evaluation of in vitro tests to detect nucleic acid sequences of HIV1.
This document, although it addresses issues pertinent to HIV1 is also expected -- the criteria listed in this document are also expected to apply to tests that are being developed for other viruses that are screened for in blood.
I should also mention that this document is currently available at a CBER Web site, so people that are interested should be able to access it. And it's currently been published for comment, so we're taking comments about the criteria at this time.
The second document, which is the draft Federal Register Notice on FDA's proposed approach in regulating nucleic acid tests for plasma pools is still under review within the FDA. And I will not be discussing this particular document since it's been presented at a blood products advisory committee meeting in the past.
Some general criteria apply to all in vitro tests and they are listed on this slide. Specifically, there should be some rationale for the design and format of the test: for example, the type of specimen that is being tested, the controls, calibrators, the cutoff. These are examples of design that impact the design and the format of the assay.
The stability of specimens under various processing conditions such as collection, transport, and storage should be determined. The components and reagents used in the test method or kit should be subject to appropriate quality control methods and their stability determined as well.
The performance of the assay should be defined in terms of reproducibility, analytic and clinical sensitivity and specificity, and any instruments including software, that is used to perform the test method or to calculate results should be validated for their function.
Most importantly, statistical methods should be used to validate results from laboratory and clinical studies.
The design of the assay -- I'm actually going to get into each of the points I had on my previous slide and I'll discuss the key points in regard to each of those criteria.
The design of the assay should take into account several features, most critical of which are selecting the appropriate primer and probe sequences and the target region for nucleic acid assays. For example, one may wish to base the selection on the degree of conservation of the particular region so that multiple sub-types may be detected.
Inclusion of appropriate controls is necessary to ensure that the reset results are valid. These controls would include internal and external controls such as controls for ultracentrifugation if such a step is involved in sample processing, for extraction, and positive controls of course, that would define the validity of the run.
Additional controls that are used in nucleic acid amplification methods are controls for false positive reactions that might arise from contamination as you heard this morning, during the pooling process, or cross contamination during the PCR amplification process.
Finally, if an assay has a quantitative format, calibrators and quantitation standards should be designed to have an acceptable range and limit.
Specimen stability has an impact on assay performance -- and I think you heard a little bit about the stability studies this morning; they're actually quite impressive.
The conditions of stability need to be defined. These conditions include collection, transport, and storage of the original specimen; the specimen undergoing extraction by a specific protocol -- as well as the stability of extracted material on storage.
For such studies, FDA is willing to accept the use of well characterized specimens, well uncharacterized spiked materials, but the use of clinical specimens, naturally occurring clinical specimens, is highly recommended.
The analytic sensitivity of the pool test may be determined by testing dilutional panels of known positive clinical specimens. Testing of sero conversion panels and low titer RNA specimens are also useful in evaluating analytic sensitivity.
FDA's current proposed limit is 100 copies per ml for the full test. This proposed limit may be modified in the future depending on the actual experience in regard to performance of pool tests in the field.
And we've decided to start with this limit because as you'll hear from some of the studies that Dr. Yu will discuss later, that this 100 copy limit is actually achievable by many of the tests that are currently in use. In all cases, tests should be run in parallel with a comparator or reference assay.
All tests have to be evaluated for their performance on clinical samples to determine their specificity and sensitivity. Specificity is established by testing random blood donor specimens with follow-up testing to resolve reactive results.
Sensitivity is established by testing known positive specimens. Sensitivity studies should include testing of genetic sub-types and specimens from persons at different stages of disease and possibly different risk groups. And this morning we heard about differential detection rates among ethnic and racial groups, so those are some things to consider at assay validation of the HCV tests, in particular.
Specimens that may be expected to cause interference in the assay by producing false positive or false negative results should also be tested to determine their impact on analytic specificity of the test.
And finally, the reproducibility of the assay should be determined by testing at multiple sites, a given panel or a given number of specimens that have viral copies -- that have deferring viral copies -- including samples that have low copy numbers, to determine how reproducible the assay is across the linear range.
Some issues are of special concern in regard to pool testing. They are pool size -- we'll start with pool size which we know has a substantial impact on the sensitivity of the test. And the derivation of the pool size should be based on the ability to achieve acceptable limits of sensitivity in the final assay.
Another aspect is the analytic sensitivity of the pool tests. And both of these parameters are extremely important, particularly in light of the need to demonstrate equivalence or enhanced sensitivity of testing pools to currently licensed methods.
An issue that should also be addressed but may not be as much as big an issue as we had thought originally, is the generation of interference due to matrix effects that might result from pooling of specimens which may cause interference in the assay.
And I think we wish to continue to see data addressing matrix effects because there are different pooling regimens and different pooling schemes, so there's going to be a continued need to see if in fact, matrix effects are an issue with pool testing.
There should be mechanisms for logging and tracking the specimens and quality assurance of systems that are used to pool and test pool specimens, as well as to trace back positive results to the original donation and the donor.
And this of course will include instrumentation and software validation since instrumentation is expected to play a major role in the implementation of pool testing.
So in summary, implementation of nucleic acid testing of plasma pools should further reduce the risk from window period donations. FDA still continues to view pool testing as an interim step towards single donation testing due to technology development.
Assay validation should occur under the IND PLA mechanism which provides adequate control of manufacturing procedures and procedures for tracking inventory and resolution of reactive results.
The use of reference reagents established here at CBER and elsewhere should be helpful in determining assay performance and laboratory proficiency, as well as in lot release testing of the pool test method or test kits that are used to test pools.
Laboratory testing and clinical trials will need to take place to establish sensitivity, specificity, and reproducibility. And finally, appropriate procedures for donor and product management are expected to be in place at the time of implementation.
Thank you for your attention.
CHAIRPERSON TABOR: Thank you. The next speaker will be Dr. Nubling from Paul Ehrlich Institute.
Return to Table of Contents
DR. NUBLING: First of all I also want to thank for an invitation to this interesting meeting. I will start with a short overview about the current situation in Germany concerning NAT testing and then we'll switch to the validation issue.
It's meanwhile, nearly three years ago that the first blood banks introduced NAT screening on a voluntary basis in Germany, and as you can imagine, this implementation into the routine screening program pushed discussions quite strongly if such NAT testing, first is feasible in general, and also if it makes sense concerning increasing of viral safety.
And just one year ago there was a meeting of the Paul Ehrlich Institute where all blood banks from Germany participated and the results obtained in the meantime were given at this meeting.
And this is just a summary of the results concerning HCV NAT testing. Up to this time there were five blood banks which had introduced NAT testing for HCV on this voluntary basis.
And as you can see, quite different numbers of donations had been tested until end of last year, and also the pool size chosen for testing was quite different between these different blood banks.
Nevertheless, more than one-and-a-half million donations were tested in total by HCV NAT and the results for entity-positive but antibody-negative donations were 13 among this number.
This results in incidence among German blood donors for NAT-positive, antibody-negative donations, approximately 1 in 120,000. And is much higher than for example, for HIV or HBV. This may explain why the diagnostic window phase -- we have discussed this early in the morning -- and another important feature is the titer during the window phase which is much higher for HCV compared to HIV or HBV.
These facts were a reason for the decision of Paul Ehrlich Institute to introduce HCV NAT from the first of April of next year. A sensitivity limit for the single donation was given with 5,000 initial units per ml, and that's the main topic of my talk today, validation is required of the methods and the documents of the validation are order to be given to the Paul Ehrlich Institute until end of this year.
First the question, what's the reason for this sensitivity limit on the single donation phases? We've performed quantitative analysis of quite many sero conversion panels. Here are seven sero conversion panels and make conductive PCR.
And as you can see, here's the logarithmic scale of copy numbers per ml. All PCR-positive donations among these sero conversion panels would be picked up by a method which is able to detect 5,000 international units permitted under single donation basis.
And what you have also seen this morning already is between the last PCR-negative donation and the first PCR-positive donation there's a quite steep increase in titer. So that we expect quite few samples which are under their 5,000 international units limit.
This is also confirmed by quantitation of single donations which have already been identified by nucleic acid screening. Here's a limit again, and we quantified different genotypes, different single donations, and all of them are quite higher than the 5,000 international units limit, and the sensitivity limit used for the identification of these donations was even lower than the 5,000 international units.
So that's the background for our decision. Now to the validation. In principle we differentiate between in-house tests and commercial tests, and in-house tests may be used by the blood banks if they are validated.
For in-house tests the features and the modifications of the tests have to be validated by the user. For commercial tests we accept the validation performed by the manufacturer. Independent, if the commercial tests are complete tests starting with extraction and ending with detection of the amplification products, or if there are only parts of the NAT procedure -- for example, only extraction kits.
Modifications of these commercial tests of course, have again, to be validated either by the user who introduced the modification, or by the manufacturer if he recommends this modification.
Documents which were relevant for our interpretation of validation requirements are mainly the ICH documents which have been mentioned also in the morning here. There are several other documents available which are quite useful when a PCR system is established, but the requirements validation are based on these ICH notes for guidance.
HCV NAT in blood bank setting is considered as a limit test and these five points are the main points which should be validated, starting with specificity.
Specificity means the identify of the amplification products and I think it's obvious that it's strongly dependent on the choice of primers, probes, assay stringency. And before establishing a PCR NAT system, a databank comparison of primers with target sequences can avoid false positive results.
During validation the amplification product has to be characterized either by size restriction, hybridization, or sequence. And for validation of specificity we require at least 100 samples; either negative samples if single donation PCR is performed or 100 negative pools if pool PCR is performed. Both is possible, of course.
Sensitivity, I think it's the most important point during validation. In principle there are different definitions of detection limits. Some people speak about the detection limit if 100 percent hit rate is obtained; 95 percent hit rate we have heard today; also others defined detection limits with 50 percent hit rate.
When the statement or the requirement for 5,000 international units, a single donation was made, it was meant in the way that it should be detected in more than 95 percent. This detection limit should be guaranteed for the most prevalent virus types in Europe or in Germany. These are mainly virus type 1, but also virus type 3.
And of course, the detection limit should be controlled on a routine basis by a positive controller.
Now it's the question, how should the detection limit be determined? We recommend to perform 3 half log10 dilution series of the WHO standards, HCV RNA, or a reference preparation which is calibrated against this gold standard, and to perform per dilution point, 8 replicates so that 24 values are obtained for each dilution point.
Then PROBIT analysis is performed and from the curve, the 95 percent value and 50 percent value and even 100 percent value can be obtained.
We encourage also, persons of blood centers to calibrate their own in-house reference preparations in order to create run controls, and for calibration of this in-house reference preparations we recommend the same procedure in parallel with WHO standards.
There are difference reference preparations already available. The most important of course, is WHO standards, but it's available only in limited amounts, for good reasons.
From CLB, BBI we have also meanwhile, created a hydrolyzed preparation which has been mentioned also in the morning already, and also other organizations -- also NIBSC has a working reagent. This working reagents are all accepted as calibrated material as soon as they are calibrated.
