• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Advisory Committees

  • Print
  • Share
  • E-mail

December 14, 2010: Blood Products Advisory Committee Meeting Transcript

FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
BLOOD PRODUCTS ADVISORY COMMITTEE

December 14, 2010

The Hilton Washington DC North/Gaithersburg
620 Perry Parkway


This transcript has not been edited or corrected, but appears as received from the commercial transcription service. Accordingly, the Food and Drug Administration makes no representation as to its accuracy.

Proceedings By:
CASET Associates, Ltd.
Fairfax, Virginia 22030
(703)266-8402

TABLE OF CONTENTS

 Page
Opening Remarks, Blaine Hollinger1
Statement of Conflicts of Interest1
Topic I: Risk of Dengue Virus Infection in Blood
Donors
6
Introduction and Background of Dengue Virus Infection -
Deborah Taylor
6
Dengue Virus Epidemiology in the U.S. and its Territories -
Kay Tomashek
11
Risk Model to Define Rate of Infectious Units During Dengue Virus Outbreaks: Endemic vs. Non-Endemic Areas -
Lyle Petersen
24
Update on Dengue Virus Panel Development -
Mario Rios
35
Overview of Data on Blood Donor Testing and Transfusion Transmission of Dengue Virus -
Michael Busch
44
Experience with Dengue Virus Antigen Tests -
Harold Margolis
56
Recent Experience in Testing Blood Donors in Puerto Rico and Key West, FL -
Susan Stramer
68
Open Public Hearing88
Open Committee Discussion102
Questions for the Committee135
Topic II: MLV-Related Human Retroviruses and Blood
Safety
165
Introduction and Background -
Indira Hewlett
166
Recent Studies of Epidemiology of MLV-Related Human Retroviruses U.S. Study,
Shyh-Ching Lo
174
U.S. Study, Maureen Hanson197
U.K. Study, Judy Mikovits210
Summary of Current Research on MLV-Related Human Retroviruses and Disease Association -
Jonathan Stoye
226
Animal Studies: Potential Transfusion Transmission of MLV-Related Human Retroviruses -
Francois Villinger
245
Update of Blood XMRV working Group Activities -
Graham Simmons
261
Prospective and Retrospective U.S. Donor Surveillance Studies -
Michael Busch
276
Assay Development Efforts on MLV-Related Human Retroviruses -
Rachel Bagni
287
Open Public Hearing300
Questions for the Committee340

PROCEEDINGS

Agenda Item: Opening Remarks"

LCDR EMERY: Welcome, everyone, to the Blood Products Advisory Committee. This is our December 14 and 15 meeting, and we are about to get started. I'm going to turn the time over to Dr. Hollinger, our chairperson.
DR. HOLLINGER: Thank you, Bryan. Bryan mentioned that this is the 99th meeting. Can you imagine? Next time it will be the 100th meeting. We should have a party or something for the 100th meeting. Maybe FDA can fund that, maybe where it's warmer.
We're glad you are here. We have a really full day today. I think to start, we will go around with the committee members and have the committee members introduce themselves and where they are from.
(Introductions)
DR. HOLLINGER: I'm going to ask Bryan to read the conflicts of interest and any other announcements.

Table of Contents

1

Agenda Item: Statement of Conflicts of Interest

EMERY: Welcome. This is the FDA conflict-of-interest disclosure statement, Blood Products Advisory Committee, December 14 and 15, 2010. The Food and Drug Administration is convening the December 14 and 15, 2010 meeting of the Blood Products Advisory Committee under the authority of the Federal Advisory Committee Act of 1972. With the exception of the industry representative, all participants of the committee are special government employees or regular federal employees from other agencies and are subject to the federal conflict-of-interest laws and regulations.

The following information on the status of this advisory committee's compliance with federal ethics and conflict-of-interest laws, including but not limited to 18 USC and 712 of the Federal Food, Drug, and Cosmetic Act, is being provided to participants at this meeting and to the public. FDA has determined that all members of this advisory committee are in compliance with federal ethics and conflict-of-interest laws. Under 18 USC, Congress has authorized FDA to grant waivers to special government employees and regular government employees who have financial conflicts when it is determined that the agency's need for a particular individual's service outweighs his or her potential financial conflict of interest. Under 712 of the Food, Drug, and Cosmetic Act, Congress has authorized FDA to grant waivers to special government employees and regular government employees with potential financial conflicts when necessary to afford the committee their essential expertise. Related to the discussions at this meeting, members and consultants of this committee have been screened for potential financial conflicts of interest of their own, as well as those imputed to them, including those of their spouses and minor children, and, for the purposes of 18 USC, their employers. These interests may include investments, consulting, expert witness testimony, contracts and grants, CRADAs, teaching, speaking, writing, patents and royalties, and primary employment.

The committee will discuss for topic I the risk of dengue virus infection in blood donors. This is a particular matter of general applicability. For topic II, the committee will discuss xenotropic murine leukemia virus, related human retroviruses, XMRV, and blood safety. This is a particular matter of general applicability. The committee will hear an overview of the research programs in the Laboratories of Hemostasis and Plasma Derivatives, Division of Hematology, Office of Blood Research and Review. This is a non-particular matter. In addition, the committee will hear updates on several topics. The updates are not for discussion by the committee, and therefore committee members were not screened for financial interests relating to the updates. Based on the agenda and all financial interests reported by members and consultants, no conflict-of-interest waivers were issued under 18 USC or 712 of the Food, Drug, and Cosmetic Act.

Dr. Celso Bianco is serving as the industry representative, acting on behalf of all related industry. Dr. Bianco is employed by America's Blood Centers in Washington, D.C. Industry representatives are not special government employees and do not vote. This conflict-of-interest statement will be available for review at the registration table. We would like to remind members, consultants, and participants that if the discussions involve any other products or firms not already on the agenda for which an FDA participant has a personal or imputed financial interest, the participants need to exclude themselves from such involvement, and their exclusion will be noted for the record. FDA encourages all participants to advise the committee of any financial relationships that you may have with any firm, its products, and, if known, its direct competitors.
Thank you.

DR. HOLLINGER: Thank you, Brian. I do want to welcome everyone here today. We do have, as I said, a full day, this morning particularly, with dengue virus, and this afternoon with XMRV or MLV-related retroviruses. I want to just remind the committee that we deal with scientific issues and evidence, and provide advice to the FDA regarding questions that are put before this committee. We have asked the speakers if they will leave one or two minutes at the end of the assigned time that they have so that the committee members can ask some questions right at that time. We won't be able to fully get into that. We'll do this later on in our discussion. But I do want to leave just a little time if somebody has a question that they want to ask the speakers for the day. We will keep, as much as possible, everyone on time, so that we can move forward, just because there is so much to talk about today.

With that, we'll begin with the first topic, which is on the risk of dengue virus infection in blood donors. Again, I want to emphasize that this committee deals with the safety of the blood supply in this country. That's what we are going to deal with today, regarding these particular topics. Without any further comment, we'll start with our first talk today, "Introduction and Background of Dengue Virus Infection," and that will be by Dr. Deborah Taylor.

Table of Contents

6

Agenda Item: Topic I: Risk of Dengue Virus Infection in Blood Donors
Introduction and Background of Dengue Virus Infection

DR. TAYLOR: The FDA is seeking the advice of the Blood Products Advisory Committee on an appropriate strategy, including donor deferral, to mitigate the risk of dengue virus infection in recipients of blood and blood components for transfusion, and the potential utility of donor screening either by a nucleic acid-based test, an antibody test, and/or an antigen test, given the current technological limitations.

Dengue virus is in the flavivirus family, closely related to West Nile virus, yellow fever virus. It is composed of four separate viruses that are closely related, dengue 1 through 4. Its genome is about 11 kb, and it produces 10 proteins, three structural proteins and seven nonstructural proteins. I just point out NS1 for your information. I'll talk about it later. Dengue virus is the most common vector-borne virus. It threatens 2.5 million people worldwide. More than 50 million infections occur each year and more than 24,000 deaths per year. It has expanded its range and increased worldwide, due to urbanization, population migration, population growth, global travel, climate change, and ineffective mosquito control. Dr. Tomashek will also talk about the points that have caused dengue to expand. Autochthonous transmission has occurred in the U.S., in Texas, in Puerto Rico and the Virgin Islands, in the Pacific Islands, and in Florida.

There are two major mosquito vectors -- both of these occur in the U.S. -- Aedes aegypti and Aedes albopictus. You can see by the map below that both occur in the eastern and southern U.S. Again, Dr. Tomashek will give us updated information and talk more about the vector. In primary dengue infection, the incubation period can be 15 days, with an unknown level and duration of viremia, and 50 to 80 percent of infections are asymptomatic. The viral NS1 protein is secreted early in infection and usually correlates with symptom onset and maybe severity. IgM is usually detectable at three to five days after infection and declines at two weeks. IgG increases after 10 days and is lifelong. A rapid increase in IgG to higher levels within two weeks is characteristic of secondary dengue infection, and IgM may be very low or absent in secondary infection.

Persons can acquire a secondary dengue infection with a different serotype. Secondary infections place them at a greater risk for severe dengue. Severe dengue will also be covered by Dr. Tomashek when she talks about the outcome of the viral infection. Transfusion-transmitted dengue virus has been reported involving all blood components, including fresh frozen plasma, red blood cells, and platelets. There have been five reported transmissions from three donations, although we will hear about more transmissions from some of the other speakers, and we will hear more detail about these transmissions. The rate of transmissibility by transfusion may be inaccurate due to the high proportion of asymptomatic infections, the high incidence during outbreaks, the unknown duration of viremia, the lack of a licensed test, lack of recognition by clinicians, and lack of surveillance and reporting.

There are several potential screening methods. Currently there are no FDA-approved or licensed tests for diagnosis or screening for dengue virus. Dengue virus can be detected by viral isolation, antigen, and nucleic acid tests. Antibody tests may detect a past or present infection. Nucleic acid tests, both RT-PCR and TMA assays, are highly sensitive for detecting viral RNA from all four serotypes early in infection. The dengue NS1 antigen ELISA may be useful for the detection of virus early in infection to identify asymptomatic infected people. Immunoglobulin may not reflect an active infection, but may confirm a past or present infection.

We will have several presentations. Dr. Tomashek will talk about dengue virus epidemiology in the U.S. and its territories. Dr. Petersen will talk about a risk model to define the rate of infectious units during dengue virus outbreaks in endemic versus non-endemic areas. Dr. Rios will talk about dengue virus panel development. Dr. Busch will talk about an overview of data on blood donor testing and transfusion transmission of dengue virus. Dr. Margolis will talk about experience with dengue virus antigen tests. Dr. Stramer will talk about recent experience in testing blood donors in Puerto Rico and Key West, Florida.

We propose four questions to the committee:

  • Are safety interventions warranted to address the risk of transfusion transmission of dengue viruses in the U.S.?
  • Do the available scientific data support a temporary deferral for donors living in U.S. non-endemic/non-outbreak areas who have traveled to an outbreak and/or endemic area? If so, what deferral period should FDA consider?
  • Please discuss what interventions, if any, would be appropriate for donors who are residents of endemic or outbreak regions within the U.S. and its territories.
  • Please comment on the potential utility of dengue virus tests -- nucleic acid, antibody, or antigen tests -- in donor screening.

We'll go through those questions after we hear the scientific evidence from the speakers that we have.
DR. HOLLINGER: Thank you, Dr. Taylor. Let's move on, then, to the next speaker, Dr. Tomashek, who is going to speak on dengue virus epidemiology in the U.S. and its territories.

 

DR. BOWER: Mr. Chairman, are we going to have a chance for questions or just at the end?
DR. HOLLINGER: We'll have questions. Do you have questions for Dr. Taylor?
DR. BOWER: I think I just missed something. Did you show the viremic period from inoculation until virus is cleared, or is that not known? Maybe somebody else will do that.
DR. TAYLOR: Dr. Tomashek will cover that.
DR. HOLLINGER: I thought they would probably cover it later on. Thank you.

Table of Contents 

11

Agenda Item: Dengue Virus Epidemiology in the U.S. and Its Territories

DR. TOMASHEK: Good morning. Today I have been to talk about the epidemiology, but also to comment on the transmission of dengue virus, as well as the clinical presentation. I'll start with transmission.
As most of you know, dengue is primarily a mosquito-borne disease. Aedes aegypti is the most common vector, but Aedes albopictus can also transmit the virus. The transmission cycle typically starts with a viremic human being. The mosquito bites the human being. It gets a blood meal, and dengue virus is in the blood meal if the person is viremic. Then there is replication within the mosquito for about 8 to 12 days, which is the extrinsic incubation period. The mosquito then is infected and can go on and infect other human beings.

What's problematic is that the mosquito remains infected for her lifetime, which is about a month in non-experimental settings. That female mosquito can live in your house and infect other members of the household or people that are visiting your house throughout her lifetime. As you know, both mosquito vectors are found in the United States. Aedes aegypti distribution as of 2007, as reported by the states and reported by ArboNET, is essentially in some of the southern states, from Georgia, Florida, Louisiana, Texas, all the way to Arizona.

But albopictus has a little broader spread of late. These are reported from the states that have vector surveillance. As you can see, primarily it's in the southeastern United States.
But dengue -- and this is why we're here today is also a blood-borne pathogen. There is evidence of this through various means of transmission. There has been evidence through receipt of blood donor tissue and organs. There are also reports of receipt of blood products and dengue being transmitted by receipt of blood products. There have also been seven reports of transmission following nosocomial injuries. Primarily these are needle-stick injuries. Then there have been several reports in the literature and several cases in Puerto Rico this year of vertical transmission -- a mother who was viremic with dengue at the time of delivery or shortly before delivery actually passing on the virus to her fetus or her newborn infant, and the infant then developing dengue within two weeks of being born.

In terms of the clinical spectrum of the disease, I would like to start with dengue virus infection. The incidence is roughly around 5 percent. As most of you are aware, the majority of infections, like in West Nile, are asymptomatic. It ranges -- and I'll go over that next -- about 75 percent asymptomatic, 25 percent symptomatic, and of those that are symptomatic, there is a broad range, again, of symptomatic infections. Up to 5 percent of infections are severe infections. In the past these have been referred to as dengue hemorrhagic fever and dengue shock syndrome. Of those that have severe dengue, the majority actually end up surviving, so this is good. Case fatality rates can be as high as 10 percent in places that are not aware of dengue or that don't have good treatment protocols. But currently mortality is 5 percent or less in most countries.

In terms of the course of a dengue virus infection, as I said before, there is an incubation period within the mosquito and then within the human being. Within the human being, it's about 3 to 14 days after being bitten. That's why travel clinics ask you, "In the last two weeks, have you been to a dengue-endemic country?"

Viremia in humans begins about 24 to 48 hours before symptom onset. This is based on studies from the 1920s by Siller (phonetic). Then the viremia lasts for about one week, based on methods that we have for detection of the virus today. Many people, as I said before, are asymptomatic. Traditionally, we have thought that most asymptomatic people are children or those with primary infections. We know from studies in asymptomatic blood donors that viremia can be very high -- in fact, higher than most West Nile infections, on the order of 105, 109 viral copies per mL.

This question came up in discussions last year with FDA about asymptomatic infections in adults. As many of you know, most prospective cohort studies in the dengue world have been conducted among children only, not adults. In fact, there are seven of them to date. They document the symptomatic-to-asymptomatic ratio of about 1-to-1 to 1-to-13, or about 47 percent to about 98 percent of recent infections among children can be asymptomatic. There is only one prospective cohort study among adults. There are other studies that have attempted to better understand asymptomatic infection among adults which I'll present right now.

These are the studies. There is one in Malaysia, published in 2008, where they took about 2,500 pregnant women and they bled them at the time of delivery. They found that 63 of these 2,500 women, or about 2.5 percent, had evidence of recent dengue infection. They asked them, "Have you been sick in the last three months?" Pregnant women are pretty good at being able to say, "I had a fever," or, "I didn't have a fever." The overwhelming majority -- in fact, 89 percent -- said, "No, I never had a fever. I never was seen. I was healthy." These women were not children; they were 24 to 36 years of age. This is in an endemic area, Malaysia.

The next study, Porter, which is the best study out there to date, published in 2005, in Indonesia, followed prospectively for three years textile workers and bled them every three months. This study followed again about 2,500 individuals. Of those that they bled every three months, 188, or 7.4 percent, had evidence of recent infection and 76 percent of those never had any symptoms -- no fever, nothing. So that is 76 percent asymptomatic. Those were 18 to 66, adults.

Those two studies clearly show that asymptomatic infections do occur among adults. There are other studies from Puerto Rico -- we are one of the few countries that do sero studies -- in the past. The majority of those studies have been among adults. A few kids have been volunteered to be bled, but very few. A study by Waterman in 1985 showed an asymptomatic rate of 57 percent. Another, more recent study in Taiwan, where there is an outbreak going on and they would identify a positive case and then they would bleed everybody in a 50-meter radius of that asymptomatic case, also found about 36 percent asymptomatic recent infection among adults.

This is just a sampling that asymptomatic infections occur among adults. This is one of the things that we had talked about in the past. There are factors that affect this, of course, not only in children, but also in adults. It's affected by your immune status, as well as the infecting serotype. There are a number of studies that are in the published literature that suggest that dengue 2 and dengue 4 can present asymptomatically, especially in primary infections. Dengue 1 and 3 are less likely to present asymptomatically.

Now let's turn to symptomatic disease. In 2009, WHO came out with a new case definition, as I mentioned earlier. With symptomatic dengue, there is more mild disease, or what's classically called dengue fever, and more severe disease, now called severe dengue, the old dengue hemorrhagic fever, dengue shock syndrome. Basically a person can be sick for about a week, miss maybe a week of work, a week and a half, if they have extreme pain. Many of them recover without sequelae, but there is a proportion of them that actually develops the more severe form at the time of defervescence. Severe forms can be classic dengue hemorrhagic fever, with up to seven days of fever, a drop in platelet counts -- so thrombocytopenia -- hemorrhage, as well as plasma leakage, resulting in shock. Some people can be classically DHF or just have severe plasma leakage and shock, severe bleed, or this more recently recognized other organ system damage, like encephalopathy, like what you would see with other flavivirus infections, and fulminant hepatic failure and other end-organ failure.

In terms of the clinical course, I went through this a little bit -- an incubation period for up to two weeks, a short period of asymptomatic viremia, and then this febrile phase. In the febrile phase, usually about a week, this is when they present with their symptoms. They might come to you and say, "I have muscle, joint pain, bone pain. I feel lousy. It feels like I have the flu. I have this really severe eye pain and a headache." Then they might get a spinal tap and a full workup, like has happened in the U.S. of late with some of our dengue cases. They may have rash or they may not -- usually more commonly with their first infection.

During this phase, a critical thing to do is to monitor those patients with dengue so that if they go into a more severe form of the disease, they are picked up early and treated accordingly.
During this phase, you can have complications. You can have people that present and are dehydrated and need to be rehydrated. You can have febrile seizures in children, and you can have neurological disease.

The next phase, which is called the critical phase, is a phase that -- many people never go into this phase. Those that have severe dengue do go into this phase. It's identified at the time of defervescence, or when the temperature comes down. There is a rapid decline in the platelet count, with an increase in hematocrit, which is symptomatic of the plasma leakage. They develop these types of warning signs and symptoms -- severe abdominal pain, persistent vomiting. They may have a mucosal bleed. They may come in very lethargic and restless. This is the period of time that more of you are aware of this, because this is what usually makes the news. These patients are in shock, they have end-organ damage or failure, and they can have excessive hemorrhage or bleeding, requiring blood products and long ICU stays.

Then the recovery phase is the phase where many of the people, especially if you had a severe case -- you're relaxing, you're starting to recover, hopefully you're starting to think about going back to normal life activities, and usually this is just when you are reabsorbing that fluid that you have extravasated into your abdomen and your pleural cavity. The patient generally gets better, and the platelet count recovers. But there are problems. If they are mistreated in the critical phase -- for example, given too many fluids or incorrect fluids -- then they can have fluid overload and acute pulmonary edema in the recovery phase. In fact, we see people who die because of mismanagement in the critical phase.

In summary, most people infected with dengue are either asymptomatic, like in West Nile, or are mild or moderately symptomatic. But there are those people -- up to 5 percent -- that have severe dengue. They can deteriorate rather quickly at defervescence. They can rapidly go on to shock because of this increased vascular permeability. The level of plasma leakage dictates the level of management that they get, whether they are in an inpatient ward or an ICU for a few days.

Lastly, the epidemiology. Where does dengue occur? It's throughout the tropics and subtropics. As you can see here, it's where we are located and where we like to travel to. So there is ample opportunity for exposure to dengue virus, for more people that are coming from these countries to visit or be temporary residents. This figure has increased recently from about 12 million to 37 million, in 2007. We also like to travel on spring break to Mexico and other places that have a lot of dengue going on. Mexico is still our number-one travel destination, as of 2008. About 20.3 million Americans per year travel to Mexico. More recent data from GeoSentinel and from other surveillance systems show that dengue is more prevalent than malaria, from people returning to the U.S. from the Caribbean, South America, Southeast Asia, and South Central Asia.

Travel-associated dengue -- one frustration for us as epidemiologists is that it's often not recognized and it's often not reported. But as of January 2010, it is a nationally notifiable disease by CSTE. But there is still risk of local transmission, as I showed you earlier. There are mosquito vectors in the U.S., and there is a largely susceptible population here. There have been seven outbreaks in Texas since 1980. There was a nice outbreak in Hawaii in 2001. Also there is a recent outbreak among people returning from Haiti that we are investigating. There is no sustainable, effective vector control and there is no vaccine at the moment.

DR. HOLLINGER: Could you conclude, please, Dr. Tomashek?
DR. TOMASHEK: Yes, I will conclude. You can see these slides. They are in your handout. Basically, travel-associated dengue is mostly among adults. It's severe disease. Forty percent are hospitalized. There have been about 40 cases reported in the United States this year. The most have been from Florida. Most of the imported cases from Florida have been places nearby, like Mexico, and 36 of those cases were from Puerto Rico. Dengue-endemic places in the U.S. -- mainly Puerto Rico, U.S. Virgin Islands, U.S. Pacific Islands, and perhaps Florida. Florida has had 63 cases so far this year. In Puerto Rico we have had 10,500 lab-positive cases so far this year. As I said before, our cases are among all age groups. Fifty percent of them are in adults. Many of our lab-confirmed deaths are also among adults. We have had 31 lab-confirmed deaths this year, for 3.8 million people.

When does it occur? This is to answer our question about deferral and when we would do this. Dengue can happen year-round. As you can see from these outbreaks we have had most recently in Puerto Rico, all of our outbreaks are called by June. But it can happen anytime of the year. This year we had an outbreak called as early as February. It depends on what hemisphere you are in. With that, I'll stop.DR. HOLLINGER: Thank you very much. Time for one question. Does anybody have a question for Dr. Tomashek? Yes, Dr. Rentas.

DR. RENTAS: If I look at your presentation here, I think it makes me believe that if we defer people for three or four weeks, we can mitigate the risk of dengue virus. I think there were a lot of lessons learned with WNV that we can apply to this virus. Wouldn't you say that it's possible to detect the RNA of dengue maybe 100 days after every active donation, just like WNV?
DR. TOMASHEK: I don't know the answer to that question. We usually can detect dengue virus in a human sample for about seven days. In terms of NS1, we can detect the antigen for up to 10 days. In terms of deferral, you have the two-week period of incubation and then you have the infection, a month's worth of deferral, so saying, "Did you go to a dengue-endemic country in the last month?" If you were choosing to do a deferral, it would be a month deferral.

DR. NAKHASI: I think the studies are not there to exactly figure out how long it can persist. You will hear from Sue Stramer and Mike Busch today that there are studies where they used the TMA methodology. Maybe we'll hear more exact dates. At this point the studies are -- there have not been done that many studies that determine how long you can find the virus in the blood.
DR. TOMASHEK: Thank you for that. Ours are PCR tests, and we detect it for about a week and the NS1 for about 10 days. So it will be good to hear about the TMA results.
DR. HOLLINGER: Thank you, Dr. Tomashek. I think we will move on. We really have some time commitments. I'm going to ask the speakers to stay within your time so we can have some questions afterwards. If you have a critical slide that you want to show, you had better show it early. Otherwise, we are not going to be able to get to the important issues that we need to deal with later on.

The next talk is by Dr. Lyle Petersen, who will speak on "Risk Model to Define Rate of Infectious Units During Dengue Virus Outbreaks: Endemic vs. Non-Endemic Areas."

Table of Contents

24

Agenda Item: Risk Model to Define Rate of Infectious Units
During Dengue Virus Outbreaks: Endemic vs. Non-Endemic Areas

DR. PETERSEN: Good morning. This morning I would like to talk first about the arbovirus risk model and some of the theory behind it, then go on to talk about dengue transfusion-transmission risk in non-endemic areas -- what I would predict from the model -- followed by dengue transfusion-transmission risk in Puerto Rico, which, Dr. Tomashek has mentioned, is an endemic area of the United States, some limitations of the model, and conclusions.

In its simplest form, the risk of an arbovirus being transmitted in the blood supply can simply be the prevalence of donors who are viremic, the transmission rate to a recipient from a viremic donor, times the proportion of recipients who are susceptible to infection. When we look at dengue and other arboviruses, we can make several assumptions about these three factors. First, looking at the prevalence of donors who are viremic, our model assumes that symptomatic persons don't donate, since most flaviviruses will cause a febrile illness, which theoretically should prevent people from donating. The other assumption is that donors are similar to the population at large with respect to mosquito exposure and infection risk. We assume that mosquitoes are equally likely to bite blood donors as anybody else.

The second thing is the transmission rate to susceptible recipients. We know that with the mosquito-borne flaviviruses, the mosquito injects a very small amount of virus, and nearly everybody gets infected. So what we are assuming is that with a blood transfusion, where you get a much larger amount of virus, the transmission rate is about 100 percent.

The final factor is the proportion of recipients who are susceptible to infection. Unfortunately, for dengue in most situations, it's really impossible to determine how many people are susceptible or not susceptible, particularly in an endemic area. So for the purposes of the discussion this morning, I'm assuming 100 percent, which we know is an overestimate in many areas.

Our general arbovirus risk model: The risk derives from viremic persons who remain asymptomatic and donate or who donate before becoming symptomatic. The transfusion-transmission risk is simply the prevalence of viremic and asymptomatic persons in the population at large. That's assuming that blood donors are like everybody else. We know from basic epidemiology that prevalence is equal to the incidence times the duration of viremia.

This is the arbovirus risk model in its simplest form. Simply, what we have is that the risk is equal to the incidence of infection in the population at large times the proportion of infections that are asymptomatic -- that's the incidence of asymptomatic infection times the duration of viremia of asymptomatic people before they get symptoms -- plus the incidence of infection in the population at large times the proportion of people who are symptomatic, times the total duration of viremia.

If you look at the variables in this risk model, there are a couple that are highly variable and uncertain. One is the incidence of infection. We know in various populations that from year to year the incidence in the population can range from less than 1 percent -- this is in endemic areas -- to more than 30 percent per year. The other variable that is highly variable and uncertain, as Dr. Tomashek touched on, is the proportion of infections that are symptomatic, which varies by the infecting strain and the previous exposure to heterologous dengue viruses. As Dr. Tomashek mentioned, most of those studies indicate that about 25 to 50 percent of the people become symptomatic.

The less variable and more certainly variables are the duration of viremia, which is approximately five days after the onset of symptoms, and the -- we don't know the exact duration of viremia before symptoms, but we know it's very short and thought to be about one day or less. We don't know what this duration of viremia is for asymptomatics, but presumably it's similar to that of symptomatic persons. One thing that is worth noting is that the duration of viremia is likely to be less than the duration of NAT positivity, which we showed with West Nile virus.

Let's take a non-endemic area. As many of you know, there has been an ongoing outbreak in Key West, Florida, now going on for almost two years. This is the first dengue outbreak in Florida since 1934. Key West has a population of 23,000. The first case onset was on July 19, 2009, and there were 27 infections identified through October 19. A household-based serosurvey was conducted in September in the Old Town area of Key West, and that time it was determined that 4.9 percent were determined to have recent infection.

So applying the transfusion risk model, we are assuming that:

  • 4.9 percent of the population became infected in 70 days, or the incidence is about 7 per 10,000 per day.
  • 33 percent of the infections were symptomatic.
  • One-day viremia before the symptoms developed.
  • A six-day total period of viremia.

Plugging all those numbers into that model, you get an estimated average transfusion risk during that outbreak period of about 18.7 per 10,000.

 

I would like to add that this risk may be overestimated, because the 4.9 percent infection cumulative incidence in 70 days may not apply to all of Key West. The survey was only done in the Old Town area. And the outbreak duration may be longer than 70 days, because we only knew when we could identify the first case. But the outbreak could have been going on longer.