Next point concerning sensitivity is the avoidance of false negative results. We require inhibition control for a pool size of bigger than 50 donations. Also internal control or a parallel spiking control to be performed with an assay.
The next is genotype sensitivity. It can be obtained before establishment of the PCR system by databank comparison of the primer sequences with the sequences of the available genotypes. And of course, it has to be checked with genotype samples; for example with panels available already.
Precision means laboratory internal variation. It is dependent on variation of persons, equipment, and also should be investigated on different days. We require that for in-house PCR reagents have to be defined with a shelf life and also quality control of new batches of primers, enzymes, dNTPs, etc., has to be performed.
Reproducibility means precision between different labs. Here we recommend to use control panels and to participate in collaborative studies in order to estimate the results in comparison to other laboratories.
Robustness of an assay can be obtained by appropriate training of personnel, by creating meaningful, standard operating procedures. Avoidance of contamination should be validated by using also high titer standards during the validation phase, alternating with negative samples in order to detect potential carryover.
And of course in the routine setting the robustness is controlled also by appropriate controls for sensitivity, inhibition, and contamination. If a (unintelligible) should occur, one should have in mind that this can occur. A second amplification -- at least access to a second amplification system should be available in order to have no shortage on blood products.
This validation criteria I think are essential for reliability of methods, especially for in-house methods. And in Germany there are many blood banks which are creating their own in-house systems and it's a major challenge for them but I think it's really essential.
And it's of course, also a pre-requisite for approval of tests which will be performed in future in Germany, meaning that commercial tests at least then have to give authorization documents to our institutes. And I think our requirements are a compromise between what can be done and what should be done in the optimum case.
Thank you.
CHAIRPERSON TABOR: The next speaker is Dr. Rautmann from the European Pharmacopoeia Commission.
Return to Table of Contents
DR. RAUTMANN: First I would like to thank the organizer for inviting me to this meeting. And when I started preparing my talk for this presentation I had some discussion. I was also asked to actually present to you how the system operates, who the institution involves, and what is the legal background behind it.
I must admit that on the other side of the ocean it's not always understood by all the people and the system is quite complex.
Therefore, my talk today will be divided in three parts. I will briefly mention the institution involved in the implementation of NAT, the EU regulation, and what are the impact actually, on the implementation of NAT.
So actually this first slide is showing you how actually things are organized on the national level to guarantee quality, safety, and efficacy of medicinal products, and how was it switched from national level to the European framework.
To make the discussion a little bit easier I have put on the bottom of the different parts, the place where the institutions are located, which is Brussels in Belgium, London in the U.K., and Strasbourg in France.
Since the first of January '93 actually, community legislation provides for a free trade of goods within the European Union, which also implies for medicinal products. Therefore, the European Union has created a few tools.
Among those by community regulation, there has been the creation of the European Medicine Evaluation Agency in London which is in charge of making the evaluation of medicinal product proposed to be put on the market in Europe through the centralized procedure.
Which means counting Europe there are two ways. Either you have the centralized procedure -- and this is compulsory for products of list A, which are medicinal products involving a biotechnologic process.
And in this case the dossier is evaluated by the EMEA in London, which makes a recommendation and this recommendation is forwarded to the Commission and it's actually the Commission which then agrees and provides the centralized marketing authorization.
This is on the part of the licensing. On the part now, of post-marketing activity, actually the European Union has made a contract with the Council of Europe in Strasbourg asking the European Department for the Quality of Medicine, Division 4, to create and coordinate a network of the official medicine control laboratories of each of the member states.
And therefore Division 4 is in charge of coordinating this network, and this is by contract between the two institutions.
EDQM has also in charge in one of the divisions, the European Pharmacopoeia. Briefly, this slide is showing you the organization of the European Department for Quality of Medicine, which as I said, located in Strasbourg, and is department from the Council of Europe.
EDQM is organized in four divisions: 1, 2, 3, 4. I am working in Division 4. As I said, within EDQM Division 4 is actually the European Pharmacopoeia Commission which, in conjunction with this group of experts is actually drafting and adopting the monograph of the European Pharmacopoeia.
We have also Division 2 which is what we call the Publication Unit. This unit is responsible for publishing the Pharmacopoeia and other publications -- Pharma Rupa and proceedings of several meetings we're organizing.
Well, then we have Division 3 which actually is the laboratory, and this laboratory is split into two sections: one is the chemical section and the other one is the biological section.
And actually, the laboratory will serve both the European Pharmacopoeia Commission but is also supposed to serve activity within Division 4 which are on one side, the biological standardization program. And in this program we are either establishing biological reference preparation which will be used in the European Pharmacopoeia monograph.
And the second activity of Division 4 is, as I said before, coordinating the official medicine control laboratory network.
Okay, let me now switch to community regulation. This includes two types of documents: either the legally binding documents under the form of directives or regulations, or on the other side you have also non-binding documents which are called guidelines, or sometimes Note for Guidance.
The reason why community has decided to use also guidelines is to allow certain element of flexibility and not provide a too-strong legal constraint on evolution of technology. And therefore, it's always possible for manufacturer to deviate from what is actually recommended in a guideline, provided the data that manufacturer is having is supporting that he was right in deviating from the guideline.
In addition to this we have also monograph of a European Pharmacopoeia, which like directives, are binding documents. For the topic of today there are several documents or legal texts which are of importance.
We have "Council Directive 89/381", which is as I said, a binding document. In the European Pharmacopoeia we have "NAT General Method", and we have the monograph on "Human Plasma for Fractionation". Those three texts are biding documents.
Then we have guidelines, and in guidelines we have the "CPMP/BWP/269/95", which is a guideline on medicinal products derived from plasma. We have the second guideline which is "390/97", which is the guideline recommended in the implementation of NAT as of July 1st, '99.
And the last guideline here which was a guideline which was developed by an adult working group within the official medicine control authority network. And in this guidelines and this expert group which designed this guideline, we had Dr. Nubling and Dr. Saldanha who are present today.
If we rapidly go through the first document which is the "Directive 389/381", the three articles which are important on the topic of implementation of NAT -- actually the guideline in Article 3.1 asks that member states takes the necessary measure to prevent the transmission of infectious diseases.
And the text is referring that the Directive 75318, as well as the European monograph, should be applied when a member state is applying those measures.
In Article 4.1 it's stated that member states have a responsibility of ensuring that the process as well as the purification steps used to produce medicinal products derived from human blood or plasma are properly validated; that they contain batch-to-batch consistency; and that as state-of-the-art technology permits, the process guarantees the absence of viral contamination.
The last article, 4.3, actually is the one providing for control authority batch release, which means that the person responsible for marketing a product on the member states within the European Union, if the law of the member states provides for, then this person must submit to the competent authority a sample of each batch of the product which will be released on the market.
Which means that each batch will be retested by the competent authority. And in this article it's written that actually a batch will be retested only once, which implies mutual recognition within the OMCL of different member states of the European Union that once a batch has been tested and released for one of the market, one of the OMCL within the European Union, no other OMCL can retest the batch for the purpose of batch release.
I'll now go to the CPMP/BWP -- actually, this series of letters in front of a guideline means that the guideline has been developed by the biotechnology working party of the EMEA in London, proposed to the CPMP, adopted by the CPMP, and then a guideline is actually proposed to the Commission and becomes an official guideline for the European Union.
As I said, implementation date of this guideline is July 1st, 1999. The guideline states that plasma pool must be tested for HCV by NAT; that only plasma pool shown to be non-reactive can be used for the manufacturing process.
It applies also to excipient, which means that actually the date in which an excipient is either entered in the final product or used to manufacture an intermediate, will be the date compared to the implementation date of July 1st, 1999.
This has some impact also, for manufacturer in the vaccine field or for some biotech product in which we have quite a lot of albumin for instance, inside. The guideline says also that the test method must be validated and that each of the assays must use a run control calibrated against the WHO international standard, and that this run control must be equivalent to 100 international units per ml.
This guideline applies to manufacturer but as I said, within the process of official control authority batch release this guideline will also apply to official control medicine laboratory.
Therefore, within the network we created another working party having experts from the different OMCL who are actually going to perform those assays within the OMCL at the occasion of batch release. And we discussed about the implication of the implementation of this guideline.
When we had this discussion one of the items which came up was actually validation and how the different OMCLs were going to validate their in-house or their method in the OMCL.
The other point that was raised during those discussions were, some OMCL were already using NAT for quite a long time; others were just developing the methodology. Therefore, two types of studies were organized within the OMCL network.
One study is the so-called self-assessment study in which a set of samples were distributed to the different OMCLs and tested. The results were then sent back to EDQM who analyzed the data, coded the data, and provided data to all OMCL. This is the self-assessment study which is supposed to help OMCL to improve the performance when performing NAT.
The second study which will be organized at the end of this year is proficiency testing study, again, within the OMCL. A panel of sample will be distributed to the OMCL and they will test it blindly in a way as it was done by the OC study.
So I'm just briefly mentioning this NAT validation guideline which actually now is a proposal. It was released the beginning of this week for public inquiry, which means that we have all possibility to access on this guideline. And please if you have comments, send those comments back to EDQM, but not later than the first January, 1999.
As I said, this guideline was drafted by experts within the OMCL network. I intended to briefly go through specificity and detection limits but since Dr. Nubling has covered a lot of it and a lot was already described by Dr. Nubling is actually within this guideline, it would be merely a waste of time. Yes, you have seen this more or less in the slide of Micha already.
Concerning the detection limits, actually it was also stated this morning that NAT can be considered as a limit test, and therefore if one looks to the ICH guideline about limit tests, the ICH guideline will commence the use of a detection limit.
But for practical purposes the expert in the group proposed to use the positive cutoff point as defined in the European Pharmacopoeia Monograph, which means that the positive cutoff point is defined as the, more or less the 95 percent hit rate.
In the guideline there is also given a way on how actually to determine this positive cutoff. And just in addition to what was just said before by Dr. Nubling, the guidelines foresee that actually the 24 test results from each dilution may be obtained by three different ways: either by making three independent tests on three days, by using eight replicates for each dilution; alternatively, four independent tests with four replicates for each dilution; or six independent tests and four replicates. All together, having always 24 test results for each dilution.
So what I would say in conclusion is that actually the implementation dates for the European Union is July 1st, 1999. I tried to show you that actual legal provision is provided at community level, either in the CPMP Note for Guidance, or in the European Pharmacopoeia.
But official control of these batch release is still at the national level, which means this operates under the principle of subsidarity, and in the directive it is also foreseen that this operates with mutual recognition.
And this is a very important point in which we at EDQM -- for which we develop a lot of activity because what we must is foster the mutual recognition of test results in one cell, and the other in cell within the European Union.
So I thank you for your attention.
CHAIRPERSON TABOR: Thank you, Dr. Rautmann. I'd like to say that I think it's very interesting to hear the comments of Dr. Rautmann and Dr. Nubling about the European approach to this problem; something we're going to have to keep in mind here in the United States as we approach the problem from a regulatory perspective.
The next speaker is Dr. Saldanha from the NIBSC.