This compares to an estimated transfusion risk of 0.5 per 10,000 during the 2004 outbreak in Cairns, Australia. They used our model, actually, and they estimated a .19 percent cumulative infection incidence in 196 days during that outbreak. So you can see that these outbreaks in non-endemic areas have a pretty wide range of estimated transfusion-transmission risk. Now I'm going to move on to Puerto Rico, where we have much more data. The model is a bit more complicated. We use a statistical resampling approach, which allows for a couple of things. One is the temporal estimates of risk, rather than just a point estimate. We estimate the viremia prevalences in the population at large on any given day, based on observed illnesses and some assumptions made from that. We can also account for the uncertainty of model parameters. What we do is, we assume a random variation in the model parameters, within prescribed limits, to quantify the uncertainty. We run the model a bunch of times, and we can calculate 95 percent confidence intervals around the risk estimates.

Our assumptions in the model:

  • We're assuming that 33 percent of the infections are symptomatic, with a range of 25 to 50.
  • The viremia ranges from five to seven days, with an average of six.
  • The viremia before symptoms develop is about one day, with a range of .5 to 1.5.
  • And in Puerto Rico, we're assuming an underreporting of clinical cases of about 15 times, with a range of 10 to 20. So based on clinically reported cases, we're assuming that the incidence is about 15 times higher

In the model, what we get is an average risk. This was done from 1995 through 2010, through October 28. The average risk was 6.9 per 10,000 donations over a 15-year period, with a 95 percent confidence limit between about 4.5 and 9. The maximum risk during this period of time was this year, which was 51.1, with confidence limits of 43 to 60.

 

This is what the risk looks like over a 15-year period from 1985 through 2010. You can see that at no time was the risk zero. There was always a risk. There was marked seasonality, as Dr. Tomashek mentioned, but occasional massive outbreaks, with a very massive outbreak this year. This downward part here is an artifact of the model. So the risk varies greatly, with some massive increases in risk during outbreak periods.

To test the validity of the model, I compared a time period, from July to December 2007, when Sue Stramer at Red Cross tested donors using the Gen-Probe NAT test. At that time they found 29 TMA-positives out of 15,000 donors. I think Dr. Stramer will talk more about this later. During that period, our model predicted that there would be 29 positives out of 15,000 donors. So our model actually fits quite well with the empiric data.
What are the limitations of this?
First, the incidence of infection in the population at large is very difficult to determine. So we are making inferences of the true infection incidence from disease surveillance data. But there are some uncertainties about the underreporting of clinical cases, and the unapparent-to-apparent infection ratio is uncertain and probably changes over time as different viruses come in. The transfusion risk model also only considered risk of viremic donations. As I mentioned before, the true transmission rate is less because of background immunity in the recipients, which is not measurable.

So what can we conclude? There is a variation in dengue transfusion risk which mostly depends on infection incidence and the proportion of infections that are symptomatic. Both are difficult to measure and highly variable. The transfusion risk model seems plausible because the predicted risk and observed NAT yield were basically the same in 2007 in Puerto Rico, where we had empiric data. The estimated transfusion risk in Key West donors during the 2009 outbreak -- we estimated about 19 per 10,000. There were some model uncertainties.

In Puerto Rico, over a 15-year period, the average daily risk was about 7 per 10,000 donations and was highly seasonal and variable by year. The maximum daily risk was 51. That was actually this year. What's noteworthy is that the average predicted daily transfusion dengue risk over a 15-year period was similar to that which we predicted for West Nile virus in six high-incidence states during the peak of 2002 epidemics. Our estimates then were about 2.1 to 4.7 per 10,000, which have subsequently been borne out when we started NAT screening. Our model proved to be quite accurate. So this model actually led to West Nile virus screening in the U.S., and, actually, the dengue risk in Puerto Rico on any given day is going to be higher than what it was when West Nile virus epidemics hit in 2002. That concludes my talk. I can take questions.
DR. HOLLINGER: Ken?
DR. NELSON: Lyle, was your data age-adjusted or did you specifically look at the risks of people, let's say, over the age of 18 versus younger?
DR. PETERSEN: We only took people age 16 and older, the blood donation age. Sorry, I did not include that, I don't think.

DR. KRAMER: Lyle, two things. One is, in Florida, you said that you may have overestimated the risk, but you may have underestimated it if you just look at Old Town. Weren't all the cases in Old Town? So if you forget about the rest of Key West and you have about 6,000 people living in Old Town, then that's going to increase your risk.

DR. PETERSEN: Right, except I took the population denominator for all of Key West. If I had a population denominator for Old Town, I could have figured out the risk for Old Town Key West itself, which would have been higher. You're right about that.

DR. KRAMER: Then the other thing is, you made a statement -- and people always make this statement -- that mosquitoes inoculate just a little bit of virus. In studies that we have done with West Nile, they can inoculate up to 106. The median dose is actually 105. So they have a large range. It may be, when they only inoculate 10 or 1 PFU, those are the people who don't get infected or are asymptomatic. So I think you have to be careful when you say mosquitoes only inoculate a low dose.
DR. PETERSEN: Right.
DR. KRAMER: If dengue is like West Nile, which we don't know.
DR. PETERSEN: Right. Good point. One of the things we found out when we did the transfusion-transmission studies for West Nile was that the transmission rate appeared to be very close to 100 percent. Given the fact that dengue has a much higher viremia, I would assume it would be even higher.

PARTICIPANT: Can you please comment on why severity is not captured in the risk model, but prevalence is? Because in most cases, when you define risk, you look at both severity and prevalence.
DR. PETERSEN: Right. I did not attempt to look at what happens after you get an infected blood transfusion, because we simply don't know. The model only looks at the probability of transmission, rather than any kind of outcome.
DR. HOLLINGER: Okay, let's move on. Thank you, Dr. Petersen. If there are speakers that come up from the audience and are recognized, use the microphone there in the middle of the aisle and identify yourself to the audience, please. The next talk is Dr. Rios, who will talk to us about the "Update on Dengue Virus Panel Development."

Table of Contents

35

Agenda Item: Update on Dengue Virus Panel Development

DR. RIOS: Thank you. I think it's needless to go over the dengue logistics, but there are four known serotypes. Why is reference reagent needed? Because there is no current way of evaluating the performance of assays, either for diagnostic or blood screening, if one gets to play, and the reference reagents usually require large volumes from all the four serotypes, which clinical specimens are not suitable for due to the limited volume. So to have a reagent material where all the tests can be analyzed to define analytical specificities will not be suitable to use in clinical samples.

Fortunately there are laboratory strains from all serotypes that are available and can be used to develop a standard. In the blood safety section, we need to detect not only just any dengue infection, but active infection that is, there is virus being produced, released to the bloodstream, and can potentially be transmitted by transfusion. There are essentially two types of assays besides the viral culture that can detect viral production, and actually in quite early infection. One of them is the nucleic acid test. It's for direct detection of viral RNA, which is produced soon after the initiation of viral infection, while the virus is replicating and pumped up in the bloodstream, way before the production of antibody. Another way is to detect viral antigen. There are proteins made during the viral replication that are ultimately released to the bloodstream. We have examples of other viruses, such as p24 antigen for HIV. But for dengue the NS1 antigen is known to be produced rather earlier. It's a nonstructural protein, as Debbie mentioned in her talk. It's secreted to the bloodstream and can be detected early in detection. It appears a little bit around the time that RNA is also detectable and way before antibody against viral infection is produced.

Historically, NAT has been used to assure blood safety for viral RNA, such as HIV, HCV, West Nile. The detection of the genetic material is very early and most sensitive when NAT assays are used to prevent transmission by transfusion. Early epidemiological studies of the presence of dengue among blood donors were performed -- and Mike Busch probably will refer to this -- using NAT assays. So FDA proactively engaged in the production of an RNA standard reagent for dengue to facilitate development of assays and evaluation of performance of both existing assays and newly developed assays, and for future lot releases, in case of need. FDA understands that the antigen assay is also a potential for the screening of blood safety. A dengue antigen panel will be developed if needed. The isolates we have are capable of allowing that.

In an effort to develop the standard, CDC has collaborated with FDA, providing four prototypes of these four serotypes, dengue 1 from Hawaii, dengue 2 from New Guinea, and dengue 3 and 4 from the Philippines. I want to remind you that these are laboratory strains that have undergone multiple propagations after the viral isolation from the original human specimen. There is a problem with activation(?) of dengue RNA titers -- I'm referring to RNA because that's our goal in general, both in the diagnostic setting and in the scientific setting. The viral titer has been defined as a plaque assay, meaning an assay that looks at how many foci are formed when an inoculum is plated in the confluent viral cell culture, and then the number of plaques are counted. The pitfall of this technology is that the number of particles involved in the formation of one plaque can vary greatly. It can be estimated to go between 1 and 1,000, if it's in the same location that foci will be generated. So we need to have a correlation between RNA copy number and PFU. To remind you, non-infectious virus or defective virus may also be detected by NAT, so if we have a titration in PFU and then we run NAT, we could err to estimate that the NAT assay is wrong, because the infectivity assays do not correlate with the NAT titer. So copy-number assignment is necessary to define analytical sensitivity of NAT assays.

In order to achieve that, we have produced, using the four serotypes, 240 mL of viral stocks from each of the four serotypes. We grew those in mosquito cells. We did compare to Vero cells and found that mosquito cells were a better producer, so we engaged in using the mosquito cells. We did some preliminary characterization of these stocks in-house by genetic sequencing, infectivity titer by plaque assay, and also RNA concentration by determining what we call in the lab PCR-detectable units, by performing limiting dilution prior to running a TaqMan assay or quantitative assay, amplification assay.

A portion of these stocks were subjected to heat inactivation at 62 degrees for an hour. This is highly desirable because it is in shipment and handling of stock materials. We performed dilution of both live and heat-inactivated stocks and froze at -80 for further characterization. We had a preliminary round of characterization. Two concentrations of these stocks each of the live and heat-inactivated stocks were tested at FDA, NIH, in two different laboratories, at CDC, in collaboration with Robert Lanciorti (phonetic). In New York State, Dr. Kramer participated in gene probe. The heat inactivation was also independently confirmed by cultivation in four independent laboratories, and we are confident of the inactivation of the material.

This shows the preliminary data of concentration. For instance, for dengue 1, we had 1011 .7 PCR-detectable units per mL and a titer of 105 PFU per mL. Right now we are performing stability studies in the pre-characterized stocks. To do that, the evaluation of these candidates -- preparation has been diluting two different milieus. One is the tissue culture medium which was produced and was the initial test that we performed, and also diluted in base matrix, which is defibrinated, delipidated human plasma that tests negative for all agents that are commonly tested in blood donations. The reason to do that is so that the material is diluted in the same material that we will be testing, to mimic the specimen.

These were dilutions stored at virus temperature, at 4 degrees, -20, and -80. Also -20 and -80 aliquots were set aside to perform freezing and thawing instabilities of the material. These were tested at various time points. I'm going to show you up to 30 days. In the left-hand side you can see the dengue 1 and 2 preparation in the tissue culture medium and in the base matrix. We see a slight decay when it's diluted in base matrix, but essentially it's pretty tight and stable so far. The same is observed for dengue 3 and 4. We are pursuing with 90 days in subsequent tests, in various time points of freezing and thawing. The next steps and future plans: A second round of testing will be performed using stocks, only base matrix this time. We are selecting laboratories and inviting them to participate. If you have a test and you wish to participate, send us an email.

After that result is generated, we will perform data analysis of these NAT results in the second round to further refine the PCR-detectable units. We are engaged in contacting WHO Collaborating Centers to proceed with the international evaluation, to generate an international standard. After these materials are assessed internationally, the data will be analyzed, finally assigned a copy number at the international unit, and panels will be available when adequately standardized and needed. Thank you for your attention.
DR. HOLLINGER: Thank you, Dr. Rios. I have a question for you, Maria. A comment first and then a question.

The comment is on the standardization. There was a fairly large differential between the viral load and the titer, the PFU, like 1 million-to-1. That seems very high, considering that you thought there might be one to 1,000, maybe, inactive particles per active particle.
DR. RIOS: I was expecting that to come. This is depending on the MOI that we selected to infect. Our aim at this stage was not to get the most infectious material, but the highest viral load. We are planning on refining that and doing it low-MOI, meaning multiplicity of infection. Every time you infect with low viral titer, with low concentration of PFU, you generate more intact particles. However, when we infect with the high MOI, we tend to produce more virus, which is desirable in a standard situation, because we would have the same material to be tested for that case.
DR. HOLLINGER: The other question is, have you had a chance to look at testing for nucleic acid in EDTA, heparin, and serum, in terms of detectability, when those were used to collect samples in blood?
DR. RIOS: We have had in the lab the opportunity to test in EDTA. We did not compare to heparin right now, because we don't have the material to do that. We don't have access to infected individuals to be collected in heparin. It's probably a good idea to plan for the next year.

In serum, I would expect that, like any other virus that has been reported, when the clot is formed, that binds to the cellular compartment, including platelets, et cetera, in the fibrin formation -- pull down some of the viral concentration. Yes, in the past, the gold standard for diagnostic has been isolation of virus from serum. However, anybody that works in a diagnostic laboratory knows that it's a limitation. In not everybody that is infected can you get an isolate. My speculation is that that could be due to the usage of serum instead of plasma for cultivation.
DR. HOLLINGER: Yes, Ken, last question.
DR. NELSON: You mentioned you can get falsely elevated NAT testing because of defective particles that aren't really infectious. Do you have any idea how big a problem this is?
DR. RIOS: I wouldn't call it a falsely positive NAT test, because if you have an RNA that is enough for you to detect, you cannot guarantee that -- well, the individual is infected and produces the RNA anyway. You don't want to take a risk, for blood safety purposes. I don't think that a positive NAT test, unless proven to be false positive by follow-up, can be taken lightly and considered false positive, unless the individual does not seroconvert.

We and others have proven for West Nile and HIV and other viruses -- and now we are doing dengue -- that the virus can also bind to red cell compartments and some other compartments, and actually diminishes a little bit what is free in plasma. So it's a tough call to make.
DR. HOLLINGER: Thank you, Dr. Rios. The next talk is by Dr. Michael Busch, "Overview of Data on Blood Donor Testing and Transfusion Transmission of Dengue Virus." By the way, we are still going to have an opportunity to ask questions of these speakers later on. I don't want you to think that we won't have time to do this later on. Just write down your questions that you want to ask.

Table of Contents

44

Agenda Item: Overview of Data on Blood Donor
Testing and Transfusion Transmission of Dengue Virus

DR. BUSCH: Thank you, Blaine. What I'm going to do is quickly run through the early studies of dengue viremia and then talk about a pilot and a planned future study on transfusion transmission and disease penetrance. The early studies started, actually, right on the heels of the West Nile epidemic, when we moved quickly to develop, in collaboration with both Gen-Probe and Roche, nucleic acid testing that was implemented in 2003-2004. At that point, we realized that dengue might be a problem, so Jeff Linnen and the folks at Gen-Probe were good enough to begin to develop a dengue TMA assay.

This just summarizes it. It's essentially based on the West Nile prototype assay. It runs on the TIGRIS platform, so it has high throughput. This is just probit analysis to demonstrate the analytical sensitivity against the four different genotypes 1 through 4. You can see detection really down to three and one copy, with high probability. Based on probit analysis, the 95 percent detection limit is down at around 15 copies. So the 50 percent is down at around four or five copies. So this is comparably sensitive to the HIV, Hep C, West Nile assays, and detects all four serotypes with equal analytic sensitivity.

In studies that we executed in Honduras, Brazil, and Australia, and Sue Stramer and CDC in Puerto Rico, we documented moderate rates of viremia. The highest we observed was in Honduras, at almost .4 percent. In Sue's studies in Puerto Rico -- the earliest study that was published, Mohammed et al., was .07, but subsequent years detected as high as .25 percent in Puerto Rico. In all of these settings, we could detect all four serotypes and often detect the virus in the presence of antibodies, indicating a secondary infection with heterologous virus. There was virus cultured from these samples. Again, these data were published.

These are just busy tables out of the two companion papers, both essentially documenting again all four serotypes, viral loads, as alluded to, ranging from 100 to 108. In many of these cases, we could type the virus both by genotyping, by PCR methods, and by isolation. You could then determine the presence or specificity of antibodies in these viremic donations and sort the cases as to whether they were primary infections, first-time infected donors who had virus of one genotype and no antibody, versus secondary infections. In all cases, both in Puerto Rico and in Honduras and Brazil cases, the secondary viremias were always of a different type than then preexisting antibodies. So, as would be predicted, these people were getting reinfected by a different serotype.
You heard earlier -- and you will hear more from Sue Stramer -- about the fact that there have really only been three viremic donors linked to five recipients. So there is this disconnect between these remarkable rates of viremia in donors, as Lyle just modeled, and the absence of clinically evident transfusion-transmission cases. There are several major explanations that we considered.

One is simply a lack of active surveillance. You are not actually monitoring recipients for infection. In these endemic areas, there are often no windows on hospitals, et cetera, so you could actually get infected by mosquitoes at a much higher probability than transfusions. So there is really not active surveillance.
Second, you may have preexisting or co-transfused antibodies. In many of these settings, 80, 90 percent of the population is antibody-positive. Those antibodies could modulate the infection risk or the disease penetrance.
Also transfusion recipients are generally immunosuppressed, and the manifestations of dengue typically are hyperimmune reactions, particularly with secondary dengue. So it's possible immunosuppressed transfusion recipients have lower risk of disease, despite acquisition of infection.

We began trying to look at this two or three years ago in Honduras, which in our studies had the highest rates of viremia. We launched a study, in collaboration with Liz Vanelli (phonetic) and with Gen-Probe, to collect samples -- this was really a pilot study, non-funded -- the goal was to collect 11,000 donor samples during the epidemic in 2007, barcode those samples, and link them to donor age and gender and donation period, but delink from the individual donor identifiers. Likewise, on the recipient side, we enrolled recipients and got samples a week post-transfusion, as well as attempted to retrieve the pre-transfusion samples, coded all these samples, linked the donor samples to the recipient samples, but then delinked them from the individual. This was IRB-approved, but it ended up being an anonymized, if you will, donor/recipient repository.

After enrollment -- again, I just alluded to this ¬- we created this anonymized repository. Then the samples were sent out. The goal was that we would collect about 4,000 to 5,000 donor samples that were linked to enrolled recipients. Those would be screened at Gen-Probe by TMA, and if viremic, the donor and recipient samples would be characterized, as we had done in those earlier studies, with respect to viral load, antibody status, et cetera. If the donors were positive, then we would go to the post-transfusion samples from those corresponding recipients and test them, and if positive, test the pre-transfusion samples and try to prove transmission cases through genetic linkage.

During this period, we actually did collect about 9,000 donation samples. During the period, however, the recipient enrollment was much lower than we had hoped. There were only 229 recipients who enrolled and for whom we had the post-transfusion samples collected 3 to 10 days post-transfusion. We did follow these recipients up two to three weeks later to get symptom-development information on the cases. This just summarizes the 8,800. These are two different regions in Honduras that were involved. Again, we were unsuccessful at accruing the recipients as we had hoped. We ended up with a total of 383 donations that went into 229 recipients. When these were tested at Gen-Probe -- we went ahead and tested all the samples from the donors and the recipients; there were 570 samples -- we found five reactive samples, four from patients and one from a donor. But none of these linked to each other. The viremic donation did not correspond to these recipients.

We went ahead, though, and further tested. For each of the viremic recipients, we went back to the additional donation samples that had originally tested negative and retested those in duplicate. None were reactive. For the donor who was viremic, we tested the recipient again. None was reactive. So this was essentially a negative study, underpowered. The design, we felt, was reasonable. For that reason, we now have a pending grant to launch a much larger and adequately funded and powered study in Brazil. Brazil, within Latin America -- of course, it's a huge country -- has by far the largest number of dengue cases and the highest incidence. This is just a comparison across South American and Latin American countries. You can see over time that the Brazilian epidemic has become huge. Geographically, it's now throughout the country. You have all four serotypes now. Over time they have progressively seen dengue 2, then 3, then 1 entering the country. Most recently, this past year, dengue 4 has entered the Amazon region. But the highest-incidence regions now have three dengue serotypes hyperendemically transmitting annually.

There is a program that NHLBI has funded called REDS within Brazil. It involves these three regions, particularly the area of Rio de Janiero, Sao Paulo, Belo Horizonte, which is in this region, and Recife, the major public blood centers that are part of a network. Within this network, the REDS II Brazil network, we collect about 400,000 donations at these centers annually. All of that data is compiled on a routine basis into a within-country data warehouse, within-country repository, and then it's all coordinated through a U.S. coordinating center. So this existing platform of collaborators and blood centers -- by the way, hemocenters in Brazil not only collect the blood, they transfuse all the blood. They actually are in the hospitals and managing the transfusion services and the patients who get the blood. This allows us to do these kinds of transmission studies more efficiently.

Our aim was to establish the background prevalence of dengue in donors and recipients in these four hemocenters in Brazil and then, within one of the regions, to launch a transmission and disease penetrance study, which is really the same approach we took in Honduras. We also designed the study to consent the donors and recipients to allow retention of the repository for future studies of, particularly, arbovirus transmission. In terms of Aim A, we plan to collect 1,000 representative donor samples from each of the four hemocenters, probably during the end of 2011, and perform serologic testing, to just get background seroprevalence in the regions. In addition, in one region, post-epidemic, we would do a similar serosurvey to assess the effect of a new epidemic on seroincidence. In addition, on the recipient side, we are planning to establish a sickle cell disease cohort in these regions. In the regions with the highest dengue prevalence in the donor pool, the plan is to have 1,000 sickle cell disease patients who have received at least 20 units -- so highly transfused sickle cell patients -- versus age-matched sickle cell patients who were never transfused, and compare the prevalence of dengue serologic markers in highly transfused versus non-transfused sickle cell patients to evaluate, in a generic sense, the risk of transfusion transmission in these highly endemic regions: How much does transfusion contribute to background incidence attributable to community exposure?

Aim B is the most important piece, I think, which is the transmission rate and disease penetrance. This is obviously a complex study. In all regions, we will obtain IRB approval to execute a transmission study, such that, depending on the background penetrance and, as you will see, the activity of dengue as the season evolves, we would launch the study in one of these regions. This requires IRB approvals, Ministry of Health approvals. All this has to be in place before the epidemic hits. Fortunately, Brazil has launched a really effective and systematic surveillance of dengue. They have a program that predicts the upcoming year outbreak, based on incidence in the prior year, the extent of infective mosquito population density, the water and garbage infrastructure, and the serotype. They have modeled out what the predicted epidemic probability is in the upcoming year using a model based on these parameters. In addition, they track in real time, just as you saw from CDC in Puerto Rico, dengue reports and can predict quite early in the year whether they are going to have an epidemic.

So we will be tracking, in collaboration with the public health, this activity and then will launch an actual transmission study in one of these regions. The goal would be to actually launch the study in early January. During a five-month period, we would consent all donors with supplemental consent and collect an extra tube of blood in a tube that can be quickly spun and frozen for effective storage and testing. Then the recipients would be enrolled -- again, very similar to the design of the Honduras pilot study. We would acquire pre-transfusion samples and consent recipients. This would be a linked donor-recipient study, where we would retain linkage; we would not anonymize the samples. Then the recipients would be enrolled, questioned about previous dengue history; the pre-transfusion samples stored, with consent; and then discharge samples or samples collected 5 to 10 days post-transfusion. So the strategy, as before, would be to look for the post-transfusion viremia, and if the recipient was viremic and linked to an infected donor, we could go back to the pre-transfusion sample and rule out preexisting infection. We would also follow the recipients at a month post-transfusion to get the clinical outcome data. These samples -- we budgeted to test 15,000 donation and 5,000 recipient post-transfusion samples using TMA and then would perform a type-specific PCR to quantify and genotype the virus, serology to assess whether the recipients and the co-transfused units had antibodies, and then genetic studies to determine transfusion linkage.

Finally, in terms of disease outcome, we would have clinical outcome data. Questionnaires would have been administered both at the time of accrual in terms of prior history of dengue. But most important, on the post-transfusion time points, particularly in the range of a month out, we would get symptom status and we would have recipients who acquired dengue and recipients who did not, so we would be able to compare the rates of symptoms in the cases -- the recipients of viremic units who were themselves viremic -- with controls who were non-infected. We would also have mortality data -- there is actually a very good death registry in Brazil -- and also hospital records data to assess mortality attributable to dengue. Thank you.
DR. HOLLINGER: Thank you, Dr. Busch.
Questions for Dr. Busch?
DR. NELSON: What is the age distribution of dengue in Brazil? It's a fairly recent epidemic. I wonder, is it primarily children or is it adults as well?
DR. BUSCH: I don't know specifically. Certainly my understanding is, primarily children. Similar to what you saw in Puerto Rico, the proportion of individuals who have antibodies as evidence of recent infection, in places like Rio and other areas with problems with respect to garbage, et cetera, is remarkably high -- 80, 90 percent seropositivity by the age of 20 to 30. So you have a lot of preexisting antibody that could be heterotypic and not totally protective.

MS. BAKER: Thank you. Under your slide where you talk about Aim A and the prevalence in donors versus recipients, in your recipients, you are going to look at 1,000 age-matched sickle cell patients who were never transfused. Are you taking into account that they may already be infused by mosquitoes?
DR. BUSCH: That's sort of the question: Can we detect a transfusion-associated increased seroprevalence relative to the background exposure rates? You could certainly argue that the patients who are sicker, who need more blood -- i.e., the transfused group -- would be less likely to have community-acquired exposures, given that they are ill and not, perhaps, out in the community to be infected as highly, although again in Brazil, and I think probably particularly in the sickle cell population of individuals, they are all exposed. There is basically so little protection from mosquito exposure in the general population. But it's a good point. There could be some confounding between the two groups.
DR. KEY: I think you mentioned that it's possible that immunosuppressed individuals are less symptomatic. Do you have data for that or is that speculative?
DR. BUSCH: I would be interested in comments of the folks from CDC, but the only evidence is that people who are -- the manifestation of dengue secondary infection is kind of a hyperimmune reaction to the infection. But I'm not aware of specific data, either in humans or animal models where you could purposely immunosuppress and assess the relative rates of disease.
DR. HOLLINGER: I think we will move on to the next speaker. Thank you, Dr. Busch. The next speaker is Dr. Hal Margolis, who will talk to us on "Experience with Dengue Virus Antigen Tests."

Table of Contents

56

Agenda Item: Experience with Dengue Virus Antigen Tests

DR. MARGOLIS: Thank you. You already heard a bit about dengue diagnostics and NS1. I just put this up here to remind you, in terms of virus detection, dengue has gone through a lot of evolution in the last five years in terms of diagnostic testing. Prior to that, most activities, in fact, were done with virus isolation, mosquito inoculation, and only recently has there been a really wide use of nucleic acid amplification and detection and, more recently, antigen detection. Antibodies I will talk about a little bit. You have heard quite a bit about antibodies. But I think in our context, antibody testing has really not been potentially useful.

So let me tell you a little bit about dengue diagnostic challenges. Again, most everything that is around on dengue really has to do with clinical diagnostics for dengue disease. If you look at a primary infection -- and you have heard quite a bit about viremia -- we really know nothing about that pre-febrile-illness duration of viremia. There really aren't studies, and animal models, in terms of rhesus models, don't give you a real disease or an infection period that looks like the human disease. But I think one to two days before and a period of five days, maybe six, seven -- that longer tail out for detectability is not known. NS1 antigen appears to parallel viremia. There are some data now that have come out of Vietnam and some other studies, again in people with acute illness, that suggest that the antigen may persist for several days past the duration of viremia.

IgM has always been a dilemma for dengue, in that detectability is only about four to five days after the onset of fever. There really is an IgM negative window, which again, for disease detection, has been difficult. You see the same things with secondary infection. It appears that viremias are about at the same level as what you use in a primary infection. The thing that is quite different is that the IgM titers are quite a bit lower, and upwards of 20 percent of individuals with secondary infections do not have detectable IgM. I put up here that IgG antibody is actually really not useful in acute disease diagnostics. Part of this is because of the incredible cross-reactivity of dengue with itself -- so other dengues in any of the assays as well as with other flaviviruses. Yellow fever, Japanese encephalitis, and West Nile to a lesser degree all cross-react in most of these ELISAs for IgG.

A little bit about NS1 antigen. It appears to be a highly conserved soluble glycoprotein. It was pointed out by Deborah that, in contrast to the flavis in the family -- HCV that you are all used to -- this flavivirus actually has an NS1 antigen. It is secreted during virus replication. It's present, as I pointed out, in both primary and secondary infections. It is type-specific. There have been large panels of monoclonal antibodies that have been raised that can actually differentiate the dengue virus types using NS1. But there is also antibody to NS1 that is produced during the infection. The kinetics of this aren't totally worked out. Paul Young, the one who has done the most work on this, from Queensland, and Marie Flamand at Institut Pasteur previously have shown that this actually binds to NS1 and also becomes part of a problem in diagnostics in insensitivity.