Return to Table of Contents
DR. SALDANHA: Thank you. I'd like to thank the organizers for inviting me to talk at this meeting, and I'm going to talk really, about the development of working reagents and the first international standard for NAT testing of HCV.
And I think we heard this morning about the use of these reagents and I'll try and give you a very brief overview on the development and the characterization of these reagents.
So I will start off by briefly stating I think the obvious, which is why do we need standards in the first place? I think we all accept when NAT testing was first introduced that there was a tremendous variation of the sensitivity and specificity of these assays, making it very difficult to compare data from different laboratories.
As Guy Rautmann said in the previous talk, that in Europe there's going to be mutual recognition of results, and to be able to do this you have to have confidence in results from different laboratories.
Another major problem that we face, especially with the manufacturers, is the discrepant results between the manufacturers and the official control laboratory which could cause problems with the release of final products.
And finally, before we can introduce any sort of routine NAT screening such as the CPMP guidelines, we need to have standards in place.
I'll take you very briefly through the CPMP guideline which in fact, Guy Rautmann mentioned earlier on, which is the introduction of NAT testing for hepatitis C virus RNA. And there are two main points to this guideline.
The first one is that each run of the validated assays should include a suitable working reagent, or a run control. And the level of RNA in this run control should be equivalent to 100 international units.
And a non-reactive pool in this system is defined as a pool which is found non-reactive using an assay which can detect this run control.
And what's going to happen is, from the first of July, 1999, only batches derived from plasma pools tested and found non-reactive will be released by the marketing authorization holder.
So the first reagent that we developed at NIBSC was a working reagent, and this working reagent was based on the results of a collaborative study that was run in 1994. And the reagent is a 1:1000 dilution of a positive donation, which is a genotype 3 diluted in human cryosupernatant.
The RNA content is approximately 4,000 genome equivalence. This has been determined by the branch DNA assay and this is equivalent to 1,000 international units. And I'll go into this derivation later on.
To-date we've made three batches of the working reagents -- about 2,000 to 3,000 vials of each. And these have been sent out to laboratories since August 1995. And we request the laboratories to assay the reagent NAT, 1:10 and 1:100 dilutions and return the results to NIBSC so they can be analyzed.
And what I'd like to say is the 1:10 dilution is equivalent, approximately, to 100 international units. So if laboratories can pick this up regularly then they can comply with the CPMP guideline.
So the first report that came out analyzing the data between '95 and '97 shows that of the 30 laboratories returning results, the NAT reagent was not in fact, as you'd expect, detected by all assays. There were only from 50 to 92 percent of assays, depending on the type of assay.
The same with the 1:10 dilution which was from 20 to 80 percent, and the 1:100 was rarely detected. And the way this data was analyzed was to collect all the data from the laboratories and pool them to get these final figures. Which I think in a way is misleading because if you look at the data on an individual basis the results look a bit better.
So I'll go very briefly through the results of three laboratories. These bubble charts for the 1:10 dilution -- that's the LABCODE and the type of assay which is an in-house. The size of the bubble represents the number of assays done, which is written at the side. That's the time and percentage of assays.
Now, if you look at individual labs you can find that some laboratories can detect that dilution of the reagent in all assays. We have other laboratories which actually show an improvement with time, and I think in this assay this laboratory switched from a first generation amplicor to a modified amplicor using a different extraction method for the RNA. And I think it's obvious that there's an improvement in sensitivity.
And other assays really just crash and really, they can't detect the 1:10 frequently.
So the second report which looked at 19 laboratories which returned results over the last year, and in this analysis we only looked at laboratories which submitted more than 12 assay results.
Because again, one of the problems with this sort of self-assessment study I guess, is that not all laboratories return results regularly, so it's very difficult to get a complete picture of the overall specificity -- sensitivity of the assays.
So we looked at 11 laboratories which submitted 12 or more assay results for the last year, and I'll show you very briefly again the bubble charts for some of these. Again, I think you can see that 100 international units can be detected very consistently by laboratories. This is over a period from April last year to April this year.
Occasionally you find laboratories which miss an assay, and there are some laboratories in the third bubble chart which have problems consistently detecting the study up to the standard.
Now as we heard this morning there are -- sorry. For the working reagent -- to briefly conclude these results -- show that it's useful for monitoring the performance of each assay run, so it's a run control.
And it can demonstrate an occasional failure of the assay, and it can also show whether the assay is improving over a period of time or whether you need to change the assay. So I think it's quite important to use some sort of run control in routine assays.
Now we know I think, that there are several working reagents available at the moment. There's the CLB Pelispy; there's a CBER reagent; a Paul Ehrlich reagent, and the NIBSC working reagent. And I think again, you're aware that people define the unit -- the RNA in these reagents in different units.
So we have anything from copy numbers per ml to genome equivalence per ml, to PCR-detectable units per ml. And I think you're also aware that we now have an international standard.
Following several meetings over the last two or three years we decided to prepare an international standard for the WHO. And this standard was accepted by the Expert Committee for Biological Standardization in October last year.
It's a batch of 2,000 vials which contain a lyophilized preparation of genotype 1 donation diluted in cryosupernatant. And the RNA content of this sample is expressed in international units. And this is a purely artificial unit. And each vial contains 50,000 units because the concentration is 105 per ml and there's half a ml in each vial.
And I've done some preliminary calculations. I think you probably got the idea this morning that the calibration of the international unit with the genome equivalence varies, anything from two to four.
And I've done a preliminary characterization at NIBSC in which the concentration of the international standard was determined very kindly by Chiron using the branch DNA assay as 5 X 105 genome equivalence per ml. And this was an average of two assays.
And the assigned concentration of the international standard is 105 IU per ml, so that works out at one international unit being equivalent to five genome equivalence.
I've done a similar calculation for the concentration of the NIBSC working reagent 96586, where again the undiluted donation was titered by the Chiron branch DNA assay to give a titer of about 4 X 106 genome equivalence per ml. And the working reagent itself is a 1:1000 dilution of that material.
And the next thing that we did was parallel assays using the international standard and the working reagent. And the difference in titer between these two reagents was two logs, roughly, and that approximately gives one international unit is four genome equivalence.
I realize these data are preliminary, so in fact what we've done this year -- and in fact the study has just been completed and I think we saw data from one of the labs this morning -- is to set up a collaborative study to calibrate several working reagents.
The NIBSC working reagent, which is about 4,000 genome equivalence; the Paul Ehrlich working reagent which is 105 genome equivalence per ml -- and this is lyophilized; a reagent from the ISS from Rome, which is about 2,000 international units; the CBER panel 1 which is about 1,000 genome equivalence per ml; and the CLB Pelispy run control which is 3,600 genome equivalence per ml.
And these have all been calibrated against the international standard using parallel assays. I'm afraid I haven't finished the analysis of the results yet but I should be able to have those available by the end of the year. And by that time we should have all these reagents calibrated in international units against the international standard.
And finally, the last study that we are planning to do is to calibrate different genotypes against the international standard, because I think at a couple of meetings, especially the last meeting in Amsterdam, there were concerns expressed about the feasibility of calibrating different genotypes against the international standard, which is the genotype 1 virus.
So we're proposing to use this study to determine the RNA content of different genotype samples and then to determine the efficiency of the different assays for different genotypes, and then to calibrate these RNA contents against the international standard.
And so far we've collected five genotypes; we're still looking out for the 6th, and this study I hope, will get underway by the end of the year.
And I think I'd like to conclude with probably saying that working reagents and standards are important to validate individual assay runs, and they're especially essential for the introduction of routine NAT testing.
And I think we still have to bear in mind, especially that nucleic acids -- the presence of nucleic acid doesn't really necessary indicate infectivity. Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker is Dr. Lelie from the Netherlands Blood Transfusion Service.
Return to Table of Contents
DR. LELIE: Yes, thank you for allowing me to show you, give you some idea what we are doing on the CLB. And to do this I always start with the first slide which reminds me of what we are doing.
We have available, dilutional standards that are available in large amounts. From those dilutional standards, plasma standards, we produce panels -- proficiency panels. And we did that in the past with the EuroHep studies and the VQC study.
And in 1999 we also do this as part of a European concerted action of the European Society of Clinical Virology, and we also try to harmonize the activities that are done by Mrs. Dex or Ellen Alan of Australia, that also does similar studies in the Australian and Asian Pacific region.
From the data that are submitted by the laboratories that are proficient, we then can characterize the standards and we also have an instrument to validate the assays in the field.
And then this is an ongoing system that should be done every year or so. So that it then enables us to produce similar panels, actually proficiency panels or reference panels that can be used for validation of the assays.
So we now are doing a multi-center validation study with reference panels in Europe, and the creation of the standards and legalization of assays also is the instrument for defining levels of run controls that should be used in the different test options.
Just I give you an example of the HCV proficiency study of 1997. There, 81 laboratories in the world submitted about 140 datasets. And we found that the undiluted samples -- so from positive and negative control samples -- that there were 84 percent of the labs that submitted correct results.
And we had two dilution series -- the EuroHep and the VQC dilution series -- and we found that 71 and 73 percent of those labs submitted correct results. So in total, two-thirds of the laboratories that participated produced results without errors.
Then we can look at the proficient laboratories, so we then only take into account the labs that passed the inbuilt proficiency test, and for instance look at the regular Amplicor kit that is used by 27 labs. And you see that the 100 percent detection limit of the normal or of the 1.0 Version of this test has a detection limit of 200 -- it is about 1,000 international units per ml.
You can, so draw through for instance, the Amplicor results, an ideal line which is in fact the PROBIT analysis. You can use these sort of statistics to determine a 95 percent detection endpoint or the 50 percent endpoint, but you can also use the data -- if all laboratory test dilutions are standards in parallel you can, from the shift in parallel lines, calculate the relative potency.
And we did this to calibrate the EuroHep standard against the VQC standard, and the VQC standard isn't available in the larger supply and this is important for the future. You can see that if you combine all the 94 datasets of proficient labs you see a relative potency of a factor 2.3. And if we do test runs in parallel with the branch DNA at CLB we find similar results; also 2.3.
We cannot simply use the branch DNA assay because we have seen that there are half-log differences between lots of the branch DNA assay. So you really should express the results based on the standard to which you have in the past, assigned a value.
So we talked about the level, the CLB run control that is now produced and should be further characterized, is equivalent to 270 genome equivalence per ml. And these data are just derived from the WHO evaluation study in which 22 labs with good results provided about 80 test runs where the VQC standard and WHO standard were tested in parallel, and from the average results -- which are huge variation of course -- but the average result in that study was that one international unit is equivalent to 2.7 genome equivalence.
This level was -- if you look back then, to the proficiency study in 1997, we found that the dilution equivalent to this detection or this concentration, was not achieved -- or this detection limit was not achieved by roughly two-thirds of the laboratories that were proficient.
So one-third of the laboratories at least in 1997, was not able to comply with the EMEA regulation. And also if we look at the WHO study then also you look at a 95 percent hit rate cutoff point in that study, 60 percent of the laboratories were able to achieve their limits. So again, 40 percent was not.
And of course there are new methods now available such as the NucliSens extractor that is under validation of CLB; where you have sort of plastic cartridge where you can use silica-based extraction with a number of steps, and the silica particles are then trapped in with the filter and we end up with a 50 microliter L weight of RNA.
When we look at the manual, NucliSens extraction kit and the extractor we have seen in dilutions of the panels that we have that the two methods were not statistically significantly different so that they were roughly the same.
Here you see the detection limits found with this NucliSens extractor. Here you see the total number of results -- the 100 percent detection limit was here about 100. And this of course, is the NucliSens extraction method in combination with the 1.0 Version of the Amplicor kit. And we now move to the 2.0 Version, so we -- I come back on this later.
So if you then calculate as we talked about, that the 95 percent hit rate cutoff points or the 50 percent hit rate cutoff points and we look at the regular Amplicor in which in fact, a five microliter plasma equivalent is amplified, you see that the -- where we then modify this method by using the NucliSens extraction system, that you will increase the plasma equivalent which is amplified from five microliter to one ml.
And doing this you see that there's almost 100-fold increase in sensitivity to a level of about 32 to 60 genome equivalence at a 95 percent detection limit.
Now, if Gerold will help me with showing a few overheads. Well, I told you that we are coordinating a sort of multi-center validation study for not only the OMCL network but for all that are interested in looking at sensitivity.
For instance, the Qiagen robot or the NucliSens extractor, in conjunction with the Roche COBAS Amplicor system or the Roche COBAS system as it is, or the Gen-Probe TMA, and of course also in-house methods.
We use a panel that is called the Pelicheck panel where we have these dilutions with include dilutions which is the 100 international unit level or the 100 genome equivalence per ml level, etc. And you see here the data that we have found as a sort of QC test at CLB where we did four test runs, and you see that the 100 percent detection limit here was about 11.
We also have a similar sort of panel which is based on the EuroHep type 3 standard. Again you see here the genome equivalence levels and the international unit levels, and also here you see that is quite sensitive.
So the idea is that this panel can be used by multiple sites. So for instance, tests are done on this panel in some of the larger test centers in Switzerland, Germany, and England and in Finland. And it's possible to either test it eight times or 24 times by using these panels which contain multiple aliquots per dilution.
The results are an effort coordinated by Theo Kuypers at our laboratory which works together with Oregonal Teknika, where this validation is a standard with the other items which should be looked at, like specificity and robustness, and where either the type 3 or the type 1 panels are tested in parallel, or even the WHO standard in parallel with the genotype 1 standard dilution panels.
So this gives you an instrument for doing validations in your in-house system, making use of commercial test components.
And doing this then, we will at the same time, validate a number of run controls of which we then know what will be the detection rates on those dilutions in the different commercial methods that I used.
Thank you.
CHAIRPERSON TABOR: Thank you. The next speaker will be Dr. Yu from the Food and Drug Administration.
Return to Table of Contents
DR. YU: I will talk about the standardization -- CBER perspective.
Nucleic acid test NAT methods used to test plasma pools are considered by CBER to be donor screens. Current proposed limit by CBER is 100 copies per ml for testing mini-pools. So we need analytical standards to evaluate sensitivities because of diverse NAT methods and also varied pool size. They're as small as 20 mini-pool units or 1200 mini-pool units.
Currently we have two standards. The first one has already been available since 1995. This particular standard, second one, is currently being evaluated and I will talk about both standards right now.
As an interim measure the low titer -- you know, the low titer HCV RNA-positive standards, this is really an IG lot that is really spiked with the HCV RNA. But as an interim measure on December 27th, 1994, CBER instituted laboratory testing of HCV RNA by PCR on all lots of immune globulin products that have not undergone validated, viral inactivation removal steps during manufacture.
So this was formulated then, and then as I said, it's 16 percent IgG concentration spiked with mixed genotypes; in fact contains all six genotypes of HCV.
We filled 2,000 vials, three ml per vial, and store currently at 80 degrees Celsius. And we use this as a really, a low level positive control for laboratory testing. We also provided to manufacturer and other testing laboratories.
And determining a consensus level by collaborative studies, right now we haven't really finished the calculation but it's about 100 copies per ml. But because it's low, very high protein concentration it's really not suitable as a standard for plasma pool testing, so we have to formulate a new one.
We had a window as a single donor unit but it is corrected during window period. This passable unit was EIA2-negative -- EIA3-negative as well -- RIBA2 was indeterminate but RIBA3 was negative. The HCV RNA copy number is about 107 -- it's more than 107 copies per ml, and genotype is 1B. This is by sequence analysis.
So we sent this window period plasma unit to other participants, to other testing laboratories, and to ask them to quantitate for us.
First lab here, they gave us the result in terms of genome equivalence per ml, whereas all the rest of the laboratories gave us the results in terms of copies per ml.
So other than just one lab which has a very low value, most of them were around this range. So the average value, it's about 5 X 107 copies per ml. So we used this window period unit to formulate a plasma pool standards.
These are the participants: Bayer, Centeon, Chiron, NGI, Roche, and we also doing the testing ourselves with their methods as you have heard their presentations this morning. So we formulate an HCV dilutional panel consisting of ten members.
The HCV panel stock was diluted with a defibrinated human plasma pool to formulate ten panel members containing zero to 105 copies per ml of HCV RNA. And we filled the first member 4,000 vials.
Each one of them is .75 ml per vial and it's targeted to contain 1,000 copies per ml, whereas the rest of the panel members is about 2,000 vials -- we filled 2,000 vials. Again, it's about .75 ml per vial. Right now we store at -70 degrees Celsius.
So we sent these panels consisting of ten members to eight laboratories and asked them to give us qualitative results as well as the quantitative results. Now, number 2 and number 5 are just filled with the diluent, so these are negative. And all laboratories, all test negative; no false positive at all.
As you can see now, the laboratory A was able to detect all the positive, you know, members; the lowest one they can detect. And the B lab has qualitative and quantitative assay methods. The qualitative assay method is much more sensitive and they can detect all. The quantitative only can detect up to 500 copies per ml.
But as you can see here, most of -- in fact all the labs can detect 500 copies per ml -- I mean, this at the targeted level all of them were able to detect 500 copies per ml -- except this laboratory which is I. This is a branch DNA assay. The cutoff is 20,000 copies per ml so they cannot detect any of these panel members.
As you can see here, some of the laboratories were able to detect 100 copies per ml; quite a few laboratories. Now, some of them may not be 100 percent hit here. One of the laboratories, 6 out of 8 positive, and then so forth. But again, some of them can even go lower; below the 100 copies per ml.
We haven't got all the results yet, but we've got about almost all the data. Again, they provide us the mean value. We haven't really got the raw data and so forth to do the statistical analysis.
These are the participating labs that provide us the results. I just run through the tests very quickly -- I mean, the results of all the panel members, what are the levels of them.
This is panel member number one; the targeted level is 1,000 copies per ml. And this is the target level right here; solid line here. As you can see, all the labs are pretty close. I mean, the variation is about half a -- within half-log difference. These are half-log scales.
Except one lab is a little lower. Then this particular -- this is panel member number one. We sent 150 vials to Dr. John Saldanha and to participate in his collaborative studies. And I guess soon we would know how much in terms of international units for this member number one.
This is member number 3. The target level is 100,000 copies per ml so it's quite high level. As you can see all the labs were pretty close. The variation is within half a log.
This member number 4 is 10,000 copies per ml. Again, most close to the target level, 104, except one lab is over half a log lower.
This is member number six. The targeted level is 500 copies per ml. The two labs were a little lower than the half-a-log; the difference.
So in the 100 copies per ml, this is for member number 7, this is the targeted levels. You can see there are two labs cannot provide us the data. Now they have -- they can detect them positive because qualitative assay is much more sensitive, but they cannot provide us the quantitative data. So that's close.
And then number seven, 50 copies. There were four labs were still -- these are our laboratories that -- all four labs were able to provide us the data and close to the targeted level which is 50 copies per ml.
And this is ten copies per ml, and only two laboratories were able to give us the data.
And five copies per ml -- the lowest one -- the target level five copies. And again, there is only two laboratories can provide to us.
So in summary, an additional panel for HCV is being developed and a difficult sensitivity of diverse NAT can be compared, and it can be used to the laboratories of NATs for HCV RNA. Current proposed limit, 100 copies per ml for testing mini-pools is achievable.
I would like to acknowledge those that contributed to the work. Mr. Mason and Dr. Tan in our laboratory. And Dr. Fashid, Dr. Yuwen, Dr. Hsia, and Dr. Tabor. And they participated -- they are the ones in our laboratory and Dr. Tabor's laboratory collaborated this formulation and testing of these panels -- the new proposed panels.
I also would like to acknowledge all the participants -- all the participating laboratories for their collaboration. Thank you.
CHAIRPERSON TABOR: Thank you. Now Dr. Nubling will speak again; this time on standardization.
Return to Table of Contents
DR. NUBLING: Okay, for me, quite easy because I think the most important things have already been said by the previous speakers. But these, the topics remember reference preparations and I think -- the topic of this section is standardization and reference -- common reference preparations are one, option to obtain standardization; common requirements for routine assays are a second option for obtaining standardization.
Nevertheless I want to first come to the first point. We've also, as I've already mentioned, prepared so-called (unintelligible) HCV RNA. Like Dr. Yu just told some minutes ago, we also use window phase donations which was negative in all on the HCV test but positive by PCR, and was quite different a titer in the CBER preparation; namely 107 copies per ml.
This was diluted and several thousand aliquots were lyophilized. During lyophilization we had a loss of HCV RNA approximately in the range of 30 percent. By quantity PCR we have 105 copies approximately, already still available in the lyophilized ampoules, and this corresponds to approximately 3 X 104 international units but the exact unit shall be given after the evaluation of the collaborative study organized by John Saldanha.
We used this concentration for the reference preparation in order to have preparation which is appropriate for spiking of plasma pools, as well as give a suitable for preparing run controls. And we will provide this preparation once calibration is finished to all the blood donation centers which require it, at least in Germany.
Nevertheless, we also encourage blood donation centers to calibrate their own reference materials. And this is just a table how we recommend them to perform such assays: 24 replicates per dilution points. Here are two dilution series.
One is a WHO center HCV RNA dilution series -- different dilution factors, international units per ml. And the second preparation is an in-house HCV preparation of a blood donation center in Germany.
And for both preparations dilutions were performed and for each dilution point 24 replicates were analyzed until the dilution range where negative results were obtained.
By probit analysis of these data it is possible for us -- this corresponds to the WHO standard -- first to establish or to calculate the 95 percent cutoff of the assays in international units; and the second possibility is to by potency comparison of these two jobs to get the unit of the in-house reference preparation.
So that's the system how we recommend it and how it obviously, also works. The next point, how standardization is obtained is by common requirements for routine HCV NAT. As I told before, we require a run control which should control the complete NAT procedure starting from lyses ending with detection, and should contain even HCV particles.