A couple of other characteristics, which again I kind of pointed out in that cartoon. There is also some data that would suggest that NS1 antigen, or quantitative NS1 antigen, may be associated with disease severity. It has now become useful as a single specimen diagnostic tool. There have been a number of comparative studies that have looked at the two commercially available NS1 tests, one made by BioRad and the other by Panbio. Here are the numbers. The BioRad test has been looked at more extensively. You see very wide ranges in sensitivity, as low as 63 percent and as high as the upper 90s. The Panbio test has had sensitivities in the 60 percent range.

Recently -- actually, starting back in 2005 -- with the WHO and the Tropical Disease Research Network and the Pediatric Dengue Vaccine Initiative, there were a large number of undertakings that were funded to begin to look at these diagnostic tests, with a global network of laboratories set up, with reference labs -- one of those being the CDC lab in San Juan, Puerto Rico, the other one at Mahidol University in Bangkok -- a steering group, biobanking, GCLP. The objective of this was as follows: It was to assess the performance of these various diagnostic tests -- and, again, I'm giving you data that is in the diagnostic realm; it has really never been looked at in terms of blood screening -- to look at IgM anti-dengue, as well as NS1 antigen and to provide information to worldwide users of these tests -- there have been a lot of tests on the market -- and ultimately for WHO to help them put these tests into procurement.

What was the process? Companies, as identified through Web searches and other information, have been identified. Most of them have agreed to participate in these studies. They sign an agreement -- this is going to tell you one piece of what I can't tell you. I can't tell you which of the companies, when I get to the final data, because everybody hasn't seen their data back, which was part of the agreement. But I think the data is going to be useful nonetheless. Then these panels were sent to the sites from the reference labs.

The first assessment was for IgM anti-dengue tests. This has been published. It showed that of a multitude of tests out there, only three plate ELISAs were useful. In the world of dengue, rapid tests are often -- people have interest in those. It turned out, at least with IgM, that none of the rapid tests had acceptable sensitivities or specificities. Now let's get to the NS1 antigen evaluation, which is actually going on -- essentially, has finished. The rest of the data, since Liz Hunsperger from CDC, and from our lab, is the one who led this effort -- she is actually in Vietnam right now looking at the other half of this data.

There were two panels put together, the combined and, in fact, this will ultimately be combined in terms of the information. But there was a panel derived from the Americas and one derived from Asia. We had problems, because these are virus-positive, and all the issues of shipping just kind of said, "We're going to do the Americas panel in the Americas and we're going to do the Asian panel in Asia." These are all virus-positive individuals, acute dengue, days 0 to 7, and for reach of the four virus types. They are not equally allocated because of the problems of getting these kinds of specimens. The same was done for the Asian panel. Then there is a very large challenge panel that has both flavivirus-positive individuals and flavivirus-negative individuals, as well as people who are known to be infected with other flavis, depending on the region, and then other febrile illnesses that are confused with dengue. Then there was a group of challenge specimens from other systemic diseases, such as rheumatoid factor and lupus.

The tests that were looked at -- we are only going to focus on these NS1 microplate ELISAs from Panbio and BioRad and Standard Diagnostics, which is a company in Seoul, Korea, which has now been bought by Alera (phonetic), which now also owns Panbio. So you have to deal with those issues. Here's the data, based on the Americas panel. There are the three companies, A, B, and C, not necessarily the way I had them on that first slide. You see this wide range of sensitivity in the ability of these tests, as used according to the package insert, to be able to detect a positive individual. Again, these are people who have dengue during the viremic phase of the disease. All these are virus-positive by RT-PCR. There's the specificity data, positive predictive and negative predictive values.

Now I'm going to take you back and show you that there has been some change. If you went back to one of those studies that was on my original list, you saw that company A at that time, with about 200 specimens in an evaluation panel, had a sensitivity of about 83 percent, and company B -- and the two companies at that time were Panbio and BioRad; that data is already out there -- if you look at these as A and B now, there has been a flip-flop. Is this also seen in the Asian data, in the Asian panel? We don't have that data yet. That data is actually being put together with the Americas panel almost as we speak. It actually will go on tomorrow in Ho Chi Minh City. So we don't have all the results. But what I'm trying to point out is that these tests, which are used for dengue diagnostics, for acute disease, have shown a lot of variability and seem to be showing variability now, over time, by the same manufacturers.

There were some other interesting things that came out of the Americas panel in terms of looking at the data. Consistently for all three of these, they had low sensitivity to dengue virus 2. Again, depending on what is circulating in your region, you may have predominance of that virus at times, as we have seen in Puerto Rico. There were also false positives to West Nile. There were false positives to yellow fever. There were false positives to rheumatoid factor. Now, different by different companies, but the point is that these are some of the things that these kinds of challenge panels will point out when you are looking at a diagnostic test.

What are our conclusions? I still think NS1 is probably ultimately going to be a reasonable dengue diagnostic for at least acute disease, once we have some consistency and we know which of the various tests that are available perform, and again, once they are also looked at in a larger series. Remember, this is a lab-based study that has well-characterized specimens. Previous studies have shown wide variation in the sensitivity and specificity between commercially available kits. We are still tending to see this. Just remember, this is preliminary data. It's only from the Americas region. It at least suggests that there is variability in the sensitivity of available tests. Thank you.
DR. HOLLINGER: Thank you, Dr. Margolis.
Questions for Dr. Margolis? Yes, Dr. Nelson.

DR. NELSON: I was concerned about the cross-reactions with West Nile, yellow fever. As you know, in many places in Asia, like Thailand, Japanese encephalitis vaccine is routinely used in childhood. I'm just interested in what the scope of this problem is with cross-reaction. Is this known?
DR. MARGOLIS: Like I said, I have not seen the data from the Asia panel. A large proportion of their flavivirus challenge specimens are JE, both JE disease and JE vaccine. Give me a while. These data will be out, as we did with the others, in terms of it being published.

DR. DEMETRIADES: Your data showed very clearly low sensitivity and significant variability. Why do you still think that is a useful test? DR. MARGOLIS: Good point, and a good question. I think what we have seen over time -- and, as I say, I can't explain, and manufacturers may have to tell us, why there have been some flip-flops. I think what most of this has to do with is, in fact -- because, again, the prototype tests have all looked fairly good, with large panels, including good challenge panels. I think a lot of this has to do with having the right monoclonal antibodies and the blend of monoclonal antibodies.

But, look, it may wash out. All I'm saying is that at this stage of evaluation -- and it's being widely used as a diagnostic in places such as Thailand, Mexico, many parts of the world. I think we have to do more of this and figure out, is it going to be a good diagnostic? I guess I should have put "potentially is a good diagnostic." DR. HOLLINGER: Hal, the differences that you see between your assay here in the Americas and the previous assay -- there have been two other reported studies which actually compared the same tests, one in Brazil and one in Asia, which also showed this disparity in results, quite different results. Do you have any thoughts about why there might be these differences between Asia and the Americas?

DR. MARGOLIS: I will actually go back -- on many of these, if you actually look at the characteristics of the study populations, they have not been comparable. I will say one thing: At least this evaluation panel is comparable across both parts of the world, and so it may give us a better measure. In fact, as some people have felt -- and I guess I put myself in that category -- some of the variability has been due to the study designs, not necessarily to the tests. Look, now you are doing it heads-up. These are actually run -- single lots from each manufacturer in the same laboratory using the package insert methodology. The potential for study variability is fading away. Now you may be seeing differences in the tests.

DR. EPSTEIN: First, thank you, Hal, for sharing these provocative data. You were very careful to point out that these were diagnostic evaluations and that the samples came from symptomatic individuals. Kay Tomashek pointed out that the level of NS1 antigen might actually correlate with symptomatology, suggesting that these might have been higher-level samples than one might see in donors who are asymptomatic. Do you think that might be biasing us on the high side of estimated sensitivity?

DR. MARGOLIS: Very good point. As best we were able, there were some highs and lows in those acute dengues, based on PCR genome copy viremia. That was actually something that was done in the IgM tests -- high, medium, and low titers. Again, we didn't pretest any of these, obviously, with a reference NS1, but we at least tried to adjust each of those cells based on virus load from PCR. So to that degree, you are seeing highs and mediums; you never really see lows.
DR. HOLLINGER: Thank you, Hal.
Let's move on to the last talk of this morning. This will be by Dr. Susan Stramer, who will give us a presentation on "Recent Experience in Testing Blood Donors in Puerto Rico and Key West, FL."

Table of Contents

68

Agenda Item: Recent Experiences in Testing Blood Donors in Puerto Rico and Key West, FL

DR. STRAMER: Thank you. First, I would like to acknowledge my collaborators, listed on the first slide. What I hope to cover today is background regarding transfusion transmission, how we, the transfusion community, prioritize dengue. Then I'll talk about repository testing that we did in 2005 and 2007 in Puerto Rico using our blood collections, testing those by NAT, using transcription-mediated amplification, as already has been referenced. I'll talk about the documented transfusion transmissions, our selection of the BioRad NS1 antigen for investigational donation screening -- that is donation screening under IND -- in Puerto Rico and in South Florida. I'll talk about what happened this year in both of those areas and then the results of our investigational screening using NS1 antigen, including reactive rates, our confirmatory data to date in Puerto Rico and in South Florida, duration of NS1 antigen reactivity, the results in the Florida Keys, and then, lastly, donor reentry.

The AABB Transfusion-Transmitted Diseases Committee had an EID subgroup, and we listed and prioritized EIDs -- that is, emerging infectious diseases that lack a current effective intervention and may pose a potential threat to transfusion safety. We identified 68 agents. XMRV was added later. Our highest-priority agents included dengue viruses, Babesia, and vCJD. Just to transition now to Puerto Rico, where our work has been done, I remind you that Puerto Rico is part of the United States. The first large dengue outbreak was reported in Puerto Rico in 1963. Since 1998, it has been hyperendemic, meaning that all four serotypes can circulate, depending on the year. The American Red Cross collects about half the blood on the island. The remainder of the blood is collected by hospitals.

In 2005, we did an unlinked repository study, in collaboration with Kay Tomashek at the CDC. She notified us that there was a large outbreak ongoing on the island. Here you can see the epidemic curve. But by the time we initiated the study, we only tested the back half of the epidemic. We performed TMA on all the donations that are not highlighted here in this white area. We found 12 positives, of over 16,000 tested, for a frequency of about one in 1,400. Five of the 12 were positive in a pool of 16; four were positive by serotype-specific PCR performed by Jorge Munoz at the CDC in San Juan. We found 103 to 107 copies per mL. Dengue types 2 and 3 were detected, three infected mosquito cultures, letting us know that there was live virus, and one was IgM-positive.

Correcting for our mistakes in that study, this time we did a linked study on a larger outbreak area in 2007, shown here on the epidemic curve. Here units were distributed in Puerto Rico and in the United States. We export some blood from Puerto Rico into the U.S. That was linked to a look-back study. We found 29 positives. We defined positives as TMA repeat reactives, since there is no other technology that is as sensitive to confirm TMA, so we used repeat reactivity. The frequency we found was one in 529. Of the 29, 14 were positive in a pool and 12 were positive by PCR, with viral loads up to 109 copies per mL. Dengue types 1, 2, and 3 were detected, 12 of the 29 infected mosquito cultures, and six were IgM-positive.

The look-back was rather unfulfilling, in that none of the units distributed in Puerto Rico -- we actually were able to get hospitals to provide us a sample. But because all samples from diseased individuals are banked at the CDC, we were able to have one recipient sample matched to one of our donors. That was a recipient who developed dengue hemorrhagic fever. So we were able to document, as I will show you in the next slide, that one of these units did result in the transmission of dengue and the result was dengue hemorrhagic fever in the recipient.

Of the 12 units distributed in the United States, the hospitals were equally as unwilling to provide us samples, but of five individuals who, unfortunately or fortunately, all received low-titer RNA, none of the five reported signs and symptoms of dengue and for two of those we were able to get follow-up samples, which were tested at CDC and negative for antibody -- so no transmission there. We did have transfusion transmission, as I mentioned -- dengue 2, in a unit that contained 108 copies for mL. The packed red cell unit recipient developed dengue hemorrhagic fever four days post-transfusion. When we did sequence analysis of about 1,500 nucleotides from the donor and the recipient, we revealed identity for those two samples -- again, dengue 2, which is the best studied. This is the dendrogram for dengue 2 worldwide. Here you can see our donor and recipient pair. Also dengue 2 was not found that commonly on that part of the island where the recipient was from, further strengthening our case that this was transfusion transmission and not community-acquired.

There have been two other clusters of transfusion transmission already alluded to, one in Singapore, where two recipients, a red cell and an FFP recipient, developed dengue-related illness and seroconverted. The third recipient was asymptomatic. The donors in both symptomatic recipients were positive for RNA dengue 2. In the first case from Hong Kong, where one recipient developed dengue-related illness and seroconverted, both donor and recipient were positive for dengue 1 RNA. In both cases the donor was symptomatic one day post-donation. Knowing that TMA was not available for investigational screening, and we wanted to do investigational screening, and, as Dr. Margolis pointed out, the BioRad showed higher sensitivity in the earlier studies than the Panbio test, we initiated investigational screening using the BioRad test, which is a one-step, 90-minute test, a very easy EIA. We used the same repository, the 2007 repository, that I already told you yielded a rate of one in 529 TMA-positives to qualify NS1 antigen. From looking at that same repository, we saw a frequency of about one in 1,500. All three repeat reactives were confirmed by PCR. We saw a clinical difference of about threefold when we looked at RNA by TMA versus NS1 antigen. Here's a cartoon of the assay.

We began testing using the BioRad NS1 antigen in Puerto Rico on March 8. Let me remind you that Puerto Rico this was from ProMED -- had surpassed the number of dengue deaths of more than a decade, close to 20,000 cases, to November 11, with 31 deaths. The prior largest outbreak ever recorded in Puerto Rico was in 1998, when the virus killed 19 people and caused 17,000 illnesses. This year has surpassed as the highest year in Puerto Rico. We also began testing using NS1 antigen in the Keys in Florida on August 16. Just some more clips from ProMED: In 2009, 28 autochthonous cases of dengue 1 were identified in Key West, Monroe Country -- 40 years since autochthonous cases had been detected in the Keys. Then this year, health authorities once again confirmed circulation of autochthonous cases, 26, 60 cases as of 11/27. Kay reminded us today that there are now 63 cases. So again, two years in a row with circulating dengue in the Keys.

Dengue 3, a different serotype, was identified in a single patient in this year, from Broward County, Florida, and in another patient with yet a different serotype in Miami-Dade County. Again, we are seeing more cases in Florida. This is the slide that shows you our NS1 antigen data. This is the number of reported cases on the y-axis, the number of repeat reactive units. This is the 2010 epidemic, the number of cases in Puerto Rico, the epidemic curve, the historical average in orange, and then the 75 percent threshold. When two weeks in a row exceed that threshold, the CDC calls an outbreak, as they did at the end of February of this year. What we are seeing here, after screening about 48,000 donations from Puerto Rico, is that we have 32 repeat reactives, seven that confirmed, for a frequency of about one in 7,000. Here you can see the pattern of repeat reactivity -- unfortunately, it doesn't correlate with the outbreak -- and the pattern of confirmed positives, which again doesn't follow what we are seeing in clinical cases.

Looking at the confirmatory data, here you can see the collect date of the seven confirmed positives, the serotypes -- we have found dengue 1 and 4 -- the range and signal-to-cutoff ratios for NS1 antigen, from one to 11, the copies per mL generated at CDC, some as high as over 1010. Jorge Munoz at the CDC in San Juan said these are the highest-titer samples he had ever seen. These are asymptomatic blood donors at the time of presentation. Gen-Probe has already done testing. We can see the Gen-Probe signal-to-cutoff ratios, all reactive to repeat reactivity.

We are also doing a neutralization test. It was mentioned, cross-reactivity and issues with antigen testing. We are using an antigen neutralization test. For antibody, we have about four cases, IgM, IgG, indicated by the two plus or minuses. We have four cases here that were likely acute infection, three that likely were secondary infection. One thus far has infected mosquito cultures and two have been sequenced by FDA, by Maria Rios. Looking at the pattern of one of the donors in long-term follow-up who provided very informative samples, day 0 was the time of index. The program allows us to predict or back-calculate what reactivity would be. This was the donor who had virus that was isolated and sequenced, isolated in two labs and positive in mosquito culture. If you look at the PCR assay done by the CDC -- that's the red line, the NS1 antigen line is the blue line, IgM is the green line, and IgG is the orange line. I'll talk about timeframes on the next slide.

This donor did call us back one day post-donation to tell us about headache, body pain on day 1, fever, backache, and chills on day 4, and rash on day 8. I should say, this was from a survey, not donor callback. We survey all donors who enroll in follow-up. This slide is a little bit more informative. The dashed line superimposed on the lines I just showed you is the TMA data. Here we have index, day 12 of follow-up, and day 19 of follow-up, where TMA is still reactive. If you look at the duration of these markers using the midpoints to determine positivity, we have about 11 days by RT-PCR, 20 to 21 days using NS1 antigen, and at least 29 days of RNA by TMA.

Why are we seeing the discrepancy with the NS1 antigen test? Dr. Margolis certainly talked about a number of issues related to the test. We know NS1 antigen is capable of detection of dengue, perhaps at a threshold of 105 to 106, but studies haven't been done to define this. We know from some studies there is about an 82 percent agreement with PCR. These are the studies by Bissoff (phonetic) that were already referenced. Would screening with TMA look at any different? We did a small comparative study that I'll show you from September of this year for NS1-negative samples, testing 5,000 by TMA. Gen-Probe did these tests. Going back to NS1 antigen, the specificity of the test and positive predictive value, only seven of 35 repeat reactive antigen-positive donors were RNA-positive, for positive predictive value of 20 percent and a specificity of 99.94 percent.

Are the discrepancies related to the donors? We know that symptomatic individuals don't donate. The range of individuals who are asymptomatic varies. We also know that about 33 percent of symptomatic cases are in persons of greater than 15 years of age. So certainly adults will donate who are asymptomatic prior to the development of dengue symptoms or will never develop dengue symptoms. We also know that in secondary infection NS1 antigen has lesser sensitivity. This was another study that Bissoff has done, in collaboration with the CDC. We do know that high rates of donors in Puerto Rico are antibody-positive. Many that we will test, if not most, will already have heterologous antibodies.

This is the study we did at Gen-Probe, looking at over 5,000 samples that were negative by NS1 antigen, to look for TMA. Twenty-two were initially reactive; 16 were repeat reactive. Of those repeat reactives, 14 were reactive in three replicates; the remaining two were two times reactive. Ten of the 13 were also reactive using an alternate TMA test. If you look at frequencies in September, what we saw from NS1 antigen was one in 2,909 versus 18 repeat reactives with TMA, for a frequency of one in 300 -- so even higher than we saw in 2007, one in 529. In this case TMA had tenfold higher clinical sensitivity. The bars here show you the signal-to-cutoff ratios. Blue is the initial reactive S/Cos and the red is the repeat reactive S/Cos and the open blue are the IRs that did not repeat. If we look at the summary I just showed you for Puerto Rico and now switch to the Keys, we tested over 2,000 donations. Three have been repeat reactive. None have confirmed to date.

What about donor reentry? When we do screening tests, we know specificity is not 100 percent, so we're always reliant on being able to reenter donors. Our criteria for the IND require a donor follow-up sample, collected at any time following their index reactivity, to test nonreactive by NS1 antigen and viral RNA. Once the above has been met, the donor may be reinstated at 120 days post-index. We picked 120 days because that's what we do for West Nile -- nothing more complicated or sophisticated than that. No donor has been reentered to date. FDA has not given us permission yet to reenter. But of the 35, three now are eligible who have met this criterion, who are the unconfirmed group, two in the confirmed group.

Let me just show you the profiles. These donors provide us follow-up samples. These are the false positives. These false positives retain reactivity to NS1 antigen. It's not related to NS1 antigen. It's not related to dengue. It's related to mouse antibodies that these individuals have, because the reactivity can be completely quenched with increased concentrations of mouse serum. Here you can see the three who are eligible for reentry. Looking at the confirmed positives, you see a very different picture -- virus increase and then a rapid decrease, and no further evidence of RNA detection over a long period of time. Here are the two donors who can be reentered post-120 days. If you look at persistence of NS1 antigen reactivity, in unconfirmed donors it's very long, but in confirmed donors it's very short.

In summary, dengue is endemic and epidemic in Puerto Rico, with rates of donor RNA detected of one in about 1,400, in 2005, by TMA, one in 529 when we did the 2007 study, again by TMA. Looking at the same repository and NS1 antigen, all reactives confirmed by RNA, we got a rate of about one in 1,500, so about threefold lower clinical sensitivity than TMA. We know that NS1 antigen is not as sensitive as RNA -- that is, by TMA -- for blood donor screening, but does detect high RNA titers. We saw a relatively low frequency this year overall. But again, the seven that we detected all had very high viral titers. One could certainly argue that most, if not all, of those would be infectious in a recipient.

Testing over the same period for this year in September, if we look at a head-to-head study for TMA versus NS1 antigen, we did see a tenfold lower reactivity for NS1 antigen as compared to TMA. Perhaps this difference relates to what Dr. Margolis talked about in the Americas study that he just presented. The specificity of NS1 antigen is comparable to other blood donor screening tests, even though I showed you false positives. Don't kid yourself: It's no better than any other screening test we do today -- no better and no worse. We do need a neutralization test if we continue to do this.

Lastly, the duration of RNA, at least in one donor that we followed -- which is one more than available in the literature -- was 11 days to greater than 29 days, 20 to 21 days by NS1 antigen. NS1 antigen false positivity is persistent, but both NS1 antigen or RNA-confirmed and unconfirmed donors should be eligible for reentry using the criteria that I have defined. Thank you. I hope I kept within time. I talked as fast as I could.
DR. HOLLINGER: I'm worn out, Susan. Questions? Yes, Dr. Kramer?
DR. KRAMER: Thank you, Dr. Stramer. A clarification on slide number 17. There's a discrepancy between what we have in our hands and what you have.
DR. STRAMER: Yes. I fixed a typo. You have 67 percent. It should be 33 percent. What Dr. Tomashek presented today is probably higher, probably about half.

DR. HOLLINGER: Can I just go on with that? Are you sure you really didn't mean 67 percent? I thought you were suggesting -- and this is something that Dr. Nelson has mentioned -- that we have moved a little bit from seeing the disease in children more to the adult population, which would be more likely to donate blood. You had previously that about 67 percent of symptomatic cases are in persons greater than 15 years of age. I thought you were trying to tell us that there are probably more symptomatic cases in people who are older. Then you changed it to 33 percent.
DR. STRAMER: Yes, I did. Two-thirds are asymptomatic and one-third are asymptomatic and adults. That's consistent with what Dr. Petersen used in his model. I just had a typo that I realized yesterday when I was proofing my slides. But it's probably higher, as the committee has already discussed, based on data that Dr. Tomashek has already described. She said 47 percent are symptomatic in persons greater than 15 years.

DR. KEY: This does not directly relate to what you discussed, but Dr. Rios said something earlier about the virus may become cell-associated, I think red cell-associated. Can you address that at all? Perhaps this is not the right time for that, but I'm sort of interested in how it segregates in blood components.
DR. STRAMER: The short answer is, we don't know. The long answer is, we do test plasma, as Dr. Rios alluded to. For RNA, we always have used, and likely always will use, an EDTA plasma. We use a dedicated tube for RNA. But it has been shown by multiple studies, for multiple viruses, virus can bind specifically or nonspecifically to cellular components. So there may be higher concentrations of virus associated with the cellular fractions. However, we typically don't test the cellular fraction. We test what's available in plasma or representing circulating RNA in a donor.

DR. KRAMER: You mentioned that with these high counts per mL, they were asymptomatic at the time of presentation. But did they become symptomatic following -- you mentioned one, but did they all become symptomatic following --
DR. STRAMER: Puerto Rico has been very good in our region as far as donor follow-up. They have to deal with the threat of me if they are not following their donors. We followed virtually all cases for symptoms. This is the only donor that I showed who reported symptoms post-donation.

DR. NAKHASI: Sue, in slide number 6, which basically says that you picked up by TMA 29 out of 15,000, which is a prevalence of one in 529 -- when you did the pool of 16, only 14. Does it mean that if one has to test in pools, you will miss more than 50 percent?
DR. STRAMER: Yes, and that reproduced from 2005. On the prior slide, I showed the same data, where 5 of 12 were reactive in a pool of 16, and then in this study, 14 of 29. We know certainly for West Nile, at least in our studies at the Red Cross, that about 30 to 35 percent of West Nile positives are not detected in pools. That's why we trigger in the summer.

DR. NAKHASI: I understand that. With the new methodologies reducing the pool size, have you tested that like the ULTRIO Plus assay from Gen-Probe -- whether you can increase the sensitivity of this detection?
DR. STRAMER: Theoretically, I guess you could, but they are still likely to miss positives. One can argue about the infectivity, and a lot of the low-level TMA positivity we see may be antibody-positive. Whether they are heterologous or homologous antibody we don't know. We don't know about the infectivity, only the high-titer units of infected mosquito cultures. So we really don't understand the differential between those that are detected in pools, which are the ones that are quantifiable PCR-positive and infect mosquitoes -- we don't know what the low-titer RNA signals mean.

DR. RENTAS: You mentioned at the beginning that you are collecting half of the blood supply in Puerto Rico. Does that mean that the other half is not being tested at all for this?
DR. STRAMER: Yes. We are doing an investigational protocol right now. The other half of the island, which is up to 20 independent hospitals -- we have no affiliation with them.
DR. KLEIMAN: Steve Kleiman, from ABB. Sue, on the one slide where you showed the sequential TMA data in the followed donor and you showed a data point at day 19 where the TMA was still positive -- day 0 being positive and day 19 being positive -- but then you made a conclusion that you projected there would be at least 29 days of RNA positivity. I didn't follow how you came up with 29 days, since your next data point is all the way out at day 54, and it's negative.

DR. STRAMER: I took the midpoint from 0 to 10, which is 5 days, to project the ramp-up. I didn't take the midpoint here. I took another 5 days at the end here, because this was a ridiculously long inter-donation interval. I took 5 days at the beginning and 5 days at the end to come up with 29 days. It's not precise. We don't have an exact estimate. But it's certainly longer than the observed 19.
DR. KRAMER: On that same slide, that's one donor, right? With West Nile, what proportion of donors are positive out to 120 days -- or, let's say, beyond the 54 days?
DR. STRAMER: Extremely small numbers. In the modeling study we did in 2003 with 186 donors, we modeled that 99 percent viral clearance would occur at day 50. The longest donor we had ever seen positive was day 56. But in studies done by Mike Busch at Blood Systems, they saw four donors which had longer periods of time, one out to day 104, with one rep positive out of five total reps tested. I actually included that in the reentry slide to show you the specific data for why we chose 120 days for reentry.

It is, if you will, excessively long, 120 days, and all of those individuals were IgG-positive already. We know IgG positives, at least for West Nile, have not been associated with a documented transfusion transmission.
DR. HOLLINGER: Other questions?
Sue, since you are still up here, rather than bring you up right after the break, on slides 7 and 8, there is some indication here that there could be a threshold at which transmission may occur. This has to do with the NS1 and using TMA or PCR for detection of dengue. First of all, it's unfortunate -- and we have talked about this before -- that you had 32 that you could look at, but you weren't able to locate 15 of them. I understand the issue with the hospital, but I think that's a real tragedy, that we are not able to follow all of these critical patients who might be available.
DR. STRAMER: In those cases, I have talked to the hospital directors myself, because I was so bound and determined to try to get these hospitals to try to follow these recipients. Five of them actually came from one hospital that is considered a research hospital, and they wouldn't even participate.
DR. HOLLINGER: I think that is always unfortunate.

I have a question about the NS1s. The patient who developed dengue hemorrhagic fever -- the donor of that patient, was that NS1-positive?
DR. STRAMER: No. NS1 antigen wasn't done on that subset of donors.
DR. HOLLINGER: So it wasn't done.
DR. STRAMER: No, but it was 108 copies per mL.
DR. HOLLINGER: So it may have been positive.
Then you had five others in the ones distributed to the U.S. that had no signs or symptoms. Two of them were IgM/IgG-negative. You said they all had low-titer RNA. Could you tell me what the titer was of the RNA in those five individuals? And were they NS1-positive as well?
DR. STRAMER: None of the donors in the study were tested by NS1 antigen. So, like the one who donated to the recipient who had hemorrhagic fever, I don't have the NS1 results. Samples are depleted, and I can't go back and test that. But as far as the low titer, low titer was 103 or less, as reported out by the CDC. When I say "less," I mean we didn't get a quantifiable viral load.