And that's also the reason we are not very happy with the amplicor tests as it is now because there's only RNA which is used as a positive control and also is the internal control of course -- is only RNA.
The HCV RNA concentration in the run control depends on the number of donations per pool, and indicates for the regulations for erythrocytes and thrombocytes should reflect the 5,000 international units; meaning a pool of, for example, 50 donations should have 100 international units as a run control as a parallel assay for each run.
Also, negative controls are required for detection of contamination, but I think that is not the most important issue because most of the samples are expected to be negative as well.
And the last point, standardization, that we require also inhibition control for pools consisting of more than 50 donations -- and then this again. The background for this requirement, experiments which are just a time ago but nevertheless, I think are quite valid.
We spiked commercial size plasma pools which were HCV PCR-negative with low amount of viruses and with 10-fold higher amount of viruses, and then we compared to extraction methods which was first the extraction method with the amplicor HCV kit 1.0 and the second extraction method was silica columns.
And both eluents were then amplified and detected using again the amplicoquet. And as you can see here, dependent on the extraction method we get false negative results due to inhibition. And this inhibition is dependent on the RNA concentration because with the 10-fold higher RNA concentration it doesn't go to zero but it decreases.
So I think inhibition control is an important issue and is also, if it's required generally a point of standardization for assays. Inhibition control can consist either of synthetic RNA which is co-amplified in the same assay, or three percent of parallel HCV spiking control, meaning that the same sample is tested spiked and unspiked with low amount of virus.
In case of inhibition we recommend to repeat the whole procedure and if it doesn't work, if inhibition still occurs, to use a different extraction method -- at least for these samples.
Okay, I promised to make it short. This was it. Thanks.
Return to Table of Contents
CHAIRPERSON TABOR: Thank you. We have time for five or ten minutes of discussion before the break. If anyone in the audience has questions please come up to the microphone and identify yourself before the question. Yes?
MR. MARTINEZ: Hello. My name is Bill Martinez. I'm from a community blood center in Florida. And from Dr. Hewlett's presentation there was a slide that described the ways that NAT testing should be undertaken. And it described them as a home brew kit or send them out.
Can the inference from that slide be that all blood drawn in the United States should be tested by NAT under one of those three options?
DR. HEWLETT: We do have an answer for that, though. No, it isn't that those are the only options, but those are the options we've been presented with. So that's what I was addressing in my slide.
MR. MARTINEZ: So the FDA is not going to require that all blood drawn in the United States be tested for the genome of HCV? Is that correct?
DR. HEWLETT: Are we requiring testing? Nucleic acid testing?
MR. MARTINEZ: That's the question, basically.
DR. HEWLETT: I think at this point there are no formal requirements but we are basically going along with the impetus from the European community, as you've heard today. There's a lot going on in Europe and that has, to a great extent, influenced our policy in regard to nucleic acid testing, and clearly if there's public health benefit then -- I mean, that's what drives whether it should be implemented or not.
CHAIRPERSON TABOR: This is a very unusual regulatory situation in that it's been driven entirely by the industry and by the pressures from changes that are occurring very rapidly in Europe, leading ultimately to the testing that will be in place by next year.
It's not being required per se, by FDA, but there's no question that market forces and the wish for the safest blood supply possible will almost certainly result in all plasma being tested in this way in a very short time.
I think it's also going to be driven by technological changes. We've heard today only about mini-pool testing but I think all of us expect to see technological changes that will permit individual unit testing in the not-too-distant future, and that will of course, change the whole ballgame.
MR. MARTINEZ: Okay, so the FDA just plans to go along with it?
CHAIRPERSON TABOR: Well, I would word that differently. I mean, we're responding to applications to evaluate and to validate and standardize these assays, and by the wish of the blood collecting community, to apply this new measure of safety.
MR. MARTINEZ: One more thing. Is there going to be a restriction on the number of INDs that can be presented to FDA for the purposes of studying this --
CHAIRPERSON TABOR: The only restriction, in reality, is on the size of the FDA staff, which is shrinking as we sit here. Otherwise, there are no restrictions.
MR. MARTINEZ: Okay. The blood centers that send their testing out, will they have to participate in an IND?
DR. HEWLETT: I'm sorry?
MR. MARTINEZ: Are there obligations of centers that send their tests out under -- defined in terms of what their participation is likely to be? Sorry about my questions.
DR. HEWLETT: Are you asking whether the blood center has to submit an IND?
MR. MARTINEZ: Well, if a blood center sends their tests out and what would they have to do in order to help provide data under IND?
DR. HEWLETT: Well, they would be one of the sites that is being tested by this service or by --
MR. MARTINEZ: So they would be part of an IND?
DR. HEWLETT: Yes.
MR. MARTINEZ: Okay. Thank you.
DR. KLEINMAN: Steve Kleinman. I'd like to make a comment and then if there are members on the panel that want to react to it. Today we've heard about GAT testing and NAT testing. We heard about master pools, maxi-pools, and mini-pools. We've heard about genome equivalence per ml, copies per ml, and international units per ml.
I think when we talk about standardization I'd like to urge some standardization of terminology because it's getting -- I mean, I think people in this room can follow it because we all have a special interest in it, but when it gets brought out to the general community we need to come up with definitions, and hopefully those definitions and terminologies can be international so we're all not having to question whether the terminology we use is the same as the terminology of somebody else.
I know international units and genome equivalence is a different story, but I think it would be nice and I think the studies are ongoing so that we can -- I mean, even in listening to these presentations it's difficult to kind of switch your mind back and forth between one speaker presenting genome equivalence per ml, another speaker presenting a standard that's international units per ml.
And all you have to do is lose your attention for a few seconds and you're not sure you've got the right thing down. So I think that's important.
With regard to things like master pool, primary pool, maxi-pool -- I think within the U.S. we definitely need definitions. Everybody needs to use those same definitions, otherwise we're not going to know what each other are talking about. So, any comments on that?
CHAIRPERSON TABOR: Dr. Saldanha.
DR. SALDANHA: Well, I can comment on two of them. One is NAT and GAT. I think we decided to call it nucleic acid amplification technique rather than genome, because you don't necessarily amplify a genome. You can do just any bit of nucleic acid.
With units I think you made a very good case for having international units and I think that's the whole point of setting up an international standard so that we're all talking about the same thing.
It's an arbitrary unit but as long as everything is calibrated against the international standard then I think we're all working to the same sensitivity. And I agree with you, it's very confusing. Because I'm not sure about the difference between copy numbers and genome equivalence.
DR. HEWLETT: Yes, I'd just like to add a few lines to that. I agree that definition of international unit or the terms of reference in terms of international units, copy numbers, and genome equivalence does need some degree of standardization.
But I think that's exactly what's going on and we're already seeing that people are making the effort to find some common ground, and I think in a couple of months or maybe a year from now we'll see a lot more in that area.
In regard to discussions about pools and about defining primary pools and maxi-pools and so on, I think we're just beginning to, again, just beginning to see data long those lines, and as these INDs mature I think we'll learn more about what definitions are actually going to work in the field.
So I think at this point we're in the situation of doing an interim assessment and I think the issues are very pertinent but we'll learn more as we go along.
DR. BIANCO: I'm Celso Bianco. I'm not going to touch on that one. I understood from our colleagues from Europe that there has been a certain change, particularly by the Paul Ehrlich Institute. Initially there was a definition of required sensitivity for the assay that is being used for the pool -- that was if I recall, first 400 units and now 100 units.
But I understood from you of a requirement that you would have to be able to pick up 5,000 of something units in the pool. Is there a change? Is there any standard that has been defined for a sensitivity for a pick up of a unit in a pool?
DR. NUBLING: I'm sorry, that's a misunderstanding, what you said. The 100 international units for the manufacturing pools are defined for the plasma industry. And you cannot define it in a different way because the pool sizes are quite different and so on.
And the regulation in Germany concerning erthyrocytes and thrombocytes defined 5,000 international units for the single donation, and it depends on the technology which pool size is used or is usable. And even if you don't have to use pools; you can also -- some blood banks which make single donation PCRs already -- small blood banks.
DR. BIANCO: So what you're saying is that if the sensitivity of my PCR is 100 units, that I would pool a maximum of 50 units in that pool?
DR. NUBLING: Correct.
DR. BIANCO: But is that -- so you are defining the size of a pool?
DR. NUBLING: No, no. We are defining the minimal sensitivity for the single donation, but if we have a much higher sensitivity you can test by pooling, and the pool size depends on the sensitivity of the method.
CHAIRPERSON TABOR: We have time for one more question.
DR. SIMMONDS: Peter Simmonds. There are several standards and controls being discussed. One of the difficulties with any of these standards is that fact that they are derived from sero-positive donations ultimately; and obviously we know that these will concentrate differently when you spin them.
Some tests use direct extraction, but other ones are actually based on ultracentrifugation, and I don't think that a sero-positive donation or control is actually a fair test of this concentration step.
So what I was really wondering was, are there any plans to actually get some sero-negative standards sorted out? Because, you know, we can't really sensibly evaluate a test such as the new Roche test which has this ultracentrifugation set built into it.
DR. NUBLING: Yes, I agree totally with you and it's not correct; not all standards are sero-positive. For example, our standard is a window phase in a donation --
DR. SIMMONDS: Sorry, sorry.
DR. NUBLING: -- which is negative in all assays which we have used. And I had the same thinking in the morning when the person from Roche, the lady from Roche told the centrifugation step as an enrichment step, but the target samples are antibody-negative and the determination of detection limit was performed with some of the materials which are sero-positive. So it could be really a problem.
DR. SIMMONDS: I mean, was the new Roche test 24,000g for half-an-hour, is that right?
DR. NUBLING: Yes.
DR. SIMMONDS: An hour, yes. There's been actually a German paper in Vox this month actually, showing that you can't concentrate virus even spinning much harder than that if the donation is sero-negative. So I'm not sure how much -- for example, the Roche test has been tested, you know, with sero-negative control.
DR. YU: I just want to comment, you know, about ultracentrifugation. We did find when in the presence of ten percent IgG -- in the presence of antibody there is a tremendous difference between the pelleting efficiency, you know, with antibody and without.
CHAIRPERSON TABOR: Could you speak just a little louder? I think people in the back are not hearing.
DR. YU: I'm just, you know, with experience that we have in the -- what Peter just raised, the presence of antibody may have something to do with the pelleting efficiency. And so, you know, in our hands I know if it's ten percent IgG the pelleting efficiency is so poor in the absence of antibody.
But that is ten percent protein. And if it's five percent protein like what is in the plasma, I think you really need to evaluate in the presence and absence of antibody. I think the effect is not as great but it still needs to be evaluated.
DR. PSALLIDOPOULOS: Are there enough materials, you know, available to actually distribute standards which are sero-negative? I mean, Jens Bukh mentioned that there are chimpanzee pre-sero conversions -- some of those samples from different genotypes. I mean, are those suitable? What sort of volume are they? What sort of volume do you have?
DR. BUKH: Those chimpanzee pools are in two different kinds. The ones are the ones that's been tested for infectivity and they're primarily going to be used for vaccine studies or studies where infectivity is needed, so it's a cell culture. And we have about 300 ml or 300 aliquots of such.
And then we have later plasma pheresis units. Then we get slower to the sero conversion of course, that's going to be made available. And we will have 400 aliquots of each of these genotypes, 100 microliters each that were from the gatepost inoculation. And those are available for PCR testing, etc.
Now, of course they all have genome titers of about 105 to 106, so potentially dilutions could be made of them.
DR. CONRAD: I just want to say that through the INDs that we've done and looked at, we've found several hundred now and interdicted in several hundred donations that are antibody-negative and some of them are quite distal to the act of sero conversion.
And we obviously saved the whole plasma pheresis unit on those, either through Alpha, Baxter, or ourselves. And those should probably become available to you guys making standards, because those are large, large volume amounts of material that is antibody-negative.
DR. YU: For us, we also, you know, the HCV stock -- the one I'm just talking about is the window period units -- 1B and so forth. We do have quite a few, quite a lot in volumes -- can I say something? But anyway, in larger quantities. But again, you know, in CBER member number 1 we do fill 4,000 vials. But our essential purpose is really what we use as a standard for other -- for a lot of these testing or other things.
CHAIRPERSON TABOR: I think we need to move on. We're going to have a 15-minute break. When we come back would the members of the panel -- that is, those who are participating in the panel discussion -- please sit around the table prior to Dr. Bianco's talk? Thank you.
(Whereupon, the foregoing matter went off the record at 3:42 p.m. and went back on the record at 4:00 p.m.)
CHAIRPERSON TABOR: The next speaker will be Dr. Bianco from the New York Blood Center. And would all members of the panel please sit at the tables? Thank you.
Return to Table of Contents
DR. BIANCO: My role is to provide you with some general lists of issues regarding the implementation of nucleic acid testing by blood centers. And I'll try to do that since I was the only person in blood banking and transfusion directly involved -- related here, except for some of the presentations from the industry and Sue Stramer -- I'll try to raise all these generic issues.
First, Dr. Kleinman came with some subversive words a few minutes ago, and I had actually come with some -- to grapple with some of these issues. And I see that GAT is an industry favorite and NAT is FDA, and I realize also the favorite of the European regulatory agencies. We'll start calling it NAT, but for the purposes of this presentation I'm going to call it GAT/NAT.
Where are we today? I think that we had a superb review in this meeting of the status of the technology. Obviously, the technology is here in terms of the testing of each one of these samples -- be it the sample, the specimen derived from a pool or derived from an individual blood donor sample.
Unfortunately, the logistics are not here. There are a number of issues that we will have to confront. We have the benefit of having a task force that was put together under the umbrella of the American Association of Blood Banks, that has been trying to deal with some of these issues.
I'm listing here the names of the members of this task force just to tell you that there is representation from almost every sector of the transfusion bloodbanking area.
There are representatives -- Dr. McCullough is the chairman. There are representatives from America's blood centers, America's Red Cross, ABBA -- and it's been a very, very effective committee. And this group kind of kind of together -- there is a report, a short, initial report that is coming out in Transfusion, I believe next month, that lists the issues being addressed by the task force.
And among those global issues are the issues of the generic questions that we have. GAT is being implemented, particular in Europe, for the units that do not meet GAT. As plasma units they are going to be subjected to viral inactivation during their manufacture.
We are involved in collection and distribution of blood products that are going to be transfused into recipients of single units. They are not going to be inactivated by any process. They cannot be red cells at the present time or platelets being activated.
So we would like to apply this technology to the units that represent the real risk to our population of patients. What viruses should we test? Will every collecting facility have access to GAT/NAT? And how are we going to deal with some of the donor issues like autologous donors or platelet pheresis donors?
Platelet pheresis are special. There are no plasma products coming out of it; it's just the platelets. Should they be tested?
Some of these issues I think that are consensus are developing. Obviously, we are going to test for both HCV and HIV, even at the risk -- according to the data that we saw today -- is the one that will benefit the most. HIV is of such an importance in the minds of our patient population that I think that we cannot ignore it.
And there is also one of the manufacturers that has a multiplex assay and by default, that both viruses tested will be searching for the nucleic acid of both viruses, not the genome. And obviously, there of other of these type of issues.
There are many of the regulatory issues that are not exactly part of the recommendations that have so far been issued by FDA by CBER; either in written form or during our discussions. And we would like as we develop those INDs, to have a common sense among all of us of what are the best procedures? What are the procedures that are in the best interest of our donors and our recipients?
And we also have some concerns that these are tests; they are not yet licensed. We are in the process of IND; we are learning about them. And we are concerned that some of the approaches that we used today with regulated tests may be applied in a rush to some of these tests that are still experimental.
But the questions are there. Do we retrieve and quarantine prior donations? How are going to confirm? How are we going to notify? How are we going to follow up these donors? And a lot of it will be learned during the process of the IND.
A major issue with the whole blood donation is a question of timing. The plasma industry has actually the luxury in terms of safety, of retaining those units, as we heard today, for a period of months until they include these units in the processing for the manufacture of the plasma derivatives.
Our platelets have a short life -- shelf life of five days. If we look at the time cycle of everything that is happening, there is in a period of 36 to 48 hours we'll have a donor sample that will be subjected to the ELISA, and another donor sample -- probably in one of those fancy tubes that Dr. Stramer showed us -- that is going to be rushed into an airplane and flown into some laboratory somewhere.
When it gets to that lab we will have to remove repeat reactives on the ELISA. Now we'll have the results. So that the number of positive pools that we have is reduced; otherwise the frequency will be much higher.
And we would pool the negatives and then subject those samples, those pools to be GAT/NAT. And here it took us 36 to 48 hours. Even if we have in-house testing and pooling this will be close to the 36 hours.
Then we have a second cycle of 36 to 48 hours minimum; that is obviously if our pool here a few hours later is negative we can release all the negatives. But if we have to go to round 2 or round 3 to resolve which ones are the positive samples in the pool, we will come to release the negatives for transfusion at a minimum of 48 hours.
So timing in the best of all worlds, GAT/NAT results will come out in 36 to 48 hours. The GAT/NAT results may appear 48 to 96 hours before we started the process. And our platelets expire in 120 hours. And obviously there are the issues of contamination, sleet, rain, snow, and everything that affects the Post Office.
One of the issues that is being seriously considered among us is the possibility that we will have to transfuse a portion of our supply -- particularly the supply of platelets -- prior to the availability of the GAT/NAT test results; and that we will then create a system which, on a kind of a look-back but an immediate look-back following that transfusion, we would provide the physician and the recipient with the molecular technology results and help them deal with this fact.
But it will be a very serious balance. It will be -- even if the frequency of those events is rather small, as we heard the numbers today, maybe a couple of hundred a year in the whole country -- that certainly will not be an easy task.
And among the logistic issues there are many. There's the transportation. We heard some of the things about the needs for preservation of the samples; the pooling hardware. And what we see as one of the major obstacles at this point is software: is software for sample management; is software for pooling; is software for management of test results.
You heard for instance, the Red Cross that uses a single software for the lab, a single type of barcodes; that the need for special handling of those samples: the avoidance of duplicate numbers, and the management of those test results and the communications back to the laboratories.
If we size it up maybe we'll have a more realistic view even, of what this represented in terms of logistics. ABC centers and American Red Cross centers, each one of them collect about half of the -- about 45 percent each of the whole blood supply, and hospitals collect the other ten percent for those hospitals.
And the majority of them I suspect, are going to ship their samples either to the Red Cross or to centers among America's blood centers in order to have them tested by GAT/NAT.
There are, for each one of these systems, about 22,000 samples a day. ABC centers are 70 centers. They have multiple software, they have multiple communication systems, they have different barcode systems, and they have overlapping numbers. And those issues will have to be resolved between now and early next year.
The American Red Cross I commented already, but each one of these systems, depending on the pool size and our ability to deal with these test results and tests concomitantly, may have to find among the 22,000 samples, 700 repeatedly reactive samples -- about three percent or two percent, depending on what they decide to pull out -- before pooling every day.
And I don't think it makes a difference if it is 50 or 700. You look in this one here full with 22,000 samples, you probably will get to the end of those tables and you'll go look for those samples.
The number of pools actually, that is going to be run, is relatively small. For a pool of 24 you see that ABC centers in total are going to run less than 1,000 pools a day -- maybe a little bit more with the repeats and resolution.
American Red Cross centers are going to run less than 200 pools a day. So the staff for that with the Gen-Probe may be one, two people, but the staff that will be involved in the logistics and the training and the software and the investment is immense compared to the actual technology of test.
The initial plan -- and I'm giving you as an advance and without authorization directly from anybody -- but as it looks today we heard that the American Red Cross has made an association that seems to be very productive with Gen-Probe, and is preparing, of what I heard, a very nice laboratory in San Diego.
The 70 ABC centers made preliminary arrangements, both with Gen-Probe, that it apparently will set up four sites in four different labs scattered around the country. And Roche, that is planning to set up 11 labs. And we are starting to work together with those manufacturers in a collaborative manner, attempting to have as fast as possible, our INDs in very good shape and so that we can initiate those studies.
The likely, initial scenario is that most or all donors are going to be tested, ultimately, sometime next year. It's not easy or I would not feel comfortable personally, testing sample -- units that are going to be turned into recovered plasma for further manufacture, and not testing the red cells that are going to be transfused to a recipient in a hospital. I don't think it would be fair.
The size pools you heard today, American Red Cross is 128, followed by an interrogation mark. The plans from ABC, both with Roche and I don't know yet if confirmed with Gen-Probe, but the plan with Roche is 24.
The ARC plans to remove the reactive samples. America's blood centers, there's an intense search for the possibility -- in the majority of these centers, about 80 to 85 percent of the donors are repeat donors. and as you heard from Dr. Kleinman this morning, the incidence of sero conversion in repeat donors is rather small; about four per 100,000 person-years for HCV.
So if we could segregate these samples that come from first-time donors we would be running a very low risk of finding a positive pool if we tested it concomitantly with the ELISA test without removing the reactives. But this is part 2 of the plans, and this is directly linked to the pool size.
Well, we heard the orders from the European regulatory authorities and we will try to comply. We don't like April 1st so we chose March 31st.
(Laughter.)
And the European manufacturers of plasma derivatives where we ship a lot of our recovered plasma, required that we ship to them tested plasma by PCR after March 31st. And we plan to implement that testing in an IND format before that date.
However, there is a consensus that came up very strongly at this task force, and particularly after an impassioned speech by Dr. Roger Dodd, that this is an experimental test and that it's being performed under an IND and that we should not behave as if we had just received a licensed test in our laboratories, and implemented with all the fancy bells and whistles without actually knowing what the consequences are of what we're implementing.
So the implementation should be well planned and gradual, and in the first several months until the systems are fully understood, we should not rush in our opinion, to take action on results of the initial pools.