DR. HOLLINGER: Okay. Finally, on your slide 13, on the supplemental testing, it may be important to point out that, at least in these supplemental tests, all of the positives here by TMA were also positive by NS1 antigen. Is that correct?
DR. STRAMER: Yes, because NS1 antigen was the selection criterion. They were screened by NS1 antigen -- so RNA, either by CDC or by TMA as our confirmatory criterion, along with neutralization testing by BioRad.
DR. HOLLINGER: Very good. Thank you. We're going to take a break now. It's around 10:22. We'll take a break until 10:40. Then we will be back in here to start with the open public hearing. Thank you.
(Brief recess)

Table of Contents

88

Agenda Item: Open Public Hearing

DR. HOLLINGER: This is the open public hearing. There is an announcement that I need to read for all the attendants here today about the open public hearing. Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making. To ensure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is mp to understand the context of an individual's presentation. For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement to advise the committee of any financial relationship that you may have with any company or any group that is likely to be impacted by the topic of this meeting. For example, the financial information may include the company's or a group's payment of your travel, lodging, or other expenses in connection with your attendance at the meeting. Likewise, FDA encourages you at the beginning of your statement to advise the committee if you do not have any financial relationships. If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

Having said that, we have one individual who has requested to address the committee. That is Dr. Louis Katz, representing the AABB Transmitted Diseases Committee and its relationship with the American Red Cross and America's Blood Centers. DR. KATZ: Hi. My name is Louis Katz. I'm the medical director at Mississippi Valley Regional Blood Center in the Upper Midwest and chair of AABB's Transfusion-Transmitted Diseases Committee. AABB pays for my travel if they like what I say. I will take money from anybody in the room under appropriate circumstances. I have gotten research support within the past year from Roche, Ortho Clinical Diagnostics, Micronics, and, I think, Abbott, for assay development, none for dengue.

Thanks to the chairman and the committee for the opportunity to provide the thoughts of the blood community on the issue of dengue and its relevance to transfusion medicine. My statement is presented on behalf of AABB, the American Red Cross, and America's Blood Centers, together responsible for over 90 percent of the volunteer blood supply in the United States. There have been reports of transfusion-transmitted dengue from three donors to date, in Hong Kong, Singapore, and Puerto Rico. This low number is somewhat reassuring, given the fact that more than 50 million infections occur worldwide annually. However, it needs to be recognized, as you have heard, that surveillance for these events has not been a priority in endemic areas, and transfusion transmission in such areas is difficult to differentiate from mosquito transmission. No dengue transmissions have been recognized in the continental United States and Hawaii, despite endemic dengue in the Rio Grande Valley, outbreaks in Hawaii, cases in the Florida Keys, and certainly among travelers returning from endemic and epidemic areas.

We believe that the cornerstone for mitigating the risk of transfusion-transmitted dengue viruses in the U.S. is public health surveillance for autochthonous infection, with subsequent interventions scaled to estimates of risk. Dengue has become nationally reportable as of late 2009, which should facilitate this activity. We, the blood community, have encouraged the potential manufacturers of dengue blood donation screening assays to develop prototype kits that could be available for use under investigational protocols, should dengue more firmly establish itself in the continental United States and Hawaii.

Although dengue has affected a substantial proportion of the population of Puerto Rico in outbreaks since 1963, travelers to the island or to other non-malarious urban areas endemic for dengue appear to be at a much lower quantitative risk of acquiring dengue infection than do residents, due to their short length of time in Puerto Rico and likely a higher standard of housing while in the endemic area. This very low risk does not justify deferring such travelers from blood donation at this time. A risk model for travel to Singapore, a dengue-endemic area, published in 2009, estimated a risk of .17 percent during a one-week stay during the peak of a severe epidemic in 2005, versus .004 percent during the low season of a non-epidemic year. From 1996 through 2008, 1,093 laboratory-confirmed dengue cases were imported into the U.S. If 67 percent of dengue infections are asymptomatic, then a minimum of approximately 3,312 infected donors entered the U.S. over those 13 years. In the same interval, there were 18,322 reported malaria cases.

Most rural areas affected by dengue have endemic malaria as well, and travelers to those areas are already deferred for this risk. In the absence of recognized transfusion-transmitted dengue in the continental U.S. and Hawaii, we do not think a dengue-specific geographic deferral is appropriate at this time, and certainly not in the absence of a clear understanding of the impact of a putative deferral on the adequacy of the blood supply.

Available data suggest that viremia is short-lived in most, if not all, asymptomatic dengue infections. Based on extrapolation from West Nile virus infection, it is likely that infectivity to a recipient would only be present prior to the development of IgG antibody. Thus, if a travel deferral were to be considered, it should be established separate from existing malaria policies and confined to the interval of known asymptomatic viremia prior to the development of IgG. Based on available information, less than or equal to 28 days would appear to avoid viremic, infectious donors, with a substantial margin of safety. This is consistent with evolving approaches of which we are aware in the European Union and the Asia-Pacific regions. The deferral period for healthy travelers to dengue-endemic areas should not be of the same length as the currently proposed 120-day interval used to reinstate donors with documented asymptomatic -- RNA or antigen-positive -- or symptomatic dengue infection, as in these latter cases it has been established that dengue infection has actually occurred.

Dengue is established in some areas of the United States and its territories, with varying levels of transmission from year to year. Given the availability of dengue tests that could be developed for use under an IND, we think it is reasonable to recommend their use when a pre-agreed-upon threshold of activity is reached in a region, unless blood centers voluntarily implement testing prior to an agreed level of activity. During the interval before the implementation of such testing, it may be reasonable, after an assessment of the adequacy of the blood supply, to suspend collections in such a region during peak dengue incidence. Thresholds should also be established for discontinuation of such measures when appropriate.

We ask for an open dialogue with FDA and CDC on formulating definitions of the level of autochthonous infection that will define an outbreak in the U.S. and an understanding of what minimally constitutes the establishment of endemicity in regions of the U.S. and its territories, as well as who is responsible for determining and communicating that an agreed-upon threshold for intervention has been reached. Likewise, we believe that there is a need to establish in advance who will be responsible for collating worldwide data and designating areas or countries as epidemic or endemic, as local epidemiology evolves over time.

To date, the American Red Cross in Puerto Rico and an independent blood center in South Florida, in collaboration with the ARC, have voluntarily instituted blood donation screening with the best test available under IND in response to epidemic dengue in Puerto Rico and a small number of cases in the Florida Keys.

Retrospective studies performed by the American Red Cross in Puerto Rico clearly suggest that nucleic acid testing would be preferable to NS1 antigen testing, from the standpoint of sensitivity and predictive value, although antigen testing has interdicted high-titer viremic donations. Antibody assays will not be useful for donor screening, since a robust antibody response coincides with loss of infectivity.

Again, we support and encourage dialogue aimed at developing approaches to the prevention of dengue transmission in case the worldwide epidemic becomes established in the continental U.S. and Hawaii. We would like to see the successful cooperation between the blood community, diagnostic manufacturers, public health agencies, and regulators, similar to the process developed for the West Nile virus epidemic, replicated as we address the potential for a dengue epidemic in the United States.

The remaining three paragraphs describe the three organizations contributing to this statement.
Thank you.

 

DR. HOLLINGER: Thank you, Dr. Katz.
Any questions from the committee for Dr. Katz and his statement?
(No response) If not, is there anyone else in the audience who would like to take advantage of this public hearing at this time?
DR. DEMETRIADES: Dr. Katz, might I ask you, what is the current status of the blood supply in the country? How would exclusion of endemic areas affect this issue?
DR. KATZ: Are you talking about A-positive or O-negative red cells? Are you talking about apheresis platelets of what type? I think that's a very difficult question to answer. I will say, very generically and subject to many, many qualifications, the red cell supply at this point is as robust as it has been for my career in blood banking, which goes back to the 1980s. In general, the supply of red cells, if you don't factor in type-specific issues like O-negatives, B-negatives, and others, is relatively good.

Having said that, we don't yet have data on the number of additional deferrals that would occur, above and beyond what we already do for malaria, were we to add particularly the large urban areas with dengue epidemics to a deferral. It also depends, of course, on the duration of the deferral. So the short answer to your question is, we don't really know. My personal belief, looking at where people are coming back from, is that it will be substantial. We'll be getting that data over the next several months.

DR. HOLLINGER: Any other questions from the audience? Yes, Dr. Epstein?
DR. EPSTEIN: Dr. Katz, I have a question. Basically what you are saying is that you would envision a strategy in which we monitor dengue activity in the continental U.S. and trigger testing only if we reach certain predefined thresholds. What you are advocating for endemic areas where there is already defined autochthonous transmission, such as Puerto Rico, which undoubtedly would meet any predefined definition, there you are calling for investigational testing.

What do you see as a long-term strategy?
DR. KATZ: Jay, I think we both realize that to move beyond investigational testing requires a licensed assay. In some certain sense, one needs to understand what the market for the assay is. I think we use an interval of investigational testing to try to figure out where we're going, what's happening to dengue in the U.S. and around the world, and go from there. It's a new paradigm, obviously. It's not the way we have handled things previously.
DR. EPSTEIN: My concern is -- we have this chicken-and-egg problem, where the diagnostic manufacturers repeatedly say to us that nothing would more clarify the situation than an FDA mandate. What that translates into is a decision -- if we have a suitable test, would we use it now? Simply to call for investigational testing -- well, of course we need that. We have to define the performance characteristics. But I think part of the dilemma is to decide, if we have acceptable tests, would we move toward recommending them. Are they warranted, in other words?
DR. KATZ: I guess my response to that is that if we have a sense at what level FDA, the blood community, public health agree that this is enough to do it, the companies may then have the calculus they need to figure out where to go with this.

DR. EPSTEIN: My point is that, really, the joint organizations have not weighed in on that point. You haven't taken a position on whether you think the data do warrant an intervention.
DR. KATZ: I think that we support what has been done in Key West. I think that's appropriate. I guess we haven't asked that question of the TTD or the boards of the organizations. But I think that the level reached in Key West supported the actions that were taken.

DR. KLEIMAN: Just to comment further, Jay, it's another one of these regional organisms, like Babesia. Certainly, if you ask the question of whether we think it has reached the level where the entire United States blood supply should be screened -- again, I can't speak for the organization, but I can speak for myself as an individual, and I think we have had these discussions. Not many people, if anyone, think that you should recommend that dengue virus testing be done on every blood donation. I think you already know that. But the question is, has it reached a level in Puerto Rico where there should be routine testing? Again, we didn't take a position on that, but Lou gave his personal opinion, and I would agree with that.
DR. HOLLINGER: Dr. Nelson?
DR. NELSON: Maybe some of the committee members or somebody from AABB can answer this. I'm interested in what the degree of overlap is between the malaria-endemic areas and the dengue-endemic areas. My sense is that the dengue-endemic areas are considerably larger, because they would involve the whole Caribbean, et cetera. Is that correct?

DR. KATZ: When you look at the maps and you compare the dengue maps and the urban epidemics of dengue to the distribution of malaria and the relative lack of urban malaria in places like Sao Paulo and other very common tourist destinations, yes, we feel that this is a substantial additional set of deferrals, on top of what we are already seeing for malaria. We need to analyze that before we jump in with both feet. That should be a fairly simple task to do, over a period of a few months, to accrue that data set.
DR. BOWER: I would also like to get maybe Hal or Kate to comment on this estimate of approximately 3,300 infected travelers coming into the U.S. I think there is probably a huge amount of underreporting for people coming in. I just wanted to see if there is any guestimate on what the number of infected travelers returning to the U.S. would be.

DR. MARGOLIS: We do not have a better guestimate. I think Lou's approach is as good as anyone's. There have been proposed studies. I think the thing that the committee needs to continue to recognize is how under-recognized dengue is. The Key West epidemic was found two-thirds of the way into the epidemic, only because a clinician, who actually was from Colombia, said, "Gee, this looks like dengue," after this person had had the $64,000 workup in New York City. So this is the problem with dengue.

No, we don't have a better estimate at this time.
DR. HOLLINGER: I just want to make sure that there is nobody else in the audience that wants to speak to the issues.
(No response) If not, I'm going to close the public hearing. We will open it up for committee discussion now, which can direct questions to each other, even to other members of the audience if they wish. There are a couple of questions that we're going to have to deal with. It has to do with traveler deferral and what to do with the patient, what an outbreak is, and so on. I would like to start the committee discussion about traveler deferral. Then we'll deal with, maybe, patient deferral and then deal with what an outbreak is, what an epidemic is, and so on. Let's just deal initially with traveler deferral.

Just to start, I want to mention that there is a paper out this year by an author, Teo (phonetic) and colleagues, which looked at traveler deferral in various countries in the world. Seven countries have no travel deferral at all. There were several countries that used two to four weeks and a few, I think, that used up to six months. So just to start it, there are some time intervals that one can deal with. Let's start the discussion. Yes, Dr. Bower?

Table of Contents

102

Agenda Item: Open Committee Discussion

DR. BOWER: I just wanted to start off by asking the FDA, are there any other viruses, acute viruses that have a short incubation period, where we currently defer people returning from an endemic area?
DR. RENTAS: I think you hit that nail right on the head. I was just going to ask the FDA to define "traveler" for us. At what point do you become a resident, and not a traveler anymore? Is it six months? Is it a year? Is it three years? Is it five years? Is it two months? I know we do have some guidelines for malaria, but I don't know if we're talking about the same thing here.
DR. HOLLINGER: Anyone have an idea what we would define as a traveler?
DR. NAKHASI: Not me, because I'm not an epidemiologist. I would default to CDC. They may be able to tell us how they do define a traveler to a particular endemic area.

DR. NELSON: It seems the risk is being there, not what your citizenship is. So it doesn't matter whether you are a citizen or a resident or whatever, if you are where there is dengue transmitted. Isn't that what we're talking about?
DR. BOWER: If you get bit the day you are leaving, that's really what the risk is.
DR. HOLLINGER: He is asking about duration. Is a traveler somebody who spends just a week somewhere? Is it somebody who spends a month? At what point do you become not a traveler? I guess that was the issue. Jay?
DR. NELSON: Well, how long does it take a mosquito to bite?

DR. EPSTEIN: I think thinking of this analogy to malaria is an error. The issue with malaria is, how long does it take to become semi-immune, in which case you might be parasitemic for a very long duration and asymptomatic. So we differentiate the period of deferral after leaving the exposure region based on how long you have been there because of the risk of acquiring a semi-immune state. Here we're not talking about that at all. The only issue that matters is, how long has it been since you left the area? You could have been bitten in an area of activity, of course, as late as the day of departure. So it really doesn't have anything to do with how long you were present. It has nothing to do with citizenship or resident-alien position. That's irrelevant. We're just talking about time of exposure.

But I think time of exposure is not relevant in dengue. If you are in an outbreak region, and we don't know whether you had prior infection, we don't know if you are seropositive, we don't know if your last infection was secondary or tertiary -- it would only be a question of, when did you leave?
DR. HOLLINGER: But I think you would agree, though, Jay, would you not, that an expatriate, for example, or someone else who goes back home to a rural area or to an urban area, but lives in a house that is non-air-conditioned, maybe unscreened, water in the household, in flowerpots or wherever else, is probably at higher risk than someone who goes to a resort which is air-conditioned or a hotel which is air-conditioned? I'm not saying the risk is zero, but I'm saying that the risk is much less in those.

DR. EPSTEIN: I would agree with that. It would be very difficult -- again, in the case of malaria, we have had to make a decision whether to exempt travelers to resorts. It's just driven by the epidemiology. It depends at what granularity we are going to be able to differentiate risk. I think that would be very difficult for dengue, because it's also an urban transmitted infection.
DR. HOLLINGER: Dr. Tomashek?
DR. TOMASHEK: I agree with the comments that were just made. For travelers, what we do is a person who has been outside the United States in the two weeks prior to symptom onset, considering the two-week incubation. The issue with the duration of travel -- we know from a very limited number of studies that the seroconversion or the risk of being infected or getting dengue while abroad varies by duration of study, but, really, in terms of our concern here, a traveler is defined as someone who has been abroad and then returns and develops symptoms within two weeks of their return.
DR. HOLLINGER: Dr. Katz.
DR. KATZ: On the Teo paper that you cited, it's important to know that the places that defer, defer for fever, not for dengue. They defer for fever within some arbitrary interval after return to the home country.

The other thing that we tried to articulate in our statement -- but I think I probably didn't do a good job -- is that we need to understand the difference between epidemic dengue in the face of some endemic level or endemic establishment somewhere -- at what levels are we talking about triggering responses? For example, in Puerto Rico, where you saw very nice data from Dr. Tomashek over some long period of time, and from Lyle's modeling paper, low levels at long stretches during parts of the year, and over some years that didn't have serious peaks during the wet season and then you would see peaks when things got worse -- we are interested, in the blood community, in understanding, is it any established endemic dengue or at what level of established endemic dengue? If you have a low level of established endemic dengue, what kind of a peak does it take to trigger a new intervention?

DR. RIOS: I just want to point out two things here. One is that when we get concerned about dengue transmission by transfusion, we are not taking into consideration those that become sick, that in two weeks after travel, have symptoms. We are concerned about those that come and we never know, and they will never know if they have been bitten or not. Also the mosquito that is better suited for transmission bites during the daytime, and not always inside houses. Just as you walk in the parking lot or whatever, you can be bitten. I can tell you, because I'm from Bahia, Brazil, and we have an air conditioner in our home. Nevertheless, you open the door to go in and out, and you have mosquito biting.

All those precautions may be well taken in the U.S. when you go everywhere and it's air conditioners everywhere and so on. But in countries where you have a less tight environment for hot and cold, I don't know how much that would account for prevention of mosquito biting.
DR. HOLLINGER: Yes, Dr. Bianco?
DR. BIANCO: I want to just point out that there is the science and there is the counting days and hours and exposure to mosquitoes. But I think the biggest concern that we have in the blood community is a concern of impact. Dengue is so widespread, there are so many people at risk 2.5 billion, we heard from Dr. Taylor -- that if we are looking for zero risk, we are going to defer half of the world population. We would like us to consider it in the context of impact. That is the most important thing. We don't know at this point what the impact is of deferring people that visited large cities in Asia or in South America from donating blood.

DR. SPENCER: Brian Spencer, American Red Cross. I want to comment on the issue of overlap between malaria- and dengue-endemic areas. The REDS program financed by NHLBI has published on this. Data from a few years ago from the Office of Tourism suggest that about 28 million people per year travel to malaria-endemic country, any country of the 100 or so that are considered malaria-endemic by the WHO. Our estimate was that maybe about 8 million of them had traveled to areas where they could have potentially been exposed to malaria, from the very lowest end of the spectrum up to high-transmission areas in Africa or Papua New Guinea. Very few parts of the Caribbean have malaria, just Hispaniola. So you take all of the travelers to the Caribbean, outside of Haiti and the Dominican Republic, which is in the single-digit millions, and all of those people would be newly deferred, depending upon whatever interval gets established.

Likewise, Brazil, which is endemic for malaria, is endemic in the Amazonian region only. Everyone who goes to coastal Brazil -- the large cities of Rio, Sao Paulo, beaches in the northeast -- those people would be newly added, not to mention parts of Asia which have dengue, but not malaria. So the marginal impact of adding a deferral for a travel risk for dengue could be quite significant indeed.
DR. HOLLINGER: Yes, Dr. Bower?
DR. BOWER: When we are talking about these deferrals, though, it is quite different for malaria, because the deferral for malaria is for an extended period of time. The deferrals that we are talking about for dengue virus are sort of like a month. I agree, this is probably the toughest question that we have to face right now. The number of people that may come back to the United States with disease appears to be somewhat low, although I don't think that we know that. Using the precautionary principle, you would think they probably should be deferred, as long as it doesn't have an adverse impact on the blood supply. I think that's what we are quibbling about right now. I just want everybody to realize that the deferment period is not going to be as long as it would be for malaria.

DR. HOLLINGER: Yes? Name and organization?
DR. KUMAR: Sanjay Kumar, from FDA.
The parallels are being drawn again and again between overlapping malaria- and dengue-transmission areas, and it's being assumed that the risk is the same. While malaria is absent from the major metropolitan cities, in the countries where malaria is actually -- in the most endemic countries, you go to the major cities and there's no malaria there -- 18, 20 years ago -- never experienced malaria activity. On the other hand, the only place where dengue is found in the same endemic countries as malaria is only in large cities. So where do the travelers go? Mostly they go to large cities. MS. CARR-GREER: Allene Carr-Greer. I am with AABB.

One thing that we're always concerned about when we speak of new deferrals for blood donors -- deferral for a new reason -- is that -- certainly, if the occasion is called for, then that's what we should do -- when a donor is deferred, we always run the risk that we never see the donor again. Donors don't like being rejected, and so we don't see them again. Many blood donors appear to donate at blood drives. Blood drives generally don't happen any more often than every 56 days. If we were to select a reasonably short period of time for the deferral, then we still would likely not see that donor again for 56 days, if they return at all. That's what we always ask for consideration of: If you defer a donor, we may not ever see them again.

DR. PETERSEN: When we think of traveler-associated dengue, I think the natural inclination is to think of somebody going to the Caribbean or some foreign place, but I would like to remind you that the U.S.-Mexico border is like the "Great Wall of Dengue." There are tremendous amounts of dengue right on the other side of the border. About 1 million people cross the U.S. border every day into Mexico and back. Really, what we're talking about is deferring people that essentially could go five miles away from their homes and cross into a dengue-endemic area. Incidentally, there are about 22 million visitors to Mexico every year.
DR. HOLLINGER: It is interesting. In Texas, however, this year there have been only 17 cases of dengue reported to the state epidemiologist. All of them are imported, none of them from Mexico.

DR. NELSON: There was an interesting paper a couple years ago -- (Off-mic) -- there was a raging epidemic across the border, but in Brownsville there were five cases or something. They looked at the risk. Everybody in Brownsville lived in air-conditioned houses, had no standing water, et cetera. But the conditions were quite different a couple of miles away. But, of course, travelers, when they go to Mexico, maybe don't all go to air-conditioned places.
DR. KRAMER: Can I just comment on that paper? Also in that paper, which I thought was a great paper, they pointed out that a lot of the people living there were going to Mexico to the doctor's, so they weren't reporting in the U.S., because they couldn't be treated in the U.S. legally. So it was an artificially low incidence in the U.S.

DR. HOLLINGER: Dr. Alter.
DR. ALTER: Harvey Alter, NIH.
I think Lyle's point is much more important than the one I'm going to make. We haven't talked about cumulative risk. We are organizations of compromise. We are always compromising risk versus donor loss: How much can we tolerate? I think in the variant CJD model, if you rule out everybody who had been in England, we couldn't tolerate it. So we made some time in England that that became important, or seemingly important. While it's true that you can get bitten by a mosquito as you enter the plane in Puerto Rico, your chances of getting bitten -- not every mosquito is a carrier -- your chances of getting bitten increase with your duration in a facility and the conditions within that country. So one could come up with a time of exposure, as opposed to any exposure. We could debate that point. But I think we could do a controlled study where we send me to a plush resort in Puerto Rico for a month and see what my risk is and we send Blaine to a hovel in the mountains, and we see who gets it.

DR. HOLLINGER: I think you're always wanting to send me somewhere, Harvey. Along those same lines, just as a piece of trivia, the average age of a female mosquito is about 8.6 days. The extrinsic period is longer than that. Most of the female mosquitoes will be dead before they can transmit. That doesn't mean they can't bite, but most of them are not going to bite at that time. So it's always the tail-end of those mosquitoes that live longer that might cause a problem.

DR. KATZ: Our main request is that we study the impact before we move on this. It occurs to me that it's not just who has been to Sao Paulo or who has been to Jakarta or who has been to Singapore, but also when people donate after they return. For example, my first thought when I come back from vacation and I have been having fun for a week or 10 days, in fact, is not to go donate blood. My impression in the blood center is that people don't get off the plane from Mexico or the Caribbean or elsewhere and come in and donate blood. So there may, in fact -- and we don't know the answer to this question -- be a built-in protection, in that people delay donation until they have recovered from their vacation, that sort of thing. So there is this entire set of data that we need to collect before we move on geographic deferrals, in my opinion. I want to point out that the FDA in their briefing materials, in fact, proposed a 120-day deferral. We are not talking about a trivial deferral, at least accepting what is in the briefing materials.

DR. HOLLINGER: Could I clarify something also? Somebody maybe could clarify for me, from the CDC or otherwise. You mentioned, I think, in your report that there were 1,093 confirmed dengue cases imported in the U.S. during 1996 to 2008. I thought it was much more than that each year that are imported into the United States of dengue. Am I wrong? A number like 500 sits in my mind, that there are 500. Then you take the 67 percent. That would make about 900 or so cases imported. Does somebody have that data? I would like to clarify that.
DR. MARGOLIS: It's about 400 to 500.
DR. HOLLINGER: About 400 to 500 a year imported, okay.
DR. MARGOLIS: Recognized.
DR. HOLLINGER: Recognized, okay, confirmed.
So there is a lot more, I think, than here. I just wanted to clarify that.
Yes, please.
DR. BENJAMIN: Richard Benjamin, American Red Cross.

As we consider travel deferrals, can I just restate the obvious? To date, as far as we know, we have no recorded cases of transfusion-transmitted dengue in the mainland USA caused by travelers. N equals 0, as far as we know. Dr. Petersen in his model made the assumption that there was 100 percent transmission by a viremic blood product. There is no evidence to support that point. In fact, all the anecdotal evidence we have says it's probably wrong. We should remember the Chagas disease situation, where we put in place safety systems for something that turned out not to be a transfusion risk, mostly, except for platelet products. So let's be careful as we consider travel-related deferrals for something where as of yet we have no evidence of harm to donors. I think we all believe we want to protect patients from this risk, but let's not overstep the mark, when there is no evidence yet, really, of harm.

DR. HOLLINGER: I think, along those same lines, one of the issues here, of course, is not so much that -- Scott Hofstead (phonetic) would say that travelers are the ones who primarily bring a disease into an area. The question isn't whether it takes hold and whether you get a lot of cases with it. The question is, with the traveler, if he is infected, could that traveler that donates blood transmit dengue to someone else?

DR. NAKHASI: We need to focus here on the issue. It is not a question of how much time for the travel. It is basically what you heard this morning from Sue's presentation. If a person gets infected, how long will the viremia be there? We need to focus on that. If a person is going into an area where there is a chance of getting that person infected, how long after that -- we are going away from that question. It is not that it's six months of travel or two months of travel. It's the question, when the person leaves that place in an endemic area, how long is that person potentially going to be viremic, and how long should we be deferring that person? That's the time period.
That's one point.

The second point is, I want to reiterate, a comparison of Chagas versus dengue is totally apples and oranges here. We know virus is in the people. It has been transmitted. Everything has happened. It happened in Chagas. The frequency -- we did not know at that time what was the frequency. At this point, once we start looking at the cases, and if there are transmission transfusion cases, we will know. For example, in West Nile virus, if we remember clearly, we thought there was only 28 days of viremia initially. Then, when we looked at it, when the testing started, we could see virus up to 104 days.

I agree with Lou and others that the data may not be enough at this point, but what we need to focus on -- if this is a virus which can be transmitted through transfusion, which has been established -- there is a percentage of people getting more infected in certain areas what kinds of interventions do we need? The very first question is, what kinds of interventions do we need to do, and how long should that intervention last?

DR. KRAMER: That brings up one of the things that they have been doing in Australia. I have been quickly trying to review this Australia paper. They have a similar situation, where they have areas in Australia where they have a lot of activity and then there is all of Australia that doesn't. They discard the fresh products and they use viral inactivation on the plasma products. That hasn't been discussed at all, whether that's a possibility, rather than deferring -- I think they also have targeted questionnaires. I don't know whether they defer people also. I'm not familiar with it enough. But we haven't talked about that as a possibility.

DR. HOLLINGER: Here I would like to jump in a little bit on what you said. Let's assume, from the knowledge that we know, the scientific evidence that we know, that if I were to get infected the day I leave a place that has dengue, and the intrinsic incubation period, which is the incubation period going on in me, is, at the most, 15 days -- there is some data, actually, that would suggest that if it's longer than that, it may be with an attenuated virus. If it's 15 days, and then I get sick at that point and I get febrile, then the probability is that the longest that I might be infected is 10 days. If I take all that information, it would suggest that a month deferral should be sufficient. First of all, we don't see a lot of transfusion-transmitted disease. Part of that may be surveillance. But if we assume that, then a month should get us out of the woods on that, if one is going to give a deferral. We have a lot of other issues to talk about, but in terms of that, for me, it would seem that 30 days ought to be sufficient.

DR. NELSON: I agree with Dr. Katz, I think we need some data on when people donate blood and what impact this would have on the donor population. I think it's right that people probably don't get off the plane and go directly to the blood bank. But they may within a month. Before we impose a donor deferral, I think it would be useful to have some more data on blood donors.
DR. HOLLINGER: I agree with that, Ken. But would you agree, if a deferral -- I was only making the point that if a deferral was made, this would be a reasonable time to make it, not the other issue.
DR. NELSON: I agree. I think the outlier on the West Nile viremia is not really relevant when we're dealing with this, because we don't know even what the risk is. But clearly the Red Cross data has shown that in Puerto Rico some donors have pretty high levels of virus. I just can't believe that there is no transfusion risk. We have that piece of data, but what we don't have is the data on the blood donation behavior.