For instance if a pool of 128 is reactive or a pool of 24, of going back and notifying recipients of all the prior units that this donor now sero converted, we will be more careful in the early days and we won't take action against individual units or individual samples, or related to individual samples from donors that prove to be positive. But we will not try to rush into action for things for which we have low degree of certainty.
How can everybody that is in this room help? In our view, CDC and NIH, that is our public health system and the academic side, can help by continuing to support research in these areas. We need and we need a lot.
Yes, also by helping the public understand the benefit -- or relatively limited benefits -- of the GAT/NAT testing, through particularly, epidemiologic studies. And the somewhat of an illusion that exists in the mind of the public that the implementation of GAT/NAT in pools is going to close all the windows.
They will improve but it will not close the windows. And again, to help educate blood donors, blood recipients, physicians, and the American public about GAT.
The FDA also can help; it can help a lot. This meeting is an example. They also are providing substantial support to the AABB task force by validating a lot of our thinking and providing a very open and supportive discussion.
Definition of standards. I asked a question before about the 50 and one, and one -- but the issue is that the initial regulatory actions were very careful, and they referred to the sensitivity of the single assay.
But now we have been confronted with a pool, and we need support and guidance on particular issues of sensitivity related to pool size. My fantasy is very simple. I'd like Dr. Hewlett to give me a tube and say, this is HCV-positive; pick it up in your pool. If you don't pick it up your test is bad. If you pick it up your test is good.
And I wouldn't care if it is genomes or if it is copies or if it is units. And again, by supporting us in the establishment of validations. Can the industry help? Yes, if we continue behaving as we have so far. That is, we are all looking at it as a safety issue and as a scientific issues and not necessarily as a competitive issue.
We have shared a substantial amount of information. We are all supporting the inter-agency task force and we certainly are supporting the public health system. Thank you.
Return to Table of Contents
CHAIRPERSON TABOR: We are now scheduled to have a panel discussion. The panel is rather small at this point and I think maybe we should combine this with continuing discussion from the audience and questions for any of the speakers.
Maybe I'll -- while we're waiting to see if there are any questions -- I see Harvey Alter in the back; maybe we can get him to sit up here at the table. Let me just ask the people who are here now -- Dr. Seeff, Dr. Miriam Alter, Dr. Dodd, Dr. Bianco -- do you have anything to add to the discussion? Dr. Seeff.
DR. SEEFF: Well, it may be heretical to extend what is being discussed here to other areas and perhaps this is already in the process of being done. I guess I'm looking at the panel and I may be the only -- except for Harvey now -- clinician seeing patients.
And the next step of course, is our anxiety about getting standardization and licensing of tests for therapeutic purposes. This is -- again, this is perhaps the wrong venue but I think we're all aware of the fact that we're moving into a new treatment era and for the first time, genotyping and viral concentration may turn out to be critical -- actually critical -- in making decisions.
Not only is this with respect to who should be treated but for how long people should be treated. And I couldn't urge more from the clinical point of view, the need to have these tests available, standardized, licensed so we can get started and get moving for appropriate treatment.
CHAIRPERSON TABOR: Maybe -- I wonder if some of the representatives of industry who spoke this morning would comment on whether we're likely to get a test for therapeutic purposes in the foreseeable future? Do any of you have anything --
DR. MIRIAM ALTER: Why would that be different?
DR. CONRAD: Speaking for industry, yes. I think universally, industry is moving towards developing tests in the therapeutic domain, both -- I think my colleagues at Amplicor -- Roche as well as us and others are probably looking at it, or doing it now.
DR. MIRIAM ALTER: I'd like to add a comment to Dr. Seeff's in the same range, that I realize we're focused today on protecting the blood supply and making it as safe as possible.
But we can't bury our heads in the sand and not realize that this has far-ranging implications; both for reporting the test results back to the donor as well as for the public health campaign that's going to take place starting within the year, which will broaden the amount of testing for hepatitis C that's currently done in this country.
And regardless of regulations, we know that in practice PCR and it's being used widely to test and diagnose patients with HCV infection, and I think we need to recognize that as we go through this process.
Because these tests, as Dr. Seeff said, need to be made available in a standardized fashion, and if the recommendations don't come from us that patients are not getting a consistent message because they're not being interpreted in the same way from one place to another.
DR. DODD: I guess my comment would pick up with on what Miriam and indirectly, Leonard said, and also what Celso has said; and that is that I think we've learned today that we've moved a very long way down the track to implementing testing in a routine fashion, and that I think that we can probably reasonably well anticipate meeting the needs of the plasma for further manufacture component of our task within an appropriate timeframe.
And I think I'd say I'm very excited with the progress of the technologies made. And I can remember hearing Dr. Kessler say, I want you guys to tell me not if it can be done, but when it can be done and how. And I think we are further ahead than many of us thought when he uttered those immortal words.
But what I think we do need to think about very carefully, particularly as we involve labile products, red cells and platelets in this process -- and that's inevitable when we use recovered plasma -- is the human side of this.
The issues of how we're going to manage our donors, how we're going to deal with giving effective information about what is still a research test. I think we all have a lot of faith in what nucleic acid amplification really means.
But we really don't have enough mass experience to understand how often we're going to get an uninterpretable or a non-repeatable or a false positive. And I do think that we have to recognize that, at least for the single donor products, we are going to have some recipients on the other end of the process.
And I actually would like to see another workshop that deals with, if you will, the human side of NAT testing rather than the genome side.
DR. SEEFF: I'm sorry, if I could just extend what I said in fact, to again support what Miriam has to say. We are starting look-back, and this look-back is something that has real relevance to this group.
I think everyone knows here that the VA system, 172 hospitals, is starting its own look-back. And on the basis of which I am aware, it may be that something like 70 to 80 percent of everybody who attends VA hospitals is going to be tested for hepatitis C.
Once we come up with a positive result for antibody -- we've all got algorithms, and the algorithm then takes us to, are these people viremic? Do we treat those who are not viremic? And of course, most people will not do that.
We've got to have a test that can take us to the next phase. We're all busy drawing all these arrows all over the place, but if we don't have any licensed tests then I don't think that they have any meaning. This is extremely important.
CHAIRPERSON TABOR: Dr. Conrad.
DR. CONRAD: I just wanted say that after hearing Dr. Bianco's talk I'm thoroughly depressed. Thanks for pointing out the problems. But I also wanted to maybe bolster my spirits and others that, we've actually looked at three-and-a-half million donations from 100,000 people.
And to say that -- you know, there are logistical difficulties, there are enormous logistical difficulties. But I've noticed a difference from when we began this process several years ago and now, and I think that all those logistical difficulties are certainly not insurmountable.
And I think that as all of us gathered here should hear, that there is progress being made, and that process is extraordinarily rapid. And so I think that we can all acknowledge that there is difficulty and there is a human side, and we have notified donors and we have dealt with the ramifications of counseling someone that you have a virus and for 172 days didn't develop antibodies and what you do with that.
And we're learning. And I think the alternate to doing that and dealing with those difficulties is to burying our head in the sand and just not doing anything, and I think that's unacceptable. We have to understand that there's difficulties with the system but it's better than the alternative, which is to do nothing.
CHAIRPERSON TABOR: Dr. Hollinger.
DR. HOLLINGER: Yes, you know, there's also something very interesting about this whole issue. We've heard of probably a number of patients who might be out there who are in the window period and are transmitting disease. And we don't see much of this disease.
I know it sounds like heresy, almost going back to the 1970s when we told the blood bankers, you know, there is a disease out there, non-A/non-B hepatitis that's occurring. Oh no, no, there's no disease out there; we're not seeing it. And we found a lot of it.
But there are some unique issues here because everyone is looking for post-transfusion hepatitis now. There's a much greater, I think, view from the public -- at least from clinicians -- looking for this in people who have had disease.
So where is this disease in these patients? Is there something unique? We know that in these patients who are in the pre-antibody area who are obviously positive -- at least from a nucleic acid standpoint and presumably have infectious virus present -- but is there something unique about the disease that they're transmitting to individuals?
Is it more likely these patients do not become chronically infected? Maybe have a very mild or acute disease that goes on. I didn't hear from the industry -- particularly Centeon for example, produced some data that suggested perhaps one out of 1,000 of their donors were sero converting.
But what I'd really like to know is what happened to those that sero converted? I didn't hear anything about the clinical aspects of them. Did they develop acute hepatitis, did they develop jaundice, or things of that nature, on any of these groups that sero converted during following up of these plasma donors?
So there are two issues: one is what's happening to that group, but secondly, is there something unique about transmission of disease in the window period that may not be causing much in the way of significant disease? Maybe somebody can tell me if there are some patients that have been discovered over the last four years or five years from post-transfusion hepatitis C.
I've asked my colleagues, I've asked many people in this room to, you know, give me five patients that they know of -- or four patients or one patient. And I'm sure it must be true but I'd just like to hear that data.
DR. CONRAD: I can give you those patients. I mean, they're not patients -- they were donors -- but they must have gotten it acutely. I mean, we presented the window period -- well, they were PCR-negative, became PCR-positive; followed them through sero conversion.
And many of those patients developed elevated ALTs, and obviously we haven't followed them to look for the -- I mean, we saw earlier presentations where people went 20 years without significant disease. So we've had, you know, 200 days.
But I mean, I can give you lots of individuals who were HCV-negative by nucleic acid as well as antibodies; became HCV-positive by antibody -- by nucleic acids, then by antibody; and developed elevated ALTs.
DR. MIRIAM ALTER: Blaine --
DR. HOLLINGER: That's not my question. My question is not of the people who sero converted from a plasma -- I mean, I understand those will happen. I'm asking, the patients who received blood from those individuals, the recipients, what has happened to the recipients? I'm well aware that you can get the disease and you go on and get chronically infected. But I'm interested about the others.
CHAIRPERSON TABOR: Dr. Harvey Alter.
DR. HARVEY ALTER: A lot of comments. Overall, my general approach is sort of a Nike approach to this thing. Just do it and get it moving. And it's actually moving as fast as it possibly can, I think.
But Blaine raised an interesting question. I hate to agree with Blaine ever, but it is interesting, and these cases have disappeared. Since we've now followed 655 people -- it's a small number but we've followed them intensively since 1992 and there hasn't been a single case of hepatitis C. Miriam is seeing hepatitis C, acute cases, disappear in the community.
In fact, I would like to delay testing a little bit so we can get a few more acute cases to study. It's just you just can't find them. So that's one comment. But I think we have to do this, but the bottom line is that this is an intensive, expensive, complicated effort to narrow window which is going to prevent a probable handful. It's predicted a couple of hundred but it may really be a handful of cases. We're so close to zero incidence right now.
DR. DODD: Can I expand the question just a little bit and ask the clinicians among you whether you would treat these window period individuals that are going to turn up as a result of screening? Would you treat them for HepC?
DR. HARVEY ALTER: Well, the approach we've used now -- and Leonard can probably answer better -- but I think in people who have needle sticks or if we knew of somebody with acute hepatitis, would be not to treat them at the point of potential exposure but to treat them as soon as they became PCR-positive.