DR. HOLLINGER: Does anyone have an answer to the question about how many people might donate -- travelers who come back to the United States?
MR. SPENCER: Bryan Spencer, American Red Cross.
The same REDS paper on malaria deferrals actually collected information -- we had information on the date that they presented to donate and were deferred, but we also recorded the date that they had last potentially been exposed. So we do have interval data between when they returned and when they came to donate. Of course, this is for a 12-month deferral. That's our sample. But from over 2,000 blood donors deferred for malaria, 40 percent came back within the first three months.

DR. NELSON: How many donated within a month of leaving the area? Do you have those data?
MR. SPENCER: I have it, but that is not in the published paper.
DR. NELSON: You need to look at the data again maybe.
MR. SPENCER: It's 40 percent within the first quarter. If the period were as long as 120 days, you would probably be bumping up against 50 percent.
DR. DEMETRIADES: Along the same lines, what is the basis of the 120-day deferral? Do you have any science, direct or indirect?
DR. HOLLINGER: Let's deal with that in just a minute, because that's about patients. I believe the 120 days that they were suggesting has to do with somebody who has disease.
DR. NELSON: There were still some West Nile donors that were still viremic -- but a small number, I think.

DR. NAKHASI: I just want to make a clarification here. That was just a recommendation, just a top-of-the-head, as Sue said, based on the West Nile. But with the data you have, you need to make the judgment about how long we can have -- just as Dr. Hollinger systematically analyzed the data. So 120 days is not etched in stone. It was just a recommendation, depending on West Nile.
DR. HOLLINGER: It actually came from –- again, we have discussed this, and I agree with Hira. Comparing dengue with West Nile, I think, is not very appropriate. That's my personal opinion. These are different viruses. The mosquitoes are different genera. They cause different diseases. The data often came because the RNA -- not viremia, necessarily, but the RNA -- was detected out to 105 or 109 days with West Nile. That was one of the reasons, I believe, that someone threw out the 120 days.

DR. NAKHASI: If you recall, when we did the West Nile, we had recommended 28 days or 56 days. Then, as the data became available, we changed the recommendation.
As you said, we need to have more data. But at least some data points to us that up to 29 days, RNA can be detected.
DR. NELSON: Actually, the duration of IgM in West Nile and Dengue appears to be different. IgM persists quite long. So these are different viruses.
DR. KLEIMAN: Just one more further clarification about infectivity and the number of days. Even if you take the West Nile parallel, it's important to remember that in West Nile we have not seen a transmission from a unit that was viremic and IgG-positive. All of those late viremia units at 104 days -- or anything beyond about 35 days -- already have antibody.

While FDA took a very cautious view of how long a person with West Nile should be deferred, based on very small numbers of donors who had demonstrable RNA, they never showed infectivity in vivo in those kinds of donors. I think it was, in my opinion, overly cautious for West Nile, and I don't think we should get into the same extrapolation for dengue. Maria probably will disagree, but again it's in vivo and not in vitro, not infecting cells in culture.

DR. RIOS: Actually, I just want to make a comment, Steve. When you stated so bluntly that there is none, it implies that you test every single recipient and donor. Outcome of infection does not say that the infection did not occur. In fact, it's the sole reason why we are discussing here. When we think about a 15-day incubation period, we are thinking about somebody that got infected and developed disease. We are less concerned about those than those that go undetected, and we don't know how long it takes. As a matter of fact, there is a robust amount of data coming out that there is West Nile persistence after infection for years and virus being shed in urine. It comes from a very good university, the University of Texas. The scientists doing this research are very reputable. We have to take everything with a grain of salt.

DR. KLEIMAN: I think that's true, but I think you are going to hear this afternoon, on our next topic, that not all data is replicated by multiple investigators. I think to jump on one report when other people can't duplicate it is overly cautious. That would be fine if there weren't implications.

Mike Busch and I did an analysis to say how many cases of West Nile we should have seen if -- we did see clinically transmitted transfusion cases. We should have seen quite a few more if the IgG-positive units could transmit. It's kind of a model, the kind of thing that Lyle has reported. You can never prove absence of an occurrence. But I think we had a good mathematical model to suggest that those kinds of units don't infect.

Anyway, this is dengue; it isn't West Nile. Who even knows if we can make the analogies?
But I want to support -- if you do vote for a deferral, 120 days is simply not justified by the scientific data. We don't know the impact of that yet. I would encourage shorter rather than longer, at least initially.
DR. HOLLINGER: Yes, Dr. Key?
DR. KEY: I know this is not the question before the committee, but the discussion about the lack of similarity, in some respects, between West Nile virus and dengue is a little concerning to me. My question is, what do we know about dengue in plasma pools that are made to prepare clotting factors, in terms of its susceptibility to solvent detergent inactivation and so on? I know in Brazil most of their pooled plasma was being sent for contract fractionation in France. I don't know if there is anything epidemiologically to be learned from this. But there was a concern with West Nile virus in the hemophilia community with large pools of donor plasma.

DR. HOLLINGER: These are all flaviviruses, which are envelope viruses, and so are susceptible to detergent treatment or to inactivation processes, I think, which undergo plasma fractionation. I think it's probably not an issue in that group, personally. If anybody else has any----
DR. TAYLOR: I will just make a quick comment on that. The pathogen-inactivation procedures were based on BVDV, which is a flavivirus in the same genus.
DR. HOLLINGER: Thank you. Dr. Stramer?
DR. STRAMER: I just want to comment again about the relationship to detectable dengue RNA and viremia. In the studies that I showed where we did TMA RNA testing, about 25 to 50 percent of those donations were capable of infecting mosquito cells in culture. That indicates that there is infectious virus, not just RNA. If you think about mosquito culture, we're only inoculating .1 or .5 mL relative to a full unit that would go to a patient.
So I think the 25 to 50 percent should also be considered an underestimate regarding what the infectivity of some of these units actually is -- or the potential infectivity.

DR. HOLLINGER: Sue, do you have any evidence of culture positivity in TMA-positive samples after 10 days -- 10 days since onset of illness?
DR. STRAMER: No. We have not done that.
DR. HOLLINGER: I think there have been several other studies which have done that. I can't recall any that have been after seven or eight days, I think, for culture positivity.
DR. PETERSEN: I would just like to address this long duration of potential infectivity. First, the urine studies done at the University of Texas Medical Branch for West Nile virus we have not been able to replicate. So I don't know if it's true or not.

The second thing is that when we did the West Nile virus transfusion-transmission model back in 2002, we assumed a six-day period of viremia. It turned out that that model was incredibly accurate, when screening was eventually implemented. So while these long durations of viremias may be theoretically possible, they are probably not that significant from a transfusion-transmission standpoint.
DR. NELSON: You are talking about six days after onset of symptoms?
DR. PETERSEN: Six-day total period of viremia.
DR. NELSON: Including the incubation?
DR. PETERSEN: Not including the incubation.
DR. BOWER: I just want to throw one other thing out there that we haven't talked about. We have been talking about just a deferral. I wonder if we can get comments on screening people that come back from endemic areas, if they report that they had been to an endemic area within one month -- instead of just referring them, you collect their blood and you screen them for dengue.

DR. RENTAS: Just one more comment here. My biggest concern with recommending a temporary deferral to travelers is that I just think it sends the wrong message to residents of, let's say, Puerto Rico. You pretty much have to defer everyone there or you have to test everyone there, if you do that. If you are going to defer people that have only gone there for a week or so, you are going to have to do something with the people that actually reside there. Is the blood community ready to do that? That's my biggest concern with this. I don't think we have shown that we have enough evidence out there -- and it hasn't risen to the level -- and this is my personal opinion -- that we need to test or defer people coming back from endemic regions at this point.

DR. HOLLINGER: Just a question for the blood bank. If a person comes to this country who lives in another country where there is an epidemic going on, can they donate blood here? If I was a resident of Malaysia and I came to this country and I was here for two or three weeks and, for whatever reason, I wanted to donate blood, am I eligible to donate blood?
DR. KATZ: You would be eligible if we could clear you with acceptable confidence for malaria, that you had not visited endemic areas of Malaysia for the appropriate timeframe. You could conceivably be eligible. It would be an odd circumstance.

There are other areas in southern Africa, for example, where you would be much more likely, I think, than Malaysia, for example, to qualify somebody for malaria and have them carrying dengue.
DR. HOLLINGER: Of course, that does not quite get to your question, which is a resident of a commonwealth -- Puerto Rico, for example -- who comes here.
DR. NELSON: That is an interesting dilemma, because somebody living in Puerto Rico could donate blood in Puerto Rico, but if they came to New York, they couldn't. Is that right? That's sort of strange.
DR. BUSCH: I think this ethical issue here is really important. In the few countries, such as Australia, in the northeast Queensland area, where they do, during epidemic periods, preclude donations for one month following travel from high-endemic areas, but they also don't use red cells -- they don't use fresh components from those regions -- until a month after the last clinical case, they deal with their own local population in a consistent way with respect to travelers.

The same with chikungunya. When they had the outbreak in Réunion and in Italy, not only did they stop exporting blood and doing travel deferrals, but they also imported safe blood and protected the endemic community that was being exposed with respect to the sort of indigenous transmissions. Here we have Puerto Rico.
There's no way we can support the Puerto Rican transfusion needs on a continuing basis. And that's just one small U.S. territory. What about all the rest of the world that has endemic dengue? You talk about the blood problem in these countries, with three cases in the world? They think it's ridiculous. That's not a significant contributor to dengue disease, and to put substantial resources into safeguarding the blood supply, with the limited number of transfusion cases in these regions, is just not an option.

DR. HOLLINGER: Is blood still imported into this country from Puerto Rico during epidemics? It is? It's tested with the NS1.
DR. BENJAMIN: We are importing tested units. I do want to make the other point, though, that the Red Cross, with Sue Stramer's lead, has been testing -- we have been battling to stimulate industry to make those tests available to us. We are very concerned about Puerto Rico and dengue. We would like to continue testing. I do believe it would be very helpful if this committee were to make a strong recommendation that all blood in Puerto Rico, where it is endemic, is tested. It would stimulate industry to make more tests available to us, and I do believe it would protect the blood supply in the country. We are looking for some leadership here, for Puerto Rico in particular.

DR. HOLLINGER: I want to change just a little bit and maybe deal with what an outbreak is and what an epidemic is. Can someone provide some information on what they consider an outbreak or dengue? An outbreak of smallpox would be one, but what's an outbreak of dengue, or an epidemic? Are they different? That might be one of the issues. If you're going to have a travel deferral from an area in which there is an epidemic or an outbreak, then that has to be defined. It has to be defined also here in the continental U.S. as well. Yes, Dr. Bianco?
DR. BIANCO: I don't want to define it, even with a 10-foot pole. But I would like to request that the person that defines it define the beginning and the end. I think that's very, very important. With the examples that Dr. Busch gave of chikungunya or in Queensland, in Australia, there was always a beginning and an end in an outbreak, a beginning and an end to interventions.

DR. NELSON: I think the other epidemiological variable is the age of the cases, in relation to the blood transfusion issue. It's interesting. I talked about the Derek Cummings paper that I thought was quite interesting, in which he analyzed the data from Thailand and did a modeling and found that in recent years, in the last decade, the average age of cases has been increasing. It used to be all children, but it's more adults. He related this to the birth rate. The birth rate has declined dramatically, by two- or threefold, so that now there are fewer children that are available for being carriers to transmit to mosquitoes. The average age of infection is increasing. The peaks of the outbreaks are larger, but they are further apart. It used to be every two or three years.

So I think there are more epidemiologic features than just when there is an outbreak. It's who is involved in the outbreak and what the risk would be to a traveler or potential blood donor.
DR. MARGOLIS: Let me start with the outbreak/epidemic. I think the best thing is to consider those synonyms. Frankly, you go to any textbook, and you're not going to be able to sort that out. In Puerto Rico, you can define an epidemic -- and remember that dengue is a cyclical epidemic disease. In between these epidemics, you have high endemic rates that are at different incidences, but they are there. Again, Lyle showed that in his model. The next is calling an outbreak. Yes, in Puerto Rico, where you have a decade-plus worth of data, you saw in those slides that you can set a threshold at which you are 75 percent over that incidence rate, and you can easily call it. If you go around the world, in every country, there is no standard WHO definition for epidemics, so different countries set this in different ways. Then you have to deal with that.

We at CDC aren't looking out for the rest of the world in terms of setting those criteria.
Then, as Dr. Nelson pointed out, there are these differences.
Yes, for instance, in Thailand -- again, I think Kay Tomashek pointed out that most of the studies, most of the prospective studies, in dengue have all been done in children, so highly biased. Then if you go look at nationally reported data, you get these great disparities in terms of age groups. But now there are some prospective studies, for instance, in Thailand -- it almost looks like Puerto Rico -- where almost 50 percent of cases are in adults. But if you go to, for instance, reported cases in Brazil, 65, 70 percent of all cases are in adults.

Again, part of this has to do with the endemicity of these four viruses coming at different times and establishing their infection rate. So it is very difficult.
DR. NELSON: By definition, dengue must be present. There aren't long-term carriers that we know of that would maintain an epidemic. The reservoir is humans. If dengue disappeared for three months, it would go away. In order to have the next epidemic, you have to have carriers infected year-round.
DR. NAKHASI: I just want to make sure that the questions between non-endemic and endemic or outbreak -- the reason we put it there is to distinguish between Puerto Rico, where it has been going on year after year after year, and Key West, Florida or now, if there are cases in Broward County, the Miami area. In Texas, somebody earlier was saying, once in a while there is some kind of an outbreak. That's what we want to distinguish between, outbreak area versus an endemic area.
DR. HOLLINGER: We will deal more with this, but let's -- Don, did you have anything?
DR. TRUNKEY: I have question with regard to the variability within the tests that are being done. It's somewhat disappointing that the sensitivity and specificity are not as good as they are. But Dr. Stramer, I think, told us that this was common for viral illnesses.

What we're not considering is the host response. We know from studies by Carol Miller several years ago in burn patients that the host response is quite different. There are poor responders, medium responders, and high responders. That was confirmed by Mannick (phonetic). Now that we have the Glue Grant, it turns out that this is very complex, because there are multiple genes that determine this.

If anybody in FDA is doing comparisons between the sensitivity and specificity in what the immune response is, I would have real problems, quite frankly, with putting any travel restrictions on it. You don't know. We don't know yet. But I would sure urge the FDA to try to combine some of these two data. The Glue Grant has clearly been very, very useful in that point. But if you are going to prospectively study people on the sensitivity and specificity, I would look at the host response.

DR. HOLLINGER: And I guess you would think that if there are 500 or more confirmed cases a year in this country, you might have seen some transfusion-associated cases somewhere. Somebody might have made that connection. Maybe not. Let's look at the questions --
DR. NAKHASI: A clarification here. Yes, immune response -- we are really looking at the viral load assays and antigens. We have already established that immune response will be too late after the virus. So I don't think we are even focusing on that. We are focusing on the virus detection.
DR. TRUNKEY: The problem with that is that the immune response is far more complicated. I'm not just talking about B cell response; I'm talking about T cell and, of course, the macrophage.
DR. HOLLINGER: A good point, particularly in secondary infection.
Let's look at the questions. We can still have more discussion.

Table of Contents

135

Agenda Item: Questions for the Committee

We are only going to vote on the first two questions. The third and fourth questions are for advice, comments, and so on. It's the first two questions that they have asked us to provide an electronic vote on. Does someone from the FDA want to read the questions or do you want me to read the questions?
DR. NELSON: Is the answer to question 1 yes or no? Or do we have to be specific? It says, "Are safety interventions warranted?"
DR. HOLLINGER: And maybe you could tell us what you mean by "safety interventions."
DR. NELSON: We could put an ad in the newspaper saying, "If you've got dengue, don't donate blood." That's a safety intervention.
DR. HOLLINGER: I think it is a generic question.
Anyway, let's go with the first question. Then we'll open it up for any other discussion and then we'll vote on it.
DR. TAYLOR: Are safety interventions warranted to address the risk of transfusion transmission of dengue viruses in the U.S.?
DR. BOWER: I assume that also means in territories.
DR. TAYLOR: Yes.
DR. HOLLINGER: I guess that's Samoa and -- does anybody have comments or anything further about the question that is being asked?
DR. NELSON: I would probably vote yes if I didn't have to specify what those were.
DR. KRAMER: And if you are grouping together the territories and the continental U.S., it could have different answers to the two. I don't know that a simple yes or no is --
DR. TAYLOR: Would you like to separate those two?
DR. KRAMER: Yes.

DR. TAYLOR: "In the continental U.S.," is that how you would like to phrased? Are safety interventions warranted to address the risk of transfusion transmission of dengue viruses in the continental U.S.?
DR. NAKHASI: Also I think I would add the question, what interventions? First, interventions, yes or no, and if there are -- testing or anything else -- you can give us a recommendation on that.
DR. NELSON: Is further research an intervention?
DR. NAKHASI: Intervention is not research.
DR. TAYLOR: So I think we can address that in question 3. This is a yes or no, in the continental U.S.
DR. HOLLINGER: Is there any other discussion on this question? Are safety interventions warranted to address the risk of transfusion transmission of dengue viruses in the U.S.?
Jay, do you have -- and clarify "safety interventions," would you?
DR. EPSTEIN: If the committee is more comfortable restricting it to continental, then I think we have to ask two questions. Otherwise, it becomes unclear how you are advising FDA. I would suggest that we ask question 1, A, continental, B, Puerto Rico or other endemic regions.

As far as interventions, of course we support research. I don't think that's what we are asking. We are asking whether you think that the safety concern warrants the development of an intervention. That could be deferral. It could be testing. It could be selective testing. We're not sure what it will be yet. You get to discuss that. But should we act at all? This is a threshold question. Should FDA seek to act at all?
DR. HOLLINGER: You talk about the continental U.S., but now we have Key West.
DR. KRAMER: It sounds like that would be a qualification. In discussing the interventions, it could be a targeted intervention.
DR. EPSTEIN: Again, I think the concept that we have is simple enough. We have some areas where there is clearly endemicity with cyclical outbreaks. We have other areas where there is no apparent endemicity, and it's only a question of people who have visited outbreak or endemic areas.

We understand that the epidemiology is changing and that regions that have no dengue may be developing dengue. Some of it may be incidental and some of it may become endemic. The question is really simpler than that. We know we are going to have to sort out strategies for endemic regions with periodic outbreaks versus regions that do not have endemicity. Again, it's an oversimplification to call it continental U.S. and Puerto Rico, quite frankly. What we really have in mind is non-endemic U.S./endemic U.S.
DR. HOLLINGER: Okay. Maybe that is what we should do. Are safety interventions warranted to address the risk of transfusion transmission of dengue viruses? And the A part will be, in non-endemic areas of the United States. B will be, in endemic areas of the United States.
We'll vote on A, non-endemic areas of the United States.
You see a plus, a zero, and a minus. A plus is yes and a minus is no. I'll ask everyone to place your vote.

(Vote)
DR. HOLLINGER: Has everyone voted?
LCDR EMERY: The committee has voted: 3 yeses, 0 abstentions, and 8 noes.
DR. HOLLINGER: Can we get a comment from the industry rep?
DR. BIANCO: Yes. I agree with the "no" vote. Until we have some more knowledge about the impact of such interventions, we should reserve a decision.
DR. HOLLINGER: And the consumer rep?
MS. BAKER: I voted no, based on the evidence presented today.
LCDR EMERY: Okay, I am going to call the names.
Judith Baker is a no.
Dr. Kramer is a no.
Dr. Bower is a yes.
Dr. Troxel is a no.
Dr. Nelson is a no.
Dr. Hollinger is a no.
Dr. Aldrich is a yes.
Dr. Trunkey is a no.
Dr. Demetriades is a yes.
Dr. Rentas is a no.
Dr. Key is a no.
DR. HOLLINGER: Thank you.

Now let's vote on the next part, which is the same question, but B, in endemic areas of the United States. That includes commonwealth, et cetera. I'm going to ask you to place your vote now, again yes, plus; no, minus.

(Vote)
LCDR EMERY: The committee has voted. We have 10 yeses, 0 abstentions, and 1 no.
DR. HOLLINGER: Just for the record, Dr. Bianco?
DR. BIANCO: The industry rep agrees with the yeses.
LCDR EMERY: I will read for the committee.
Judith Baker is a yes.
Dr. Kramer is a yes.
Dr. Bower is a yes.
Dr. Troxel is a yes.
Dr. Nelson is a yes.
Dr. Hollinger is a yes.
Dr. Aldrich is a yes.
Dr. Trunkey is a no.
Dr. Demetriades is a yes.
Dr. Rentas is a yes.
Dr. Key is a yes.
DR. HOLLINGER: Thank you.

Let's move to the second question.
DR. TAYLOR: Do the available scientific data support a temporary deferral for donors living in U.S. non-endemic/non-outbreak areas who have traveled to an outbreak and/or endemic area? If so, what deferral period should FDA consider?
DR. HOLLINGER: Any questions or comments? Yes, Dr. Demetriades?
DR. DEMETRIADES: Should this split again into A and B?
DR. HOLLINGER: No, because it talks about people living in the U.S. non-endemic/non-outbreak areas who have traveled to an outbreak and/or endemic area, which could be the U.S., but it could be any other part of the world.
DR. DEMETRIADES: But this it says deferral periods.
DR. TAYLOR: The first is a yes or no vote.
DR. HOLLINGER: The second part is just more for comment, by the way. It's just more for comment, Dr. Demetriades. I was just informed of that.
DR. TRUNKEY: I have a question. I can't remember who presented the data, but there were cases in Washington State. Is that correct?
DR. HOLLINGER: In Washington State?
DR. TRUNKEY: I saw on one of those maps that there -- it was a traveler.
DR. HOLLINGER: Can anyone comment about that?
DR. TOMASHEK: That was the map that had travel-associated dengue reported to Arbonet this year through mid-November. There were cases reported from -- 23 cases so far. Some of them, as you were correct in saying, were from Washington State. You remember, we had a lot of people go to Haiti and then come back with dengue.
DR. HOLLINGER: Any other comments about this question? It's an important question. It talks about the available scientific data.
(No response)
Let's go ahead and cast our vote. Yes is plus and minus is no.

(Vote)
DR. HOLLINGER: This time, what would the industry rep vote -- so you can't see what --
DR. BIANCO: You want a blinded vote. No, the industry rep does not think that there is available data sufficient to recommend a deferral at this point. That's stated in the joint statement from the blood banking organizations.
DR. HOLLINGER: Which is an answer --
DR. BIANCO: It is no.
DR. HOLLINGER: I just wanted you to say that.
LCDR EMERY: The committee voted. There are 3 yeses, there is 1 abstention, and there are seven noes.
Dr. Baker abstained.
Dr. Kramer is a no.
Dr. Bower is a yes.
Dr. Troxel is a no.
Dr. Nelson is a no.
Dr. Hollinger is a no.
Dr. Aldrich is a yes.
Dr. Trunkey is a yes.
Dr. Demetriades is a no.
Dr. Rentas is a no.
Dr. Key is a no.
DR. HOLLINGER: Now how would the industry rep like to vote?
DR. BIANCO: I will try to be consistent and say no.

Let's just discuss a little bit, if there are some questions about them, the next two questions. Let's see if there is some discussion on that.
The first one.
DR. TAYLOR: Please discuss what interventions, if any, would be appropriate for donors who are residents of endemic or outbreak regions within the U.S. and its territories.
Did you want to hear the fourth question also and comment together?
DR. HOLLINGER: Maybe we should do the fourth one first.
Any issues about this? Any comments that we can give the FDA on what interventions, if any, would be appropriate for donors?
Dr. Troxel?
DR. TROXEL: It seems to me that the options that have been presented are some kind of deferral, which we have already discussed, and/or some kind of testing procedure. I think the selective testing procedures that have been developed for West Nile are perhaps a good model for this as well, recognizing that obviously they are different subjects. But that sort of targeted approach seems like it could be useful.

Having said that, we have seen very disappointing results from the tests that have been presented. The question is whether there really are tests that have good enough performance to make that reliable and worthwhile and something that we can do. I think it's clear that the antigen-based tests are not sufficiently of high quality for that to be reliable. Then the question is, going back to the nucleic acid test, what the performance is of those.
It might be helpful to review the selective testing strategy that was developed for West Nile, in order to get a sense of what those options might be. But something of that nature, I think, is probably sensible.
DR. HOLLINGER: Hira, do you have a comment about the questions?

DR. NAKHASI: Yes. I think, since we saw that the committee didn't want to do anything for the travel, and we also heard during the discussion that there is a problem in endemic areas, like Puerto Rico or Key West, we would like to have some discussion on what you want to do there. You can't defer. You cannot do testing. At the same time, people are getting sick and dying. So I would like the committee to make a decision here, to have some kind of discussion and give us guidance as to what we should be doing in areas where this thing is going on now.
DR. HOLLINGER: Dr. Nelson?
DR. NELSON: I think it would be useful to implement the most sensitive and user-friendly, adaptable test to the donors in these areas. That's question number 4.

On question number 3, I wonder if you could stratify a risk among potential donors by, let's say, adding a question to the donor deferral questionnaire: Has anybody in your household had fever with a rash in the last month? This is a human reservoir, and the mosquito that bit your child and caused dengue could also bite you. So you might be an asymptomatic carrier.
There could be some way that we could differentiate people at high risk and those at -- you could do it with how many mosquitoes they have seen and stuff. But I would think a household illness that was dengue might put this person at much higher risk.
That's just a thought.
DR. BIANCO: Just a comment. If we could lock them at home, Dr. Nelson, it would be okay. But they are exposed everywhere to the mosquitoes. I understand the density idea, and people living in slums and all that would be more likely.

I think that a most important thing here that we heard and that you noted and others noted is that we do not have available nucleic acid tests that could be used. I would like to encourage the manufacturers sitting in room and watching this discussion to just go ahead and develop them.
PARTICIPANT: They are developed.
DR. BIANCO: They are developed. Commercialize them. I'm sure that the FDA, after this discussion, is going to be very sympathetic and license them very fast, and we'll have those tests.
DR. NELSON: The FDA is never sympathetic.
DR. HOLLINGER: Yes, Jay?
DR. EPSTEIN: Perhaps it is helpful to the committee to review what we have heard about candidate interventions. One candidate intervention is non-collection. That was the model that was practiced in La Réunion, except that they introduced pathogen reduction for platelets. They stopped collecting blood to make red cells, and they didn't use plasma. They brought those products in.

Another option, which we heard about in Australia, was to collect blood, but not to use red cells and platelets, and only to use plasma for fractionation. The implication there, though, is that you have to bring in red cells and platelets from somewhere else. I think we heard that we couldn't supply Puerto Rico on that model. So that would be problematic.
I think the third option that we heard is testing. What you heard is that current testing has a problem. The antigen test is relatively insensitive, threefold to tenfold less sensitive than TMA, individual unit NAT, but that pooled NAT is also insensitive, no better than about 50 percent, compared to individual NAT.

So that's what we are confronted with today. I think where we are looking to be advised is, where do we invest to fix this problem over time? Are we advised to test now with NS1? What you have heard is that Red Cross does this now. Of course, it's under IND. Fair enough. But is that the long-term policy. I would just say that a good argument can be made that something is better than nothing. You are picking up the high-titer viremic units. Those are probably more likely to make people sick. And that's not zero value. So one could consider it.
Most people think that in the long term you would really rather have NAT, because it has proven more sensitive. But you have to deal with the issue of pooling.
I'm not suggesting the right answer to the committee, but I think it may be helpful to consider these options, run through them, and see where we are. Again, there is non-collection, there is selective use, which also means resupply, and there is testing. And you ought to talk about antigen and NAT.

I don't know what other options there are. If other people can think of other options -- Will, we have talked about not having intervention for travelers, but you did suggest an option for the non-endemic region, which is selective testing of the traveler.
DR. BOWER: Right.
DR. EPSTEIN: There are countries that do practice that strategy for malaria. They have temporary deferral and accept donors after negative tests, provided that they have been four to six months -- it varies by country -- since their malaria exposure.
So those are the available strategies that have been contemplated.
DR. BOWER: While you're up there, we can't resupply Puerto Rico, but we could resupply Key West, likely.
DR. EPSTEIN: That is correct. I guess the historic experience there -- it was suggested to the blood center in Key West that they cease collection. They pointed out that that would, in effect, put them out of business and that they wouldn't be able to then resume. Instead, there was a cooperation reached with Red Cross to obtain NS1 testing, which was the 50 percent solution.
DR. BOWER: Even just collecting plasma for the manufacturing would still put them out of business.

DR. EPSTEIN: You know, it's very difficult for me to get involved with the business question. Perhaps others here could comment. But the feedback at the time, when we raised that issue with the blood center director in Key West, was that that was a death knell. Could they have sold plasma and survived? I don't know. Celso, you can comment. You have usually told us that the sale of plasma for fractionation is a very small proportion of the revenue of a blood collection center. I don't speak to that point as an expert.
DR. BIANCO: As I remember, it is less than 5 percent of the revenue.
DR. KRAMER: I think that you brought up an important point. The individual nucleic acid testing is by far the most sensitive. It may not be epidemiologically significant. Really, "epidemiological" isn't the right term because we're talking about transfusion. But it may be that the pool testing -- the positive you will get there is the one that matters in terms of making the recipient sick.

We don't have the data for that, I assume, and I don't know whether we have the data for West Nile for that. I know we don't want to compare them directly, but I think there are comparisons that can be made, looking at pooled versus individual. So detecting it is one thing, but then the disease is the other aspect of it.
The thing about Key West is that I'm hoping that Key West does not remain endemic, that this country can control the infection, in a limited, circumscribed, very small area, so that it's not going to be a permanent problem in the U.S.
DR. HOLLINGER: Yes, Dr. Stramer?
DR. STRAMER: Just a comment about West Nile and the triggering strategy that we use. We have had some breakthrough infections, even though we do trigger during the epidemic season for West Nile. Those units were very low in titer when titers were obtained, like .01 to .4 PFU/mL. I don't remember if that was the exact number. It has been published. So those breakthrough infections occurred in spite of mini-pool testing, not in spite of single-unit testing. Many of them could have been detected by single units, but many of them weren't tested either.
DR. HOLLINGER: Yes, Dr. Demetriades?
DR. DEMETRIADES: We are all very concerned about the low sensitivity and the great variability of the tests. Can the industry perhaps tell us if there is anything exciting, anything new in the pipeline?
DR. BIANCO: The NAT test -- Susan, do you want to talk about them? You used them.

DR. STRAMER: Well, the most exciting, certainly, is what I presented, in that it was, as Dr. Epstein mentioned and I mentioned, 3 to 10 times more sensitive than NS1 antigen. It's an automated assay. Technologists are trained on the procedures. We have the instrumentation validated in our laboratories. To add collections in endemic areas of the United States, or epidemic areas of the United States, on to what we do for NAT for HIV, HCV, HBV, and West Nile would not be a difficult task. The difficult task right now is getting the manufacturer to move forward with an IND, under some reasonable cost-recovery mechanism, so that we can initiate testing.
I think it would be a fait accompli if we had an investigational test available. We wouldn't be having these discussions; we would be using it.

DR. HOLLINGER: Anyone else, any comments about either the third question or the fourth question? Yes, Dr. Busch?
DR. BUSCH: Just two comments. One, in Puerto Rico, as you heard, the Red Cross is responsible for about half the collections. They have had the initiative, in partnership with the manufacturer of the antigen test, to implement IND testing.
As someone mentioned, the other half of Puerto Rico is hospital collections -- very small collection facilities. They are not per se for-profit, but they are managed by for-profit mediators. Our organization, Blood Systems, tests at our Tampa facility the blood that comes from there. But to actually be able to implement an IND with full informed consent and monitoring is a horrendous obstacle. I think what we need is approved capacity to implemented tests that are not under the traditional IND.

The other point about the ID-NAT -- with West Nile, we started with pooled NAT, and over time we learned and we developed a very rational, progressively enhanced, targeted ID-NAT, seasonal ID-NAT. I think with dengue, it may well be that the yield during the winters and the off-season is so small -- again, the infectivity of the pool-negative, but ID-positive -- if they are anything like West Nile, probably only a fraction of those ID-only ones are infectious.
So I think we probably want to start with ID-NAT, but I think the option of potentially moving to seasonal mini-pool NAT in a surveillance manner and turning on ID-NAT during epidemics would make a lot more sense, from a sort of cost-effective perspective.
DR. EPSTEIN: I just want to put one more concept out for discussion. Perhaps there are members of the industry in the audience who could elaborate. That is the idea of multiplexing a dengue test with a test we already do. In other words, if you could multiplex it with a West Nile test or if you could multiplex with -- you have the disadvantage that we turn it on and off with ID-NAT -- or if you could multiplex with HIV/HCV/HBV. There are obvious technological issues there. But there has been some effort to develop a multiplex flavivirus test. There would be clearly an economy, and you already have an infrastructure of testing. It's another way to go.

DR. LINNEN: I'm Jeff Linnen, from Gen-Probe in San Diego.
That's something we have looked at in the research setting, doing some feasibility for a multiplex test. But, really, to address the problem as soon as possible, it would have to be the dengue test that Dr. Stramer was talking about.
DR. STRAMER: Dr. Hollinger, may I make one more comment? I think maybe Jeff was too shy to say the hurdles. If a multiplex test for West Nile, let's say, dengue, perhaps chikungunya, or the flavivirus du jour or alphavirus du jour -- it requires a tremendous amount of revalidation. I think the FDA-licensed West Nile test would have to be revalidated from scratch. When you are combining all of these agents together, the validation hurdles are considerably more challenging. Then we may be doing dengue -- well, it's true, we would be doing dengue everywhere we do West Nile. So there is an upside to that, but there is also a downside. Then, when we had reactives, we would have to further discriminate to know they were West Nile, dengue, chikungunya -- you name the agent.

DR. HOLLINGER: I guess I would disagree that you would have to do it where you do West Nile, since they are so different in terms of pathogenicity and transmission and so on.
Jay, do you want from this committee, before we break -- are there particular other areas that we haven't discussed?
DR. EPSTEIN: I think we actually haven't heard from committee members how they might answer questions 3 and 4. We have been trying to help the committee by defining the options. But what is the thinking of the committee members?

DR. HOLLINGER: Yes, Dr. Bower?
DR. BOWER: I will go ahead and answer numbers 3 and 4, for this committee member. I think it would be appropriate to do screening for residents in endemic areas or outbreak areas within the U.S. and its territories. I would prefer that it be a NAT-based test. I don't think that anybody's, as has been pointed out, would be appropriate for all, and the antigen tests seem to miss a lot.
DR. NELSON: I don't quite understand why the NAT testing for dengue seems to be much less sensitive than for West Nile. Is it true that the cutoff for positivity is a log or two or three higher for dengue than it is for West Nile? Is that correct?
DR. HOLLINGER: I thought that Dr. Busch presented some data that suggested that the amplification tests were about as sensitive as they were for HCV. Did I read that correctly, Mike, 15 or something, you put up there on the slide?
DR. BUSCH: The analytic sensitivity of the TMA test specifically is comparable for West Nile to HIV and West Nile. That's the low-end detection. The levels of viremia in infected donors that we have documented in our studies tend to be higher. But it's sort of cross-sectional. We haven't been able to do really longitudinal follow-up of these donors and get real acute dynamics. But it looks like -- and I think it's borne out by other data the levels of viremia are perhaps peaking at a log or two logs higher in dengue than West Nile.

DR. NELSON: If that is the case, then shouldn't a multiplex assay work for dengue when it doesn't work as well for West Nile? Am I incorrect on that?
DR. HOLLINGER: Did you want to respond to that, Mike? Dr. Dahl.
DR. DAHL: Mike says I can answer for him. Roger Dahl, Red Cross.
Going back to Ken Nelson's point, I think that what Ken was referring to were data on a different PCR which was used in Sue Stramer's study as part of the confirmatory algorithm, which did appear, at least from the data, to be less sensitive than the West Nile virus, comparatively speaking. But I think that what Mike said, exactly, was that the TMA test, which is really what we're talking about right now, had similar sensitivity to and is actually based on the West Nile virus test.
I think that the points about multiplexing are probably good. I would just add that we probably don't want to be gratuitously testing 14 million donations a year for dengue at this point, given the committee's response to item number 1.
DR. HOLLINGER: Yes, I think you're right. It seems like the RT-PCR has not been as sensitive as the TMA test. I think that's sort of a given.

The question is, then, which -- obviously, those tests are, quote, more sensitive for detecting early infection in the seronegative window period than the NS1 test, I think, from the data that has been presented.
DR. BUSCH: I think that's not a fair statement, Blaine. The RT-PCR -- the quantitative assays that we developed and CDC uses to quantify viral loads are not as sensitive as the TMA NAT screening test. But Roche's RT-PCR assays have achieved comparable sensitivity to the TMA.
DR. HOLLINGER: I should have said real-time testing versus the ordinary RT-PCRs, yes.
DR. KRAMER: Given equal sensitivity, your dengue patient, probably even your asymptomatic patient, has a much higher viremia in infectious virus and copies than a West Nile patient. So even though there are equal sensitivities, you are going to pick up your dengue patients much more easily, because their titers go up -- West Nile doesn't go up above, like, 102, and dengue will go up 106.
DR. NELSON: As I understand the algorithm for West Nile, they do multiplex, but when there is evidence of an outbreak locally, then they go into individual testing. That might not be necessary for dengue. Is that right? Because in a tropical area like Puerto Rico, you have dengue year-round, really.

DR. HOLLINGER: Any other questions? Does anyone want to challenge the use of NS1 versus the amplification assays, in some respect?
We just don't have a lot of data. Unfortunately, we don't have a lot of transfusion-transmitted data to really make some honest decisions, even in Puerto Rico, where we talked about one case of dengue hemorrhagic fever. That was recognized. We have looked at all the others -- and there were several, five or nine or something like this -- that were without signs and symptoms of disease. We don't see much.
So then the issue would be, do you really need a highly sensitive test to detect it? We don't know threshold values. We know that for other viruses that there are threshold levels that have to do with transmission of disease.
So there are some issues that we just don't have the data for, to be honest.

DR. NELSON: It seems like the NS1 test is a lot simpler and might be easier to implement. I don't know what would be required, but if there was an epidemic of dengue, I would rather have blood that was tested with something. The NS1 test has good specificity. It's just that the sensitivity is poor. If you knew that it was positive, you shouldn't use the blood, even though you missed some positives.
DR. HOLLINGER: Would the committees members here feel that certainly in Puerto Rico -- I'll leave Key West out for right now -- there should be testing done of donors in that situation?
DR. RENTAS: Yes, I sure do. In fact, I completely agree with the comments by Dr. Bower as to what needs to be done in a place like Puerto Rico. I think there is enough evidence out there to start testing. As he said, I would suggest that the manufacturers work together with the FDA to potentially get an FDA-licensed test out there in the near future.

DR. HOLLINGER: Any other comments from the committee?
(No response)
Hira or Jay, anything else? Dr. Kramer?
DR. KRAMER: I guess I agree with you, but I would like to reinforce what Mike Busch said. I think it should be targeted to peak times of activity, and not all the time.
DR. RENTAS: Absolutely. I think I'm a little concerned about the 50 percent of the blood collections in Puerto Rico that are not tested at all right now. By making this a requirement from the FDA, I think it will take care of the other 50 percent that are not testing right now. That's where I was coming from.
DR. HOLLINGER: That is covered in the commonwealth. A recommendation here is also covered there, about what should be tested in the blood supply. Hira?
DR. NAKHASI: I guess what we heard is that in Puerto Rico or in other endemic areas, we need to do something, and what I heard was that testing would be one of the options. That's what we were trying to get at, because otherwise we are without any answer for those things.
DR. HOLLINGER: And we would have wasted the morning, wouldn't we?
DR. TRUNKEY: What about the travel between Haiti and, say, Miami and the travel between Cuba and Miami? Then what about the people who go back and forth between Puerto Rico and New York City? Are you going to test them as well, if they give?
PARTICIPANT: (Off-mic)
DR. HOLLINGER: That's right. It was a no.
DR. NAKHASI: For Haiti, they come under malaria deferral, so they cannot donate.
DR. TRUNKEY: I understand --
DR. NAKHASI: But for Cuba, I don't know.
DR. HOLLINGER: That is a good question, Don.
We are about 40 minutes late, and we have a long afternoon. Bring your bedding. We are going to break until 1:30. We are going to try to start around 1:30 if we can, or 1:40.
(Whereupon, at 12:40 p.m., the meeting was recessed for lunch.)

AFTERNOON SESSION

DR. HOLLINGER: We do have three new members to the committee, and I would like them to introduce themselves. Dr. Coffin, Dr. Klimas, and Dr. Glynn, if you would just introduce yourselves and where you're from.
DR. COFFIN: John Coffin, Tufts University and part-time with the NCI. My research interests are in fundamental retrovirology and retroviral pathogenesis.
DR. KLIMAS: Nancy Klimas, University of Miami. I'm a clinical immunologist and I am a clinician and have an immunology laboratory, working in this field.
DR. GLYNN: I am Simone Glynn. I am the branch chief of the Transfusion Medicine and Therapeutics Branch at NHLBI/NIH.
DR. HOLLINGER: Thank you all for coming.

Table of Contents

165

Agenda Item: Topic II: MLV-Related Human Retroviruses and Blood Safety

We have a very busy afternoon on, as said before, XMLRV and MLV-related retroviruses as they affect blood safety. We have four questions for the committee, only one of which will require a vote, which is the first one. The others are for comments and discussion only.
Let's begin the second topic of the day. Dr. Indira Hewlett is going to give us an introduction and background.
By the way, I will also mention to you that we are switching two of the sections. B and C will be switched, so we'll go to the recent studies next and then follow it with a summary of current research.

Table of Contents

166

Agenda Item: Introduction and Background

DR. HEWLETT: Thank you, Dr. Hollinger, and good afternoon, everyone.
In this afternoon's session, we will be focusing on murine leukemia viruses that have been identified in humans and their potential impact on blood safety. The issues that FDA is specifically seeking the advice of the BPAC on are donor deferral based on a history or diagnosis of chronic fatigue syndrome, CFS, or prostate cancer, with which these viruses have been associated; and testing for the newly identified MLV viruses in humans, recognizing that disease causation has not yet been established.
The MLV-related retroviruses were first identified in humans in the DNA of prostate cancer tissue using a microarray of conserved viral sequences. The sequences found in prostate cancer were clustered with the xenotropic murine leukemia viruses, and therefore were termed "xenotropic murine leukemia virus-related virus," or XMRV. Subsequently, polytropic MLVs were also identified in humans. Together, these viruses will be referred to as "MLV-related viruses," for the purposes of today's discussion.

In 2009, a U.S. study reported detection of XMRV in the blood of 67 percent of CFS patients and 3.7 percent of healthy controls using PCR. These viruses have since been detected in the blood of up to 14 percent of healthy controls. However, several studies have shown a lack of detection in CFS, prostate cancer, and blood donors. Finally, although disease association has been reported, causation has not yet been established.

In July of this year, we held an informational session of the BPAC where a number of topics were discussed, including the identification and detection, or lack of detection, in CFS, prostate cancer, and blood donors. Assay validation study plans and assay development efforts were presented. We learned about the AABB recommendation for use of donor education materials on CFS and indefinite deferral for donors who voluntarily disclose their CFS diagnosis. We also heard about Health Canada's policy of indefinite deferral for prostate cancer and recent introduction of deferral for those who voluntarily disclose their CFS diagnosis.

Since then, there have been a number of reports that have presented either positive or negative findings in regard to disease. In the case of CFS, at least three reports are worthy of mention. I will not discuss them because you will be hearing about them later in this session -- in fact, right after my talk. The most important point that should be mentioned here is that, at least in one of the studies, up to 86.5 percent positivity was observed in CFS patients. This is using PCR on PBMC DNA. The virus identified was polytropic MLV. About 6.8 percent positivity was observed in U.S. healthy volunteer blood donors.

However, in contrast, a number of negative findings have also been reported. They are summarized on this slide. There are at least two studies in the U.S. and several from Europe and one from China where, by using various methods, such as PCR, Western blot, ELISA, serology, and virus culture, no XMRV was detected in CFS cases, healthy controls, and blood donors.
In regard to prostate cancer, there are at least three reports that show positive association. Petros et al. reported at the First International Workshop on XMRV the detection of XMRV sequences by PCR in three prostate cancer tissues from three separate patients. In Northern Europe, when 355 prostate cancer tissues were tested, two out of 355 were positive for XMRV by PCR. All of the PBMCs from the 40 patients for whom PBMCs were available were negative by PCR, as also were all healthy controls.

Finally, a study that was recently published by Danielson et al. reported 22 percent positivity of XMRV sequences in 144 prostate cancer patients from the southern U.S. In this study they used PCR to look for XMRV. In the same study, both normal and tumor tissues were available for a subset of cases, and XMRV sequences could be detected in both normal and tumor tissues.
Again, in contrast to the positive findings, there are a number of negative findings that have been reported, one study in the U.S. and three studies in Europe, where they looked at a number of different prostate cancer tissues using RT-PCR, immunohistochemistry, serology, and virus culture.

In regard to transfusion, transfusion transmission of these viruses is theoretically possible due to the presence in blood cells and the evidence of cell-free virus. Also low but detectable transient viremia has been reported in rhesus macaques at 4 to 14 days and seroconversion between 11 and 14 days post-intravenous inoculation.

A seroprevalence of 1.7 percent has been reported in Japanese blood donors, .1 percent in U.S. blood donors, and by PCR, up to 6.8 percent positivity has been reported.
Again, in contrast, however, no evidence of XMRV was found in two other U.S. studies of blood donors, one of 121 blood donors where PCR and serology were used and a different U.S. of up to 1,500 blood donors that were tested using a TMA assay. No XMRV sequences were detected in 105 plasma, 19 PBMCs in Cameroonian blood donors and HIV-infected individuals in Uganda.
As you can see, there are a number of conflicting findings in regard to CSF, prostate cancer, and blood donors. The potential sources of these discrepancies could arise from a number of factors, such as the differences in study populations, geographic differences in prevalence, differences in case definition criteria and stages of illness, specimen processing, sensitivity and specificity of the test methods used in the studies, potential genetic variation of the virus and other unknown factors. Therefore, a need for assay standardization and validation was identified.

During the past year, there has been a great deal of effort in this direction. This effort has been spearheaded by the National Heart, Lung, and Blood Institute, under the leadership of Simone Glynn and Jerry Holmberg, who have created a scientific research working group composed of scientists from the government, academia, and blood organizations to validate assays for transfusion-transmission studies. The Blood Systems Research Institute is coordinating the laboratory efforts of this group by putting together analytical and clinical reference panels, to be evaluated by multiple labs, including the CDC, NIH, the Whittemore Peterson Institute, Blood Systems, and two labs at the FDA, for testing of blood donors in future studies. In addition, FDA is currently developing NAT and serology panels for future lot release, if needed. Some of you have participated in these studies.

Finally, the NIH, through the NIAID, is sponsoring a proposed study on XMRV and related viruses in chronic fatigue syndrome. This study will include a number of different laboratory sites and clinical collection sites, and will use various sample-preparation methods and assays to look for XMRV in CFS. The study is being coordinated by Ian Lipkin at Columbia University, and it has a number of experts that are inputting into the design of the study.

In regard to blood donation, no direct evidence for transfusion transmission or association with a transfusion-transmitted disease exists. Therefore, FDA has not established donor policies specific to the MLV viruses. FDA regulations require donors to be in good health at the time of donation, and medical directors exercise judgment on whether CFS patients meet those criteria.

In regard to prostate cancer, a large survey in Sweden and Denmark showed no increase in disease cases in more than 800,000 blood-transfusion recipients without a prior cancer diagnosis at the index transfusion. Indefinite deferral of donors who voluntarily disclose their CFS diagnosis is in place now in Australia and New Zealand, the U.K. and parts of Canada. In the U.S., the AABB is recommending the use of donor education materials on CFS and indefinite deferral for donors who voluntarily disclose their CFS diagnosis. With that brief background, I would like to put forward the questions that we are posing to the committee:

  • First, do the scientific data support asking donors about a medical history and/or a diagnosis of CFS as a basis for indefinite deferral?
  • Second, we would like the committee to discuss whether the scientific data support asking donors about a medical history and/or a diagnosis of prostate cancer as a basis for indefinite deferral.
  • Third, we would like your comment on the scientific evidence that would be needed to justify a policy of donor testing for infection with MLV-related human retroviruses. In particular, should donor testing be considered in the absence of confirmed disease causation?
  • Four, assuming that testing is warranted, please comment on the potential utility of NAT and/or serologic testing of blood donations to ensure safety of the blood supply from transmission of MLV-related retroviruses.

To assist in the discussion, we have put together a session which has a number of presentations on MLV/XMRV. They include:

  • A review of current data on the virus, including highlights of the First International Workshop.
  • Research updates and current data on the association of these viruses with CFS.
  • Data from animal infectivity studies.
  • An update from the Blood Scientific Research Working Group on assay validation using the MLV and XMRV panels.
  • A proposed design for prospective and retrospective donor prevalence studies that is in the planning process.
  • An update on assay-development efforts on MLV-related viruses.

I'll close with that. Thank you for your attention. I will take any questions at this time.
DR. HOLLINGER: Thank you.
I think we will to the next -- because we'll be discussing much of what you presented here, Indira.
The next section we are going to discuss is "Recent Studies of Epidemiology of MLV-Related Human Retroviruses." The first talk will be by Dr. Lo, in which he will discuss a U.S. study. I want you to be sure to pay attention to the timer.

Table of Contents

174

Agenda Item: Recent Studies of Epidemiology of
MLV-related Human Retroviruses

DR. LO: Ladies and gentleman, my name is Shyh-Ching Lo.
While we are waiting for the slides, I would just like to say, because the results of our study have already been published in a PNAS article more than two months, I'm quite sure many of you have had a chance to read it. So I'm going to just very quickly summarize our study results.
I would very much like to start by saying, why did we get into these XMRV studies? Because I transferred to FDA only about two years ago. Before I came to the FDA, I served as a staff pathologist at AFIP, the Armed Forces Institute of Pathology, for more than 23 years. During that time, in the early 1990s, a research study in my laboratory at AFIP led to the discovery and the characterization of previously unknown human mycoplasmas in patients with AIDS. Subsequently, several laboratories' studies reported that these mycoplasmas, which we called AIDS-associated mycoplasmas, were associated with the development of chronic fatigue syndrome. That was in the mid-1990s.

So blood samples from CFS patients followed by specialized centers or individual clinicians -- they sent their samples to us for the mycoplasma studies.
However, our subsequent study concluded that we did not really see any evidence of an association between this mycoplasma that we call Mycoplasma fermentans or Mycoplasma penetrans and the development of CFS.
Portions of the CFS blood samples sent to us had been maintained in frozen storage at -80. We thought this provided a very unique opportunity for us to study for the evidence, if any, of XMRV infections in CFS patients. This is obviously in the early part of this year, after the many conflicts in results had been reported of XMRV in the patients with CFS.

Overall, most of our CFS patients were sent to us by Dr. Komaroff, at Brigham & Women's Hospital, Harvard Medical School. He provided us 29 blood samples obtained from 25 CFS patients. In addition, we have 12 blood samples also sent to us from different centers and clinicians. So in this particular reported study, we studied 41 blood samples obtained from 37 patients with CFS. All these specimens had been frozen at -80 since the mid-1990s.
Our collaborator, Dr. Harvey Alter, at the Clinical Center, NIH, provided previously frozen PBMCs from 44 volunteer healthy blood donors. These samples were collected between 2003 and 2006.

The results, as we reported in PNAS: We found that a very high percentage of the patients tested positive by a nested PCR targeting the murine leukemia virus gag gene. In comparison, about three out of 44 -- so 6 to 7 percent -- of the volunteer healthy blood donors also tested positive.
This is the gel electrophoresis of the PCR amplification products. As you can see, using this primer design originally reported in a Science article -- using the primer sets, amplify 731 bp products. Most of our products come out slightly larger than the expected size of the 730 bp. These all turned out to be 736 base pairs, because we did not do this as specific deletions identifying the XMRVs.

Coupling with the second round of PCR, with 45 cycles, using the internal primer set -- this is the in-house primer set we have, NP116 and NP117 -- the product of 380 bp was amplified -- more than 80 percent of the patients, anyway.
So in the first round of PCR, we identified about 40 percent from the first 25 CFS patients, PBMC. This was provided by Dr. Komaroff.
I would also like to mention that virus RNA could also be detected by RT-PCR. However, the sensitivity apparently is lower, and we required a second round of PCR to detect them, and also fewer patients tested positive by RT-PCR. But this could also be because our current protocol of RT-PCR assay is less sensitive than our PCR assay for the PBMC DNA.
This is for comparison for 25 healthy blood donors, PBMC DNA. In the first round of 40-cycle PCR amplification, we only see one of the patients that has positive amplicons.

Coupling with a second round of PCR, the 45 cycles using internal primer PCR amplification -- these are the same patients -- only this patient could be confirmed. Also this particular patient's product could be confirmed.
As you can see, there is a mini side bands that have been amplified, and all this band was a close size to the target 413 PB. In this case, they have all been retrieved for sequencing. All the sequences turn out to be human sequences. Only this one and this one is murine leukemia virus gag gene sequence.
This is the alignment of the amplified gag gene sequences. This is from the CFS patients from type 1, type 2, type 3. Most of those sequences belong to the type 1. But as you can see, all these sequences missing this specific 15 nucleotide deletion are reported in the XMRV sequence, and the sequence alignment certainly is much more closely aligned with polytropic murine leukemia virus and has more variation with the XMRV sequence. But this is only the first 184 base-pair alignments.

Using the phylogenetic analysis, as you can see, all of the sequences we obtain from the CFS patients cannot be grouped in the same cluster as the XMRV. The cluster is actually more closely related to the polytropic virus sequence, including the blood donors' sequences.
When we translated those sequences into amino acid sequences, this is again to show all these amino acid sequences cluster better with the polytropic virus, but also closer to the XMRV. If we compare to the ecotropic virus, the murine leukemia virus, this is the virus often being used in the laboratory to study murine leukemia viruses, like Friend leukemia virus and Moloney murine leukemia virus -- very distant from the sequences we obtained.

In some of the patients -- a very few of them -- we successfully amplify the envelope gene. The envelope gene sequence, again, clusters better with the polytropic murine leukemia virus, and much more distant from the XMRV sequence. Once we had these results, we had to decide what we were going to do with the findings. Certainly we debated for a while. This is basically based on a PCR amplification assay, and in all PCR-based studies we should have a concern of contamination. In this particular case of the XMRV study and MLV-related viruses study, we have three main concerns of contamination.

The first one is, can this be a contamination by the PCR amplicons because of so many sequences being amplified in each amplification? Can that be a contamination source? Since we see more than six different sequences in this study, the variations certainly don't(?) favor that this can be coming from the PCR amplification contamination. During this study, we built in 300 to 400 negative controls, running in parallel. We have never amplified any MLV sequences in our negative controls.

Can this be a contamination by murine leukemia viruses commonly used in the laboratory? We also thought this was not quite likely because the sequences are quite distant from all those ecotropic viruses commonly studied in the laboratory. It's very different from the viral vectors normally studied in the laboratory.

The third question is the more difficult one: Can this be a contamination by mouse DNA? The mouse DNA genome contains endogenously many copies of the related retrovirus sequences. We thought this would be necessary for us to develop an assay to verify that there is no contamination of mouse DNA in the assay system we use and also in the clinical samples that test positive for MLV-like virus gene sequences. We developed a highly sensitive nucleic acid PCR assay targeting mouse-specific DNA. In this case, it's a mitochondria DNA, because we need a target gene sequence that is very well conserved among all the different species of mouse and also has multiple copies in each of the mouse cells. That's why we selected mitochondria DNA, because each cell carries 200 to close to 2,000 copies of these genes.

I'm not going to go into the way we designed it, but I just want to show you that the semi-nested PCR assay targeting the mouse-specific mitochondria DNA allows us to detect 10 femtograms of mouse DNA in the first round of the PCR, without a nesting step. This can detect 10 femtograms in the presence of the background 35 nanograms of human DNA. In parallel reaction, as you can see here, our gag gene PCR assay would be able to amplify -- very faint band in the 10th picogram of mouse DNA. So it's almost 1,000 times more sensitive than our gag gene PCR amplification to amplify the mouse DNA.

Coupling with internal nested PCR, this mitochondria-specific PCR, allows us to detect 2 to 3 femtograms of mouse DNA. In the background again is a spike into the 35 nanograms of human DNA. In this case, the internal PCR amplification of the gag gene allows us to detect about 500 femtograms of the mouse DNA.
Using this assay, which is close to 1,000 times more sensitive than the gag gene PCR assay, we come back to verify that all the DNA from those CFS patients, particularly those where we can detect murine leukemia virus-related gene sequences in the first round of PCR -- that's supposed to have higher titers of this viral gene sequence -- none of them show evidence of the mouse DNA contamination. In a parallel run, we showed the assay did detect 10 femtograms of mouse DNA in the background of 35 nanograms of human DNA in the first round of PCR. Coupling with the second round of PCR again, we can detect 2 to 3 femtograms of mouse DNA in the background of 35 nanograms of human DNA. Again, we did not detect any mouse DNA in all of the CFS patients and the healthy blood donors that tested positive for MLV-like gag gene sequence.