I think that's a -- there haven't been good studies to prove that's efficacious but there have been some hints of that, and if I were the one infected I'd want to be treated very early.
DR. SEEFF: Well, there's no answer to this. Well, there are some studies I guess, in Europe where people with acute diseases were treated and the sense was that there was a reduction in progression to chronic liver disease. But there really -- we don't have enough data.
I know that Miriam has been trying to set up a study, a needle stick study, and others. But I'd go along with what Harvey says; that if I were exposed I would have myself followed and the minute I became PCR-positive I would like to be treated. I don't know how, but I would like to be treated and I think the --
(Laughter.)
DR. MIRIAM ALTER: Which brings the issue back to where we were originally in that, although we're here to protect the blood supply, in fact, post-transfusion HCV infection is an extraordinarily rare event.
And as you two just brought up, this is driving an issue that really has an impact -- has a much broader impact in other settings. And that's the diagnostic setting, not the blood safety setting.
And it's because of this that perhaps we're going to get licensed tests that can be used to test the healthcare worker after the needle stick, and to screen other people who have risk factors, and to use in the clinical setting. In the absence of an antigen, an IgM test and a variety of other things that tells whether or not a person's viremic.
CHAIRPERSON TABOR: Dr. Lelie.
DR. LELIE: Yes, I want to comment on this discussion. I think there -- we talked about standardization of blood screening tests, but there's also a need for standardization of diagnostic tests. And we tried to make that link by of course, having our standards calibrated against the WHO standard as the primary reference preparation, but not only as the primary reference reagent for blood screening tests but also diagnostic tests.
And so the proficiency studies that are planned, it's also industrial use own group, that contacts of -- brings us in contact with the manufacturers of the NAT kits. So that there is the panels that are then approved of also by the industry.
And so if we use this diagnostic test also for confirmation I think -- and just an idea, but I'd like to have a comment on this. Since we are now collecting PPT samples that are in fact, PCR (unintelligible), this opens the possibility to use the same cubes to immediately, after doing RIBA or Western Blot testing, also use the same test on the NAT kits for confirmation.
And this would certainly improve the quality of donor counseling I think, in this country. Because I think there are a lot of donors counseled, even though they are false-positive in a sort of confirmatory test procedure that is applied now.
And I think if those PCR kits can be used for confirmation testing this would certainly improve the quality of counseling donors in the United States.
DR. MIRIAM ALTER: I'd like to just follow up. I'm glad you brought that up because I'd like to introduce a word of caution in the use of the term "false positive". It means that the person did not test positive for the virus.
It does not mean that the antibody was a false positive. It does not mean that the person was not infected with HCV in the past. And I think that we have a responsibility to the individual we're testing to tell them what -- the meaning of their test results and not just that they weren't virus positive on the day that we happened to test them.
One, it is with this type of disease this person could develop chronic hepatitis down the line at some point, if in fact that antibody is truly positive. And so I think that the RIBA or other supplemental assay still plays a major role in the counseling messages that you give an individual and in their need for further medical evaluation. Maybe not in protecting the blood supply but certainly for the donor.
DR. KLEINMAN: Yes, Steve Kleinman. I'd like to follow up on two of the points -- the first one that Miriam just made -- and that is, I think what Nico is saying is that amongst our antibody-positive, PCR-negatives there are probably two groups of people: those who have been exposed in the past and cleared the infection, and those who are truly false positive.
I mean, they have never been infected. Our RIBA tests are not perfect. They use the same antigens as our EIA tests; they just spread them out in a different format. I'm not sure that confirmatory is necessarily the right word for them. They're supplementary for sure.
So I think that's probably only a small number of individuals but we haven't really proven whether that antibody-positive person by RIBA who's PCR-negative and he comes back in and remains PCR-negative, whether he's ever had past HCV infection. That's an inference. I agree with you; we should still counsel people.
The other things was, in Blaine's comment about where are the transfusion associated hepatitis cases. I think the numbers from modeling would suggest that obviously those numbers are very small. We're talking about 100 units a year or so that are donated in the window phase, maybe made into an average of 1.5 components each; therefore exposing 150 recipients in the entire United States.
We know that 60 percent of those units go to people who never survive their initial hospitalization. So now we're down to 70 or so recipients. We know that three-quarters of the cases of non-A/non-B are anicteric. I mean, we're really talking about 15 cases a year throughout the United States.
Now, do they get reported back as post-transfusion, non-A/non-B -- as post-transfusion hepatitis? Well, I don't know what the reporting rate is, you know, but it would have to be -- if it's less than half -- I mean, very few of these cases are going to be reported back to the blood center and very few clinicians are going to see them.
So I think it's consistent with the fact that we don't see them. We wouldn't expect to see them at this point, and it certainly supports the point that it's a rare event and yet everybody has sort of, I think agreed, at least tacitly if not explicitly, that because we can narrow the window it's something that we should do. But that's why we don't see them.
DR. MIRIAM ALTER: Actually, I have a question for both the agency as well as industry. And that is, are the proposed studies that are going to be done only going to address pool testing or are they also going to address individual testing, such that they would be useable in the clinical setting? Diagnostic setting?
CHAIRPERSON TABOR: Will you respond to that, Dr. Ticehurst?
DR. TICEHURST: Yes, more or less. I'm John Ticehurst and I work at the Center for Devices and Radiological Health in FDA, as well as Johns Hopkins. And I'm here today largely as an observer out of interest, but a lot of the issues that particularly Drs. Alter and Seeff are bringing up are what we deal with at CDRH.
This is sort of internal FDA stuff that isn't very interesting but it's very relevant to the discussion. Which is that for these particular assays, the indications that pertain to the diagnosis and management of individual patients is going to be evaluated at CDRH.
That doesn't matter for anything except that it's administratively handled separately and the companies have to deal with it separately. And to just make a comment: we're not really hearing from the companies on this.
There's been very little of that and I know from my Hopkins point of view and from talking with colleagues and going to virology meetings, I'd say what was stated before a little more strongly. HCV RNA is a standard of medical care in this country right now but it's an unstandardized standard of medical care.
CHAIRPERSON TABOR: Miriam was saying something sotto voce and maybe if you could say it in a microphone?
DR. MIRIAM ALTER: You mean you can't say it? Is there anyone from industry who'd like to respond?
MR. WESOLOWSKI: I'm Alex Wesolowski. I'm Senior Director of Regulatory Affairs at Roche Molecular Systems. In response to Dr. Ticehurst's comments, I do know a couple of members of industry who had met with the Device Center -- Roche is one of them -- discussing these very tests.
We're looking at submission of a qualitative and a quantitative test for hepatitis C, diagnostic testing as well as the work that we're doing in blood screening. So they are on the way. I know Dr. Ticehurst can't talk about specific cases; however, since we're very close we're usually a little more joyful about making these announcements.
We're in the late stage of doing some diagnostic, clinical studies and we expect that these products will be available, approved by the FDA relatively soon.
CHAIRPERSON TABOR: Dr. Bianco.
DR. BIANCO: Miriam, we have a deadline and we want full attention from the agency, so the diagnostics will have to wait a little bit.
(Laughter.)
CHAIRPERSON TABOR: Dr. Alter.
DR. HARVEY ALTER: I'm President of Alter, Incorporated, and I am working on --
(Laughter.)
We are going to hopefully, do individual -- since we have such a small donor population we're hoping to do individual testing if I can get the right cooperation from companies, just as a pilot to see how this would work on line in getting platelets out, etc. That's one thing.
The second thing is that not too far behind, as I talked about before, is viral inactivation of cellular products, which if given the same impetus as this kind of testing, could move just as rapidly.
And the data are now -- we just finished the chimp data so I'm saying this with a bias because I'm involved in it -- but we've just finished the chimp data which show that this viral inactivation will kill HBV and HCV and prevent disease in the chimp model.
And there's already been a plethora of in vitro data. So I don't know how that's going to impact. You're going to have two powerful systems for preventing a minuscule number of infections. And will the system support both or will we have to choose one from the other? I think the next year is going to be very interesting.
CHAIRPERSON TABOR: The capacity of the system to support these things is astonishing. If you think back a few years when everyone was happy that we were detecting hepatitis C for the first time, now the -- basically the operating principle is to have zero risk if at all possible.
DR. SEEFF: Do you have a means of doing whole body viral inactivation of humans?
DR. HARVEY ALTER: Oh, yes, yes we do. We're working on it. My goal is to go from zero risk to zero meetings.
DR. BIANCO: But certainly, I think that at least an approach I think that -- I find pooling is a very interesting intermediate step because it's giving everybody the opportunity to think, to confront the issues, to confront the technological issues in a reduced number of samples.
That is, we saw from 50,000 samples a day we are going to reduce everything maybe to 1,000. And that is going to allow these systems to move.
DR. HARVEY ALTER: Yes, I agree it should be done that way, but we'd only be testing one pool a day.
Return to Table of Contents
CHAIRPERSON TABOR: Are there any final comments? I have a -- I wanted to show one or two slides just in closing.
I think it's important to take a moment to think about the fact that the concept of pooling in a series of pools, starting with a group of donors making one pool and taking -- say if these were pools of 25 donors each -- taking several pools of 25 donors each and making a tertiary pool and so forth and then testing it, and then working back from a positive test to identify the actual donor is a real paradigm shift.
I mean, it's one of those concepts, one of those simple concepts that changes the way we do things. And this is certainly going -- is obviously going to have that effect and it's basically the availability of this concept that's allowed us to do the kinds of things that have led to this meeting today.
And this is just based on a slide of Dr. Conrad's showing the cubicle concept that they use and working back. And I think it's possible that Dr. Conrad may even be the one who's responsible for this concept, but it's really enabled us to do this kind of testing.
And even if we're only eliminating a small number transfusion-transmitted cases, we're able to do something to improve the safety of blood that is what the public wants today.
I'd like to say just a few words in closing about the FDA's approach to nucleic acid testing. I think it's fair to say, at least unofficially, this is really just an interim measure until we're able to test individual units.
But we do feel strongly that it should be regulated as donor testing. We feel that it's part of our responsibility to the donors and to the concept of donor safety that donors should be notified if they've been found to be positive.
And it's interesting that in the last year-and-a-half or so that we've been discussing the concepts of nucleic acid testing with industry that this concept of nucleic acid testing of mini-pools as donor testing seems to be fairly -- pretty much accepted by everybody involved.
And I'd like to also point out that there are really important public health benefits from notifying the donors who would be identified through nucleic acid testing in mini-pools.
First of all, we would be able not only to prevent the donation -- the transfusion of the donated materials at that time, but we would be able to prevent repeated donations during the window period that might affect other individuals.
Notification of these donors would allow us to prevent transmission to close contacts, and in addition it would allow donors to seek therapy early.
Well, I think we've had a very long day and I think the talks have been terrific and the discussion has been stimulating. I think on that note I'll close the meeting and I'll see many of you at the Blood Products Advisory Committee tomorrow.
(Whereupon, the meeting was adjourned at 4:52 p.m.)
Return to Table of Contents
|