We also went back to check -- all of those patients' blood DNA was preserved, stored since the mid-1990s. Eight of these CFS patients' previous frozen blood samples tested positive for MLV-like gene sequence. We went back to get their blood. This was by Dr. Komaroff. Seven of these eight repeat blood samples -- this has been processed freshly and tested freshly -- also tested positive for the PCR assay for this murine leukemia-like gag gene sequence. However, the viral gene copy number in the repeat blood samples did not really increase. On the contrary, they appeared to be lower than those found in the previous blood samples of these patients. The gag gene sequence showed a quite noticeable variation in the blood obtained from most of these patients 15 years later. The details of that comparative analysis will be reported separately.

I'm not going to go into this. After our publication, there were several publications, some that could not find XMRV in chronic fatigue syndrome patients, could not find in HIV patients by some studies, as described, but found in prostate cancer patients. This depends on specific PCR conditions. I would like to conclude our study by saying that our results support the earlier finding that MLV-related virus gene sequence could be found in the blood of many patients with CFS. These viral sequences could also be detected in a small fraction of volunteer healthy blood donors. But differing from the reported finding of near genetic identity of all XMRVs in patients with CFS, in prostate cancer, and also in healthy blood donors, our analysis of the viral gene sequence revealed a more genetically diverse group of MLV-like viruses. The viral gene sequences were more closely related to the polytropic murine leukemia virus.

So our study shows that the zoonotic MLV-related viruses are infecting some human beings and that the disease association and possibility of blood transfusion by this murine leukemia virus-like retroviruses in human will require further studies. I would like to acknowledge those scientists, Dr. Li, Dr. Pripuzova, and Dr. Hung, in our laboratory doing all the work. Our collaborators, Dr. Harvey Alter and Dr. Wang, at the Clinical Center provided us those healthy blood donors' samples. Dr. Komaroff, from Harvard Medical School, provided most of the CFS patients. Dr. Puri and Dr. Witten are my superiors in my office, OCTGT. I would also like to acknowledge Dr. Epstein for his strong support of our work.
Thank you very much.
DR. HOLLINGER: Thank you, Dr. Lo.
Questions for Dr. Lo? He is going to have to leave. Yes, Ken?

DR. NELSON: Two questions. I'm interested in the blood donors who tested positive who were healthy. Was there any epidemiologic data from them that they had any association with people who had chronic fatigue syndrome, household or otherwise?
Secondly, apparently you didn't look for an immune response to the XMLV -- or maybe you did; I don't know. But I think another paper reported this. I wonder if you would comment on that.

DR. LO: Yes. For the first question -- did those healthy blood donors have any association with or were related to any CFS disease patients? -- we could not really answer that. That's just random healthy blood donor samples provided by Dr. Alter. Dr. Alter, in later discussion, I'm quite sure, can answer more about this. On the antibody response, I think there later on will be speakers who will talk about it, particularly in the healthy blood donors.
DR. COFFIN: We have had some of this discussion before, but I would like to raise some issues again, which will be familiar to you, for the sake of the rest of the committee.

First, I have to point out, of course, that unless you had something new to tell us, as yet you have no virus that goes with these sequences. Is that correct?
DR. LO: There's no virus isolation --
DR. COFFIN: You have only PCR-isolated sequences?
DR. LO: Right.
DR. COFFIN: Second, although I think I may have originally suggested mitochondrial DNA to you as a way to test for contamination, I worry very much about it, for a couple of reasons. One is that, of course, if there is contamination of DNA from some source -- you point out that there is a wide variation in the number of mitochondria per cell -- you have no idea what the cells are giving rise to that contamination, much less any idea of how many mitochondria you might have. PBMCs, for example, have relatively small amounts; muscle cells or oocytes have huge numbers of mitochondria. You have no idea.

Second is that your extraction procedures may, in fact, differentially extract mitochondrial DNA from genomic DNA.
For these reasons, my lab has developed an assay, based on another repeated element, that presents about 1,000 copies per cell that I think gives much better results and can detect mouse DNA contamination in many samples where mitochondrial DNA doesn't. In fact, using that assay, I have two collaborations ongoing which have led to papers that are going to be published by the end of this month in Retrovirology that show, in a sense, results similar to yours. In one case, two sets of samples, one from CFS patients and one from non-contemporaneous controls in fact, like you have -- showed that one set was highly positive for almost exactly the same sequences in fact, I think in some cases exactly the same sequences -- that you report and in the other case was completely negative. In that particular study, those were again taken at different times. In fact, it was the CFS patients that were largely negative and the control samples that were taken later that were largely positive. In all cases in that study and in another one in which I'm a collaborator, the samples that proved to be positive were positive for the IAP sequence control that we use for mouse DNA contamination. We attribute those results -- I think the only reasonable thing to do is to attribute those results to contamination with some laboratory reagent or another, with very minute traces of mouse DNA. One-one hundredth of a cell is enough to contain an amplifiable provirus, although your assay seems to be quite a bit less sensitive than that.

Not all of the samples were positive for mitochondrial DNA, in agreement with the idea that this is actually a less sensitive assay than IAP for mouse DNA contamination. So I worry very much about that. There are a couple of things that you could do that would help reassure me. One is -- and again, I think we have had this discussion -- the sequences that you obtain are from bulk PCR product. A hallmark of the sequences that you would be amplifying from contaminating mouse DNA -- there would be a wide variety of them, each one of which might correspond to a previously sequenced and identified provirus, within each band. You would not see that in a bulk sequence. You would just a computer-generated sort of average of what was in there, a recombinant of some kind of what was in there. Have you actually gone back and done that, cloned or done limiting dilution amplification on those to show that those are monomorphic?

DR. LO: On mitochondria DNA and IAP sequence PCR. We did compare the sensitivity of our assay. Obviously, in our hands, and I think in some of the earlier presentations from your associated group -- I think the mitochondria PCR assay is much more sensitive in detecting the mouse DNA if we run in parallel. That's why we continue to use the mitochondria DNA. In this case we did not use the high-mitochondria-containing mouse cell, like the muscle cells. We are using the spleen cells. That's the most common that people use for the laboratory study. It's hard for me to believe that we had a selective isolation of the mitochondria DNA and did not have that nuclear DNA being isolated. In fact, those kits many publications said isolation is equally efficient with the nuclear DNA and mitochondria DNA, in previously published studies.

Of course, the contamination always is a concern, like you said, because the sequence is so closely related to some of the genomic sequencing. But I would also like to say that we amplify and sequence many of those mouse DNA amplifications and we very rarely see the intact product from the PCR sequence. They always have many different interruptions of the stop codon. Of course, you can say maybe it's amplified -- for so many different sequences, they are specifically amplifying particular sequences. We cannot completely rule that out. But essentially all the sequences that were obtained from the patients have the intact reading frame.
DR. COFFIN: But there are some internal inconsistency, I think, in your results. You claimed a 1,000-fold greater sensitivity of mitochondrial DNA compared to MLV-based amplification, but yet there 100 MLV proviruses. At a 1,000-fold greater amount, that would require 100,000 mitochondrial DNA copies. I don't think anybody believes there are that many mitochondrial DNA copies per cell.
DR. LO: That's why we use the spleen cells. The spleen cells obviously have a very low concentration, a very low number of mitochondrial DNA. We did not select those, like muscle cells, that have very high copy --

DR. COFFIN: The data you showed were from the spleen cells? You were claiming 1,000-fold difference.
DR. LO: No, no, no. We are saying the reason we select the mitochondria is because mitochondria have multiple copies in each cell. That's why we use those targets. The reason we select spleen cells, of course, is because the mitochondria concentration is much lower. We use that as our assay.
DR. COFFIN: But you can still see it in 2 to 3 femtograms of spleen cell DNA, where it takes 500 femtograms to see --
DR. LO: Right, putting it into 35 nanograms of human DNA, we can detect 3 femtograms of those mouse -- most of the DNA, of course, is coming from the nuclear DNA. It's not from the mitochondria DNA.
DR. COFFIN: You didn't reply to the second half of my question, which is, have you gone back and actually looked at subclones from the PCR products that you amplified to see if those are, in fact, reasonably monomorphic as compared to what you had seen in mouse DNA?
DR. LO: I do not really understand your question. You say, if we go back to --

DR. COFFIN: When we amplified products that we attribute to contaminating mouse DNA, what we found was that in many cases the amplification gave rise to mixtures of products, and so you needed to do further dilution or subcloning of the product to actually get clean sequences. But if you just take the sequence that is fed back to you, you won't notice that until you actually subclone all of the mixtures of DNA that might be in that. Have you done that?

DR. LO: For most of the sequences, we directly read -- and when we subcloned it, of course, the majority of the sequences are exactly the same that we directly read. We do have, for example, 10 percent, 15 percent -- they will have a mutation or replacement of nucleotide here and there. Essentially, the -- DR. COFFIN: Have you done that with all of them? Whether the band is monomorphic or not will depend exactly on the quantity that you start with. DR. LO: The majority of those sequences -- the predominant sequence is the one we actually read. We also, following your suggestion, will go back to do the single-gene amplification. Essentially, the majority of the sequences are consistent with what we have.

DR. HOLLINGER: I just want to ask one question. You said you had eight donors that were positive in your assay, and then you recalled those donors --
DR. LO: No, no. Previously, in the mid-1990s, those 37 patients tested positive, and 25 of those patients originally came from Dr. Komaroff's clinic. He went back to trace eight of those patients.
DR. HOLLINGER: These did not come from blood donors?
DR. LO: These were not blood donors. These are the patients with CFS.
DR. HOLLINGER: These were not blood donors that you looked at later.
DR. LO: That's correct.
DR. HOLLINGER: So they have not been looked at over a period of time.
DR. LO: Right.

DR. HOLLINGER: Dr. Klimas.
DR. KLIMAS: Two questions. The eight follow-up subjects, do you know if they were more or less ill than they were the first time?
DR. LO: Actually, I think Dr. Komaroff should be the one to answer this question. I asked him, and he said those patients are still sick. There is a variation among those patients, but most of those patients continue to be sick over this period of time.
DR. KLIMAS: The samples that were in the freezer all these many years, had they been manipulated -- frozen/thawed or in any way --
DR. LO: No. Regional samples in this study have been kept frozen for all this period of time. They have never been thawed out to use for different studies.
DR. GLYNN: I just have a question for John, if you don't mind. If there was contamination, how do you explain the differential between patients with chronic fatigue syndrome and the normal donors?
DR. COFFIN: I would have a lot of trouble explaining that if the samples have been taken contemporaneously and handled identically and blinded. In the case of the studies that I am involved in as well -- and it's the simplest thing to do, starting out -- the samples from the patients and from the donors were taken completely separately, at different times, different places.

I'll share with you my nightmare. That is that somewhere there is a warehouse that has a big pile of sodium phosphate sitting there. At night, mice come in and crawl over this and do whatever mice do. Then during the day somebody comes in with a shovel and grabs a few kilos of this, and that ends up being you reagent-grade PBS. That sometimes happens and sometimes doesn't. The sensitivity of these assays is good -- a 100th of a cell is enough to -- this is the kind of thing we have to be extremely vigilant about in these kinds of assays, where you are just taking the sequence, where you don't detect a virus. I'm not saying that Dr. Lo is wrong on this. I'm just saying that the standard of proof has to be extraordinarily high.
DR. LO: I totally agree.
DR. COFFIN: I'm not completely convinced that the mitochondrial DNA is --
DR. LO: I totally agree. In this kind of assay, obviously we have to be extremely careful. I just want to say, yes, we did pay attention to that. Every time we run all these assays, we put it in enough negative results and run it together. We have to say, in our system, we have never seen that.
 

DR. HOLLINGER: Dr. Nakhasi.
DR. NAKHASI: In the second group of patients that you recently got from Dr. Komaroff, seven or eight people, did you attempt to isolate the virus? I know you told us earlier that you could not from the frozen samples. Was there any attempt to isolate the virus from these samples?
DR. LO: We did that. We obviously tried very hard to isolate the virus, because that will be the key question. When we have the samples, we set up all kinds of assays and tried to isolate virus.
At the present time, I have to say that we have not successfully isolated virus. We did see some infections in the signal, documented by PCR, but we could not have the virus isolated.
DR. HOLLINGER: One comment, and then we're going to move on.

DR. ALTER: I just want to perhaps clarify something. In terms of contamination, Shyh has covered most of it. To me, it was very important to get these eight samples from 15 years ago. These were drawn in the clinic, sent directly to Dr. Lo, and analyzed immediately. If a contamination was coming from a reagent, these were samples drawn differently, the sequences had changed over time, and the controls were again negative. It just doesn't speak to me of a contamination. It's not just the original two samples, the donors coming from one place and the patients coming from another place. This is a third sampling now from the same patients, drawn and tested immediately, leaving less room for contamination. You can't rule it out.
But I also wanted to ask Shyh, didn't you use Dr. Coffin's method on these samples as well?
DR. LO: What do you mean, Dr. Coffin's method?
DR. ALTER: For the IPA for the mouse contamination.

DR. LO: Yes. We checked the IPA. As I described, in our hands, the sensitivity of the IAP assay to amplify lots of a side band from the human sequence and the sensitivity when we compare with the mitochondria PCR the same mitochondria PCR for the same set of human DNA -- with a spike in mouse DNA, we certainly see that the mitochondria PCR assay, in our hands, is much more sensitive. That's all we can say.
DR. ALTER: But did you test our samples by that method?
DR. LO: Right. That sensitivity was much lower, and we could not really document the IAP sequence there.
DR. HOLLINGER: Thank you, Dr. Lo.

Table of Contents

197

Agenda Item: U.S. Study

We will move on to the next U.S. study, Maureen Hanson, from Cornell.
DR. HANSON: Thank you.
I am going to be calling this study the 20/20 study. It's a small study, funded by an exploratory grant from the NIH, an R21. This study is entirely unpublished. It's still in progress. We haven't been able to complete all the tests at this point.

I would like to start with acknowledging my collaborators. In my own lab, I have a postdoc, Li Ling Lee, and a part-time technician on this project. We have been assisted by David Ruppert, a statistician. The clinician who has provided the samples is David Bell. He has been assisted by his son, David Bell.
Forty subjects were selected by Dr. Bell. Dr. Bell presided over one of the outbreaks of CFS that occurred in the mid-1980s in his area of western New York. We have 10 what we call severe CFS, 10 recovered CFS, and 20 healthy controls. The Bells administered a number of survey instruments to test the health status of these healthy controls and the severe and recovered CFS. The samples were received at Cornell. We were blinded to the health status of the individuals providing the samples. I had to be unblinded in September in order to present -- in August, actually -- to present the talk at the XMRV meeting at NIH. But my lab members are still blinded because the study is not yet finished.

I would like to describe the very severe CFS. These people are very ill. They are housebound. They are often bedbound. They have less than three hours daily of upright activity. The recovered CFS all met the CDC definition of chronic fatigue syndrome at one time, but they all consider themselves either recovered or nearly recovered. They have many more hours of upright activity, 13.5. Six of these 10 were part of the outbreak that occurred in Lyndonville, New York. Of the healthy controls, all of them live in the same geographic area, western New York. They were screened so that they had never lived with a person with CFS, fibromyalgia, or prostate cancer. However, some of these people are close friends of CFS patients. These people have 15.5 hours of upright activity, and they appear to be completely healthy.

If we graph the scores on the survey instruments, you can see that the blue here, the severe CFS patients, have significantly different scores than the healthy controls. But what's interesting is that the recovered patients, although they consider themselves recovered and perfectly healthy, actually have some lower scores on many of these instruments, significantly lower on the SF-36 test.
These people do feel well. They feel well enough to donate blood, and a number of them, in fact, have done so.
We are still working out our assays for looking for the MLV-like viruses. We have a number of assays that we have tried, and I'm going to mention these. But we're still not sure which are the best ones to use. We have results from all the different assays, but we are still, as I said, in progress.

We have made PBMCs from blood collected in EDTA and made nucleic acids from these PMBCs that are immediately frozen or cultured for 5 to 10 days. We make RNA with TRIzol and make cDNA. We make DNA with CTAB. Then we do PCR on these samples. We observe DNA bands on the gel. We never just conclude from the size of the band that we have a gag sequence. We always sequence it before we conclude that we have a positive gag sequence.
We have also recently been doing PCR assays on whole blood. We are using the Qiagen kit -- again, making DNA, doing PCR, observing the bands in the gel, and sequencing.
We have also been looking for virus in plasma by incubating it with the prostate cancer cell line LNCaP cells. Again, we make RNA, convert it to cDNA, make genomic DNA, and do PCR.

We do take precautions against contamination. We have been using mouse mitochondrial DNA controls, using cox2 primers. We are actually thinking of switching to the ones that Lo described, because he has validated how sensitive those are. We also do all of our work in labs that have never had any sort of mouse cell line or mouse work done in the lab. In fact, we put our hood into a plant growth room. All of our blood is collected in this former plant growth room. We UV-irradiate all of our tubes before we use them. We carry out our first reaction PCR in one of these hoods that has a UV light in it to clean up any DNA that may be in there. We have a second hood for our second PCR. This is all in a lab that's separate from my lab, where we work with the amplicons.

This is just an example of some of our results. This is actually genomic DNA from these LNCaP cells, incubated with plasma and then assayed after four transfers. We have not had very good luck in any of our assays with just a single round of PCR. Usually you get a gel like this, which is blank. But in second-round PCR, we have been able to detect gag sequences. This is after six transfers. These are different samples. They show again that we can detect gag when we sequence this. We often, like Dr. Lo described, get nonspecific human DNA bands of different sizes. Whenever we sequence a band that is the wrong size, it turns out to be some accidental amplification of human sequences.

Our sequences that we have obtained are very much like the ones described by Dr. Lo and the Alter group. We, in fact, have one of their blood donors up here. In comparison to XMRV, our sequences look a lot more like polytropic MLV. One thing in particular is this deletion right here in XMRV in comparison to polytropic viruses. We don't have that deletion, nor does the Lo, Alter paper either. This deletion was first described in the prostate cancer study, present in the glyco-gag region of XMRV.

This is a summary of our current data. We have 7 out of 10 severe CFS patients, 7 out of 10 recovered CFS patients, and 4 out of 20 healthy controls. These are only the numbers where we have done two different tests and gotten positives. We actually have some additional samples that are positive, but we only did it once, and so I'm not reporting that here at this talk. You notice that we have rather a high healthy control number here. It's my suspicion that this may reflect the fact that the samples were taken from an outbreak area, and although these people weren't living with people with CFS, as I said, many of them were friends and associates of people with CFS.

Why are there so many reports of failure to detect these viruses? One thing I would like to point out is that some of the negative reports -- the people used primers that span this deletion. A primer that spans this deletion is not going to amplify the virus that either Dr. Lo reported or that we are reporting here. This primer will not work. But it was used in at least two of the negative studies. The other possibility is that all of the labs with negative reports optimized their PCR conditions for detecting VP62 XMRV or XMRV from a culture and they didn't optimize for any MLV-like virus that might be present. In my lab we haven't had any XMRV in the lab. We have just optimized according to what's in the samples. We have tried lots of different conditions to figure out good conditions to use to look for virus in the samples, not using XMRV as our control.

For example, just to illustrate what different results you can get with different conditions, here's a case where we have four samples, and these two samples are negative under these conditions and these two are positive. But using these conditions, we get all four of these samples positive -- these, incidentally, being the conditions that were used in the Lombardi et al. paper. The other thing that we have noticed is that really very small differences in your methods can make a huge difference in the results. We have three different brands of PCR machines in our lab. We have seen differences in which PCR machine is used, in our own lab, as to what results we get as far as detecting things or not detecting things.

I would like to end by presenting some results that were not -- these data were not obtained in my lab. These were obtained by Rachel Bagni's group at NIH Frederick and at the WPI. Samples were sent from Dr. Bell to these groups to look for antibodies in the sera of these same people. Of the seven severe CFS samples in which we detected virus, at NIH six were detected as having at least one XMRV-reacting antibody. The WPI detected the seventh. Clearly the NIH and the WPI tests are a bit different. You are going to be hearing about what those tests are later on in this session. Of the seven recovered CFS, NIH detected four of them as having an antibody to at least one of these antigens, and WPI detected the other three. Of the four controls -- this was quite reassuring to us -- in which we detected virus, NIH detected three as having one antigen and WPI also detected one of those three.

I would just like to mention that, of course, antibodies are less specific than PCR, and an antibody that reacts to XMRV should also be likely to be reacting with an MLV-like virus. Our conclusions are:

  • We have recovered patients who feel well, but they are actually significantly different from healthy controls.
  • We have gag sequences similar to polytropic MLV.
  • We have been able to infect LNCaP cells with patient plasma and get gag sequences after four or six transfers.
  • Our virus sequences are highly similar to those reported in the PNAS paper by Lo et al.

Thank you.
DR. HOLLINGER: Thank, Dr. Hanson.
Questions for Dr. Hanson from the committee? Dr. Coffin?
DR. COFFIN: You do not have a virus that goes with these sequences at this point. Is that correct?
DR. HANSON: We only have gag sequences at this point. We don't yet have the virus.

 

DR. COFFIN: I will ask another one of the same questions I asked him. The sequences that you report, are those based on bulk sequencing of bands or are they based on cloning or limiting dilution of PCR?
DR. HANSON: These are based on bulk sequencing. However, we don't merely just take the sequence that the facility gives you. In addition to getting the actual text sequence, we also get the actual traces. We know from our other projects in the lab -- my lab works on RNA editing as well -- that if there is about 15 percent of a different sequence, we can see it just in the bulk sequences. So in some of our bulk sequences, we can see some single-nucleotide polymorphisms. Sometimes we can see a G in the provirus and see a G and an A in the cDNA. We can see that. I think if we had more than 15 percent of a different sequence, we would actually be able to see it in the bulk sequences.
DR. COFFIN: But you do see some.
DR. HANSON: We do see some, but single-nucleotide polymorphisms, not --
 

We are at the limits of detection. We will have a sample positive one time and then negative another and then positive the third time. The PCR is really tricky. As I said, sometimes when we got inconsistent results, we discovered that one person was using one PCR machine and another person was using the other. So you really are at the limits of detection with the PCR. That's why we want to get at least two positive results before we count a sample as positive.
DR. HOLLINGER: So basically they were tested multiple times?
DR. HANSON: Yes.
DR. HOLLINGER: Dr. Nelson?
DR. NELSON: I was just thinking of the causative criteria, the Koch hypothesis, et cetera. You had a group who had severe disease and another one who had recovered and then another one that was healthy. I guess this is totally qualitative data that you have. But I would just wonder, is there any way to quantitate whether or not the signal you get is different. Does that correlate with the then, obviously, the other criterion is temporality. In order to establish a cause, the exposure has to come before the outcome. I just wonder if there is any way to get at that, in terms of the age of the patients or when symptoms occurred or the age of the controls versus the cases. Can you see any way to look at these two criteria namely, temporality, which I think is the most important, and a quantitative relationship?

DR. HANSON: First of all, as far as characterizing the patients, we were quantitative in giving them survey instruments, so we do know their health status. Many of these patients have
DR. NELSON: But what about the virus?
DR. HANSON: Many of these patients have been ill for 25 years. Many of even the recovered people got sick in that 1984-to-1987 outbreak and then they recovered, most of them five or six years later. As far as the quantitative PCR assay, we are not there yet. I know there are groups, represented here, trying to develop a quantitative PCR assay. We don't have a quantitative assay for the virus at this point.
DR. KLIMAS: Congratulations on such a nice study. It's really well done.
Your question is a very good one. The idea of using this epidemic group and comparing it to a cross-section of everything else that is out there would be very interesting. I think that would be good. I think it's very interesting that these so-called recovered patients are equal to the ill patients in the numbers you are calling positive.

At the XMRV meeting, I came away thinking that the LNCaP cell line, the prostate cancer cell line, wasn't as good for MLV as it was for XMRV. Did I hear that wrong?
DR. HANSON: That is the conventional wisdom. We have actually never seen what is called classic XMRV. We have only gotten these polytropic-like sequences from the LNCaP cells. I can't explain that at this point, but, in fact, what we are seeing is polytropic-like virus in the LNCaP cells.
Let me also say one thing about epidemic versus non-epidemic. Not all of these individuals were people from the outbreak. There were people who you would call sporadic cases that appeared since then. There are some more recent people mixed in with the severe and the recovered. It wasn't just the 1985 outbreak.
DR. KLIMAS: Did they scatter into the positives?
DR. HANSON: With the small numbers that we have, I don't think we would have any statistical power to tell you that.
DR. HOLLINGER: Thank you very much.

Table of Contents

210

Agenda Item: U.K. study

We'll move to the third talk in this section. It's a U.K. study. Dr. Mikovits will present this.
DR. MIKOVITS: Thank you. Today I have been asked to talk about a study we presented at the XMRV workshop for a group of patients in the U.K. Since the first isolation of XMRV from the blood of CFS patients, my collaborators and I at the NCI, SAIC, and Whittemore Peterson Institute have been working to develop more sensitive assays for detection of infectious virus. The rationale for these studies is that in our work over the past year we have developed more sensitive methods for both the biological and molecular amplification of human MLV-related viruses, which in this talk will be called HMRV, in the blood and plasma. We have developed these technologies. These methods were used to determine the incidence of HMRVs in a U.K. cohort of ME/CFS. Importantly, this cohort was diagnosed using the more rigorous Canadian Consensus Criteria. Those are the criteria that Tony Komaroff and the study you heard from Dr. Lo, David Bell, as well as Dan Peterson from the original Science paper, used throughout their patient populations. It's important, because of the heterogeneity of the disease that we heard from Dr. Hewlett earlier, that the cases were similar.

Let's talk a little bit more about the study cohort we used from the London area of the U.K. They had gotten diagnosis of what they call myalgic encephalomyelitis. Or, often, post-viral fatigue is how it's diagnosed in the U.K. primarily. All of these patients do meet the Canadian Consensus Criteria of CFS that are used in those studies that we talked about today and in our original study and throughout the WPI. The disease duration in this patient population was 9 to 26 years, with greater than 50 percent of the patients actually housebound, and many bedbound. The onset of disease could often occur in childhood or puberty. We won't go into the possible reasons for that today. In addition to that profound post-malaise/fatigue that is really the sine qua non diagnosis of CFS that meets the Canadian Consensus Criteria, the other symptoms included severe cognitive dysfunction, multi-joint pain, the onsets of new and frequent migraine headaches, vertigo, dizziness, lymphadenopathy, profound mitochondrial dysfunction, which might explain the energy. Many of these patients have GI disturbance and dysbiosis, an inability to absorb nutrients, and medications as well. They have chronic infections. As we have often heard, well, those CFS patients have everything. Yes, they do, and that's the point. A healthy immune system doesn't have chronic EBV or chronic HHV-6. We see shingles in a 30-year-old often in these patient populations.

Importantly, many report a flu-like onset. They knew the day they got sick, and didn't recover. The current age of the study participants that were used was 19 to 70. Interestingly, in the 50 that were done randomly in this first pilot study, equal numbers were male and female. The study design we used is similar to that which we used in other studies done at the WPI. We have the blood drawn by Phlebotomy Services International, which is an independent certified phlebotomist group that goes around the world. PSI codes and ships those samples. In this case they were shipped to the NCI, where they were processed in a laboratory that had no previous XMRV work nor any previous murine research -- a human lab that had not done XMRV research previously. The plasma and the PBMCs were isolated two days after the blood collection. That was largely a matter of the shipping of the samples from the U.K. All the samples were tested in two independent labs, blinded. We blinded in 50 healthy controls taken from blood donors by our collaborator, Jonathan Kerr in London, in the mid-2000s, 2005 to 2008. We didn't have fresh draws from those blood donors, but these were blinded into the study, as those were the controls that we had available.

The samples were tested for the four methods. I'll go through it very carefully:

  • For plasma XMRV RNA.
  • For cell-free transmission from the plasma to the LNCaP cells, which we have heard about earlier. I'll describe that assay in detail.
  • We looked for plasma antibodies to HMRV viral proteins.
  • We Western-confirmed the positive cases from those transmission studies. Finally, we did sequence characterization of the HMRV isolates.

First, we'll talk about the plasma PCR. We had never before done direct plasma PCR. We had been working with the Blood Working Group and thought that perhaps delayed processing, which had been done in other studies, might increase our ability to see viral RNA which may have been associated with other blood components and actually released into the plasma. When we did this plasma RNA from 140 µL of the plasma from 48 percent or 24 percent of the patients, we could see in this top, using the Lombardi nested primers and conditions, a very strong band for the gag. These were all sequence-confirmed to be gag of XMRV, but this amplicon wouldn't distinguish the polytropic sequences. It was very small.

Secondly, we used the Lo primers, went back to these samples that had not been previously frozen and thawed, because we aliquot them into .5 mL aliquots. You can see that two patients who were negative using our PCR were positive using the PCR protocol of Dr. Lo, suggesting that that also sequenced. But we didn't see it as polytropic. Maybe that's just our phylogenetic analysis. But all were confirmed by sequences and again highlight that subtle differences in PCR protocols can give you really big differences in results, as you would have found far fewer of these patients positive by the Lo protocol. Importantly -- and, unfortunately, it doesn't show well here -- we used 5 femtograms of DNA from the murine cell line that we have in the lab. We could see no mitochondria-specific amplicon, as Dr. Lo described.

The other thing that we do in all of our studies you see here number 2767 -- those are patients from the original WPI Science study, where we consistently and over time -- over three years' time -- can both detect plasma viremia and isolate from that patient. We carry this sample throughout these studies. We do that with several samples in every study. The control samples, as I mentioned, that we blinded in from Dr. Kerr -- you can see that very few actually had XMRV or HMRV RNA in the plasma. But importantly, two out of 50 that were reportedly from the healthy blood donors -- was 4 percent of the population there in the U.K. I do want to remark that of the two negative studies that had come out at this time from the U.K., they had absolutely zero incidence in controls or patient population of XMRV.

Since we are not a PCR lab and neither we nor the Science paper nor our work focuses on PCR, we went to doing the culture techniques that you have heard about today. I'll describe them in a bit more detail to show the isolation and characterization -- that these were indeed representing infectious virus. In the assay that we used in the Lombardi study, shown in the top line here, we take plasma or activated -- this is dividing peripheral blood and mononuclear cells -- from the patients, and we co-culture them on the prostate cancer cell line, which was responsive to androgens and inflammatory cytokines. This is important because we know we have characterized the LTR, in Steve Goff's lab and Bob Silverman's lab, and we know that there are hormone-responsive elements there that would be an on switch to make the virus replicate more in the cells that were responsive to androgens.

We culture these for 21 to 42 days. Dr. Hanson mentioned four passages. This is a lot of cultures. Carefully looking at other negative studies, they might culture them for a week. We follow them, in addition, to PCR by Western blots from antibodies. These are monoclonal antibodies which were described in the original Science paper. This rat monoclonal to the envelope of this spleen focus-forming virus, which is a polytropic, xenotropic virus -- importantly, this antibody was characterized by Sandy Ruscetti all the way back in 1982. But this surface unit I show you here in the Western for the transmission of three of these U.K. patients -- this antibody recognizes all polytropic, xenotropic, and ecotropic viruses. This antibody -- and that may be why our numbers were so high, because our original paper didn't originally rely on just the PCR, when, in fact, this antibody could detect all of the viruses. I'll refer you to that paper. If you look at Figure 1, you will see PCR-negative patients who clearly we could culture virus in, detect it from the antibody, and sequence whole virus.

This is the assay, which is quite labor-intensive and cumbersome. You know that it took us quite a while to do these 50-odd samples and 50 controls. We have been developing -- and you heard this from Dr. Le Grice at the last BPAC meeting in July, so I won't go into detail -- an assay in which an MLV vector has an inactivated green fluorescent protein in it. That vector is packaged by either XMRV or any MLV-related virus. You then infect those cells with that virus, if it's in the plasma, and in only 4 to 18 days, you can see green cells, representing infectious virus. In order for this vector to go from inactive to activated, it needs both reverse transcriptase and integrase. So it's important that this assay is an assay for infectious virus.

We show here that you can also quantitate it by flow cytometry and clearly see and count the green cells. Hopefully, that has been speeding things up a lot. When we use it in the U.K. samples, here is a positive control. Only 11 percent are positive. But that is due to viral interference and other things about this assay. But clearly a negative and clearly a positive. Both of these samples, if you go back and look at that first figure, were plasma PCR-positive, suggesting, but not proving, that it's infectious virus. Now we can see that 78 percent, 39 out of the 50, were positive in this infectious assay. This is just showing you other numbers that were negative and positive in the same assay, as you can see here. When we confirmed all of the samples -- and I show you here only the positive -- we confirmed by Western analysis, using an anti-MVL envelope. This is a xeno, so it is not the monoclonal I showed you, and then a Gag antibody as well. You can see that we can detect both Gag proteins and envelope proteins in these Western analyses, confirming that we had, in fact, transmitted the virus from the plasma of these patients to the LNCaP -- we call these DERSE cells.

Importantly, I show you that 2767 positive control that we carry throughout these studies. We next amplified a wider range, shown in the box on the top, of the envelope. When you look at small amounts of envelope, maybe due to the diversity that is wider than we originally anticipated -- when we actually did a PCR in the pol-pro region, extending down 600 base pairs of product into envelope, and then we sequenced -- I show you here, representative, three of these U.K. samples we could see that they were indeed more similar to XMRV than to the polytropic viruses we have been hearing about this morning.

Interestingly, this patient, U.K. 1023, was negative in all of the other assays, but we could actually, from the LNCaP, which was a DERSE assay -- it was only 2 percent, but we could actually see by Western that there was indeed virus there, and we could clone it out. We are doing full-length sequencing of as many of these viruses as we can from single cells at this time.

We next talk about the serology assay in these patients. This is the assay that you heard from Maureen Hanson and that was described in detail in the Science paper. We use a cell line that expresses the murine spleen focus-forming virus envelope, that same region that I showed you. The antibody recognized all known polytropic, xenotropic, and ecotropic viruses. We take a plasma from a patient sample and incubate it with the non-expressing cell line. You see nothing in this histogram, which shows increasing fluorescence and density of the binding of a secondary fluorescently labeled IgG. But here, with the patient sample on the envelope-expressing cell line -- clearly suggested that there is antibody to the envelope in the patient's plasma. You can compete that using that monoclonal antibody. If we co-culture the monoclonal antibody with the patient sample, you see that you can compete either at 1-to-10 or 1-to-50, demonstrating the specificity of this antibody and that indeed the patient samples do contain antibodies to spleen focus-forming virus envelope.

This is how the shift looks. We did this in all of the patient samples and controls. When we compare the detection of antibody reactivity with virus isolation from the plasma, you can see a concordance there in essentially all of them. There were only five samples where we could isolate virus, but could not detect presence of an antibody -- we don't know why that is -- and a few samples where we could detect antibody and not actually isolate virus. In summary, then:

  • We could detect gag in the RNA in the plasma in 58 percent of the 50 patients.
  • We could transmit 78 percent of the patient samples to the LNCaP cells.
  • We see antibody reactivity in 68 percent of those 50 patients.
  • We could sequence the envelope products, showing that the predominant HMRV in this U.K. cohort is indeed XMRV.

We conclude that multiple methods are necessary to detect evidence of XMRV infection. In this case, in a very well-defined cohort, similar to the positive studies, we could detect it in greater than 70 percent. With that, I'll thank my collaborators and funding and you for your attention. DR. HOLLINGER: Thank you.

Questions?
DR. COFFIN: The virus is growing out in the DERSE cells. How much of that have you sequenced?
DR. MIKOVITS: We have sequenced large parts of the envelope and the gag. We --
DR. COFFIN: I'm just curious, because one expectation in those cells is that what might happen -- the virus that grows out may not actually be the virus that you originally started with.
DR. MIKOVITS: That's why we -- DR. COFFIN: It could be a recombinant that has picked up useful sequences from the vector, like the LTR. It would be nice to see if that's happening.
DR. MIKOVITS: Indeed, and that is why we also run those LNCaPs without that vector. We have been sequencing the virus out of the LNCaPs where I showed you the Westerns. We run both assays, because we recognize that that might happen. It would be interesting if it did indeed happen.
DR. RUSCETTI: Can I follow up on the Coffin question?
DR. HOLLINGER: Could you give your name?
DR. RUSCETTI: Ruscetti, NCI.
We have done LNCaP and the DERSE cell on several isolates and found no difference, at least in the envelope and gag regions.
DR. COFFIN: I was thinking particularly about the LTR, which might well have been exchanged.

DR. RUSCETTI: We are just beginning to work to look at the LTR. We have been pushed into it by Jonathan, who asked us at every meeting to look at it. So we are now looking at it. But we don't have any results for it.
DR. MIKOVITS: And it could well be a key to the reservoir. The LTR is really a key, maybe, why we can't grow this virus in these cells. We are looking at other cell types right now.
DR. COFFIN: That harks back, in a sense, to the other talks. To my knowledge, nobody has ever grown a virus that has a polytropic or a modified polytropic-type LTR in it, unless somebody in the room has done that recently and I haven't heard about it. DR. MIKOVITS: Usually these are seen with xenotropic, actually mobilizing the polytropic. The polytropic indeed becomes the pathogenesis. But they need the xenotropic to be --
DR. COFFIN: In mice that seems to be the case.

DR. NELSON: These are pretty convincing data, to me, that this infection is real. You showed antibodies. One issue that remains is, is this the cause or is this the result? It's quite possible that they have something else - the immune system crashes or whatever -- and they get infected with whatever is around, a mouse or whatever. Have these patients been studied for other chronic viral infections, HHV-6, HTLV-2, or EBV? Have these patients had an extensive virologic/infectious disease workup?

DR. MIKOVITS: These patients in the U.K. have not. It is a psychosomatic disease in the U.K., and they can't get those types of medical treatments easily and maintain their benefits. In our study in Science, the answer is yes. These patients have multiple chronic active infections -- EBV, HHV-6, CMV, as I mentioned, shingles. We see everything -- mycoplasma. It looks to us like an AIDS patient, with an obvious hypothesis being that the retrovirus causes the underlying immune deficiency. But it alone can't cause the disease. It needs the co-pathogens. You can have HIV without having AIDS, but you can't AIDS without having HIV and one of 25-odd co-pathogens.
DR. NELSON: So it looks like the next step is to there are several repositories, and this disease may be frequent enough that one could identify an infection with this agent. You have 4 percent of blood donors or whatever. These people need to be followed. These people need to be followed to answer the question: Does this infection occur before the chronic fatigue syndrome or afterwards? If it occurs afterwards, then it's just a passenger; it isn't the case. I think that's a critical question.

There are some repositories, the NHANES, Washington County, Maryland, where there are large numbers of samples that are stored and frozen, to go back and look at incident disease after that. It seems like this should be of some priority at this point.
DR. MIKOVITS: It is a critical question, but we can't do it with the existing assays. Part of the reason we were developing these, hopefully, more high-throughput assays is so that we could do the large-scale epidemiological studies necessary. I agree completely.
DR. HOLLINGER: Dr. Coffin, one more.
DR. COFFIN: In that same vein, it's also not out of the question that this virus infects very large numbers of people and remains in some very-difficult-to-detect form, as it does, apparently, in macaques after they are infected, as we'll see, and that there is something about a condition like chronic fatigue that allows it to appear and replicate in that fraction of people, despite the fact that it's in almost everybody to begin with, or in a very large fraction of people to begin with
DR. MIKOVITS: If it were, we might find it a little more easily than we do.

DR. COFFIN: We might, but it's not so easy to find Epstein-Barr virus if you don't have an antibody response either in infected people. It's very, very infrequent in latent infection, but it's there. It might be that there is -- I think it's a remote possibility, but a possibility that in some people there is some hidden reservoir somewhere that only becomes visible due to some condition associated -- immune deficiency or whatever -- associated with chronic fatigue.
DR. HOLLINGER: Thank you, Dr. Mikovits.
Now we'll go back to the one we were going to start with. Dr. Jonathan Stoye is going to talk to us about a "Summary of the Current Research on MLV-Related Human Retroviruses and Disease Association."

Table of Contents

226

Agenda Item: Summary of Current Research on MLV-
Related Human Retroviruses and Disease Association

DR. STOYE: I thought I was giving the introduction. I'm not. Some of my slides may be repetitious. I apologize for that, but there's not much I could do about it.
It is, actually, rather better than something that happened one time I was in the States before, where I had flown over for a day to give a talk on a certain topic and found myself pushed into a discussion session. It got later and later and later in the day. It was about 6:00 or 7:00 at night, U.S. time, and it was well past midnight my time. This time at least I'm awake.

So I'm going to give you a brief introduction. I want to say that I am the head of the Division of Virology at the MRC's National Institute for Medical Research. I have worked with murine leukemia viruses for longer than I care to remember, certainly more than half my life. What I was going to do was summarize recent data on the detection of murine leukemia viruses in human samples and describe their association with disease. Then I was going to consider and present my perspective on the data that I have described. I think this is probably still relevant to you, even though you will have heard at least some of it before.

This field started in 2006, with a paper from Bob Silverman and his collaborators, who isolated a virus they called XMRV from prostate cancers. They found it in around 10 percent of prostate cancer tumors, associated with a deficiency in an immune defense gene. Since it was similar by sequence, they called it XMRV.

This stimulated a number of different studies that I'll tell you about in a few minutes, but first I want to say a little bit about some of these strange words -- "gammaretroviruses," "xenotropic" -- just to make certain you have been listening. I will test you afterwards to see whether you know what they mean.

This slide shows a tree of retroviral phylogeny. There are seven genera of retroviruses. You are familiar with HIV, which falls in the lentivirus genus, or HTLV-1, which is in the deltaretrovirus genus, and XMRV falls, with the murine leukemia viruses, within the gammaretroviruses.

These gammaretroviruses are simple retroviruses. They don't have the accessory genes that are present in some of the viruses, like HIV or HTLV-1. They are widespread in nature. They have been isolated from many mammals and koalas. Infection is lifelong, with persistence of infected cells established very soon after infection. They have been associated with a large number of diseases, frequently with cancers.

These viruses can give rise to what are called endogenous retroviruses. These are retroviruses that could infect the germline and are present in the germline of essentially all mammalian species. Different endogenous retroviruses are present in different numbers within these groups, and it's perhaps not well appreciated that, in fact, we have more viral sequence in us than we have coding genes. So we are more virus than human.

The first recognizable endogenous retroviruses were inserted more than 50 million years ago. So back in the age of dinosaurs, retroviruses were first floating around the world infecting our precursors. These viruses have, on the whole, suffered from various different inactivating mutations that don't give rise to virus. Basically, we have lived in a sea of retroviruses for more than 50 million years, and there has been constant conflict between retroviruses and us.

We have developed various ways of trying to protect against these things. Very few of them, in fact, can give rise to infectious virus. So the defense mechanisms have succeeded in inactivating these long enough for random mutations to have occurred. They can give rise to infectious virus occasionally or can contribute sequences to replicating viruses to change their properties.

Xenotropic viruses were originally defined on their ability to replicate only in species other than mice, due to polymorphisms in the receptor gene. This idea may, in fact, have to change slightly. Recently it has been shown that certain mice do contain functional receptors for xenotropic virus. But it's likely that this name, as a label for a class of sequences inherited as endogenous forms, will persist.

Some of these proviruses -- for instance, Bxv1 -- are intact and infectious and frequently have been shown to infect human tumor cell lines passaged through nude mice. None of these xenotropic viruses is directly pathogenic in mice, but often recombine and contribute to the generation of oncogenic recombinant viruses. They can cause a wide variety of diseases.

The pathogenicity in other species is unknown. And they are closely related to XMRV, but none is identical.

Let me start the meat of this talk by presenting the association between these viruses and prostate cancer. There have been 11 published, peer-reviewed studies dealing with different cohorts of patients that have been examined for the presence of XMRV. There seem to be basically some groups that find around about 20 to 25 percent positive for XMRV. They constitute about a third of these. In the rest of them, there are zero or between zero and 1 percent detection rates. So you have these two schools: You either find it or you don't find it. That seems to be true for CFS in much the same way as it's true for prostate cancer.

Some studies show an association with RNase L, but, in fact, it doesn't look as though that association is holding up. In one study, with a relatively small set of samples, they were analyzed by three different techniques simultaneously. Five of these seven were positive by all three techniques. By contrast, two were negative by all three.

However, in the largest study showing an association between XMRV and prostate cancer, a number of samples were positive by one technique, but not by another technique. This is very worrying. One has to be concerned about the techniques that are being used to detect the virus. There are now half a dozen different techniques that are being used to detect virus. If you get a result with one but not another, one does have to worry.

This just compares one recent negative study, one recent positive study. Both used very similar techniques for detecting virus. One used a qPCR for envelope, the other a nested PCR for envelope. They worked with similar DNA. They have similar sensitivities. But one had zero out of 161 positive for XMRV; the other, 32 out of 144. Interestingly, the group that found positives, if they dropped the amount of input DNA for their PCR reactions, also detected a much lower frequency of DNAs, suggesting very strongly that it's very tough to detect this virus and you have to push your PCRs in order to find XMRV.

Which brings me to one of the questions I would ask: If the limit of detection is around one copy per 1,000 cells, one has to come up with plausible disease mechanisms if one is to invoke these as causative agents for disease.

The first of the two studies on the previous slide, the one by Aloia et al., also used immunohistochemistry to try and detect the virus. They looked at nearly 600 samples. All were negative. They also state, but do not document, that they obtained negative results with sections from specimens found to be positive in the previous study.

Again, were enough cells studied? If the methods used here were correct, then there must be something wrong with the other study.

A key point in this saga was the paper that came out from Lombardi and various people you have already heard from on the detection of a virus, XMRV, in 67 percent of patients with chronic fatigue syndrome and in 4 percent of controls. This virus was found by PCR, serology, and virus isolation, and appeared to be robust.

If this figure of 4 percent of control donors is correct, then we clearly have a problem, because this is higher than the number of HIV-infected individuals in the world. And if this virus can cause different pathologies, it's a real problem.

However, various people have tried to replicate that study. And here I have to confess to a conflict of interest. I'm an author on one of these papers. Nobody, other than that first paper, with the exception of the Lo paper, which I'll come to in a moment, has been able to replicate Lombardi. So neither the initial association with CFS or controls has been substantiated.

I'll mention two other groups of studies before I go back to CFS. There are a couple of positive papers -- well, one is a paper and one is a patent -- claiming that XMRV is present in the respiratory tract of immunocompetent and immunocompromised individuals. Similarly, immunohistochemical studies of breast tumors suggested that a fraction of these also contained XMRV.

Unfortunately, these studies remain to be confirmed by other techniques. It will be interesting to see what comes out. These are other searches for XMRV in groups of individuals where you might have risk of transmission through blood or sexually transmitted. They include several groups of HIV-positive individuals, a group of Hep C individuals, and also, interestingly, given claims of an association between XMRV and autism, there is a negative autism study. In this case some of the patients were autistic children of CFS-positive mothers.

In all cases the assays look pretty good here. I'm surprised that they are zero, if virus is really present in the number of individuals for which it has been claimed.

Let me go to the Lo paper. As you have heard, they detected MLV-related sequences in 32 out of 37 patients and 3 out of 48 normals. These were most similar, by sequence identity, not to XMRV, but other endogenous proviruses. And I must actually insist, since I coined this phrase, that what they are most similar to are modified polytropic viruses, not polytropic viruses.

They concluded that "our results clearly support the central argument by Lombardi et al. that MLV-related viruses are associated with CFS and are present in some blood donors."
I have concerns about that statement, for three independent reasons.

The first is that there is no virus isolation or evidence for viral insertion into human DNA. The second is -- well, let me talk about the first one first. Let me describe the case of a virus that was at one point called HRV5, human retrovirus 5.

It was originally cloned by PCR techniques for a patient with Sjögren's syndrome. It was related to betaretroviruses of rodents. It was found in multiple labs, in various different conditions, and it was not found in controls. Independently, several different labs discovered this same virus. Some patients had serum antibodies that reacted with HRV5 gag, but no virus was identified. When, finally, integration sites were cloned from clinical samples, they were shown to be rabbit, and not human. It's still not clear how multiple labs detected a rabbit endogenous virus in human tumor samples, but not in controls. But that's what happened.

To date, no infectious replication-competent viruses which are nonrecombinant, belonging to the endogenous PMV class, have been isolated. We need urgently to see a more detailed characterization of the viruses described by Lo et al. In particular, I would like to see those eight new sequences which were reported two months ago, but we still haven't seen details of.

Let me talk about the question of variants. It has been suggested that xenotropic and polytropic viruses are simply variants, that they are more closely related than certain isolates of HIV-1. However, I would argue that that is not the case, for a couple of reasons.

Firstly, the four classes of endogenous virus that have been described in the mice have fairly defined and fixed properties, including a number of restriction sites and insertions shown here, but even at the nucleotide level. You can tell pretty much which one is which just by looking at a few hundred nucleotides of sequence. You don't see intermediates between these. They seem to be fairly fixed.

Certainly -- this is data from the recent review by Christine Kozak -- they evolved in different parts of the world, apparently to form a fairly stable form of virus. So xenotropic viruses and polytropic viruses came from different parts of the world.

I'm not convinced that these can be thought of as parts of a continuum. Rather, they are distinct viruses. Yes, they can recombine with one another, but they are not part of a continuum. So I would say these viruses are therefore unlikely to represent variants of XMRV.

If their results are correct, then it would imply that you have independent viruses, with independent transmission into the human population.

Lastly, I want to talk about a comparison of the sequence methodology used. This is a bit difficult. I have tried to do this a number of times with papers from a number of authors. Different primers have been used in PCR techniques, and different DNAs have been used. But there is a very small subset of the Lombardi paper in which they used the same first-round primers as the Lo paper -- so these 419F/1154R. This, remarkably, had a 50 percent incidence of positive results. This, I believe, was only 7 percent. This virus here was xenotropic virus; this virus here was a polytropic virus. Interestingly, the primers used here -- this primer has mismatches with polytropic viruses. It's identical to XMRV. So if these samples had XMRV, one would have expected that this would amplify XMRV and not PMV.

So I have great difficulty in explaining how this could be. Based on these results, it is improbable -- highly improbable, in my opinion -- that these samples have XMRV. That result one can take further by considering the second-round nested PCRs, but I haven't got time to do that.

I would also compare the Lo data with that from another paper, by Henrich et al., which has just appeared. As far as I can make out, the technologies were similar, the primers were similar, the sensitivities were similar, the patients came from the same part of the world. Yet in one case you have negative, in another case you have positive data, both in the control and the CFS patients.

You have seen a version of this slide already. There have been a number of explanations suggested for these discrepant results -- differences in diagnostic criteria, differences in the geographic distribution of XMRV, PMV, and all the other ones, how the samples are processed, what sample was actually taken, the possibility of mutations introduced by APOBEC or other agents. But I don't think they cover all the cases all the time.

I think one gets very close to this problem, and the word that everyone wants to avoid using, which is "contamination." I might, in this context, change the word "improbable" to "unthinkable." But, nevertheless, I'm going to think it because of this problem.

I would stress that working at the limits of detection presents a significant risk of having problems. There are at least these five potential problems, any one of which might give you false positives.

The first is that you fail to sequence the products. Often you see human bands that contaminate gels. I'm not saying that these things have occurred multiple times, but I know of at least one occasion in the last year when each one of these has happened.

You can have amplification of endogenous retroviruses from contaminated mouse DNAs. You can have amplification of plasmid DNA or vector controls. You can have amplification of viral nucleic acids from "outer space." That means things that come up positive once, but not again. If you get a virus like XMRV, which will go through LNCaP cells like a rocket, I don't believe that this kind of thing can be -- that such sporadic isolation can be explained. I would argue very strongly that having positive controls, such as virus or plasmid, actually increases these risks. Unless you use a separate lab for each experiment -- you're going to run out of rooms in your building. I did wonder once whether you could just go down the corridor.

So I think there are real problems there. I'm going to conclude with three quick conclusion slides. There are various things we need, which include validated samples, better reagents, better case definition, and the results from the blinded split samples that I hope we'll learn about in the next few hours.

There are lots of things that I think we still don't know. We don't know where this virus came from, whether it crossed once or more. If the results are all correct, then it has to have multiple cross-species transmissions. We don't know about roles of disease. We don't know of incidence, distribution, mode of transmission.

So my take on all this is as follows. I don't think anyone would disagree with this first one: These levels are extremely low and hard to detect. I would argue that the samples studied by Lo et al. did not contain XMRV, and thus do not corroborate the data of Lombardi et al. There will be those who argue with that conclusion. However, if one wants to reconcile them, there have to be different transmission events, and if that is the case, then there have to be multiple origins for chronic fatigue syndrome. In other words, the samples that Lo et al. looked at have to be different from the ones from the WPI.

I think it remains unclear whether the XMRV presents any problem to the blood supply. I must say that the study that I was part of and other studies I know of in the U.K. have not seen the same results that we heard reported earlier.

I think I'll stop at that point.
DR. HOLLINGER: Thank you, Dr. Stoye.
Questions? Dr. Klimas?
DR. KLIMAS: I am left with just these big puzzles. One is, why are the controls negative in the positive studies? The other is, why is there antibody? It's not just a question of virus, contaminated virus, and so on.

The third is, my experience as an HIV doc for an awful lot of years -- I remember going to a conference where Dr. Ho announced that one in 10,000 cells were positive for HIV. The next year, it was one in 1,000, and the next year, it was one in 100. Finally, it got down to one in 10, over the course of just four years of improving assays and improving method.

So here we are with these assays that everyone agrees are just dreadful. Yet you can say with confidence how much virus is there? It doesn't make any sense.

DR. STOYE: I think that these techniques are ultrasensitive, these PCR techniques. If there was one copy for 1,000 cells, you would find it -- not one in 100, not one in 10, but one in 1,000. You would find it. So that's the first thing.

The second thing is, precedent says there have been instances in the past where control populations were negative and others were positive. We don't know why that is. But that has happened before. So there is precedent for it. Certainly if the samples are not completely intermingled with one another, one does worry very much about somebody knowing what a positive is and what a negative is.

That does bring up a practical problem, in that I say one shouldn't have known positives in among negatives. I don't know what one does about that.

What was the -- oh, the sera. I'm not a serologist, but my understanding is that you can sometimes get cross-reactivity. There have been cases where the antibody was directed against something different than you thought it was, even though it reacts with it. It has been known on more than one occasion.

DR. NELSON: You showed a comparison of a couple of studies that used the same primers, the same methods. Have any of these tested the same specimens?
DR. STOYE: I think we will hear about that in the study from the Blood Working Group.
DR. NELSON: It could be that XMRV causes something one place and HHA-6 causes it somewhere else.
DR. STOYE: I cannot rule that out. But I think we will hear in the next talk or two about studies on the same set of samples.
DR. COFFIN: I would like to reemphasize a point that Jonathan touched on, and that is that our current information on the association of XMRV with disease is based on about five different things, including detection by FISH, for example, of small numbers of infected cells in prostate cancer, including isolation of XMRV from both prostate cancer and chronic fatigue syndrome, including detection by PCR of MLV-like sequences, including some serology and so on, and the detection by immunohistochemistry of probable viral antigens in prostate cancer.

I think it's very important at this point not to conflate those observations together. I often hear people talk about these as though they were all tips of the same iceberg. I think there is no reason to make that assumption right now. I think we have to consider each of these observations separately and independently.

As Jonathan eloquently pointed out, I don't think we can consider, for example, that the Lo et al. study confirms or refutes or really impacts the conclusions from the Lombardi study, because we are looking at two different things. Each of these has its own possible sources of possible problems, such as contamination, cross-reactivity, and things of that sort. While I would very much like to have seen this subject go forward to understanding what the pathogenesis of the virus might be, we are still hung up, and we're hung up because we don't yet in any of these cases have the sort of convincing data that would reassure us that our concerns about cross-reactivity, contamination, and so on aren't true.

For example, if XMRV samples showed significant genetic variation within a patient and significant variation in the way that HIV does from patient to patient, we wouldn't be having this conversation. There wouldn't be an issue here. If there was a complete infectious virus that went with these MLV-like sequences, our conversation would be very different. But in none of these cases have there been -- although none of the experiments that have been done conclusively show that these, in fact, are due to these various kinds of artifactual situations, there is a lack of experimental evidence in any of these cases that reassures us that they are not. That's what is bothersome here. We need to keep moving forward and try to get evidence one way or another. But we are certainly not at that point yet.

DR. STOYE: If I may just add one thing to that comment, there were 17 sequences that were described of envelope sequences in the Danielson paper. There were about 10 nucleotide differences in total between those 653 times 17 sequences. Now, by my calculations, assuming an error frequency per replication cycle at 10-5, that means that all those 17 cases had to be linked by less than a total of 100 replication cycles with one virus. It's, to me, implausible to see how that can occur.
DR. COFFIN: But not quite impossible, unfortunately.
DR. STOYE: By no means impossible, but implausible.
DR. RENTAS: You mentioned at the end that one of the possible issues is the lack of validated positive samples. How would you define that?
DR. STOYE: That is a very good question. You have to have a sample from a patient that multiple individuals can receive and all agree that this sample is undoubtedly positive. I don't think it's there yet.
DR. HOLLINGER: Thank you, Dr. Stoye.
Let's move on. The next one is going to be about "Animal Studies: Potential Transfusion Transmission of MLV-Related Human Retroviruses," by Dr. Villinger

Table of Contents

245

Agenda Item: Animal Studies: Potential
Transfusion Transmission of MLV-related Human Retroviruses

DR. VILLINGER: Thank you very much. Good afternoon. It's a tough act to follow, to say the least. My cohort is probably the only one that is smaller than Dr. Hanson's. But the price of monkeys is so high, on a weight basis.

Thank you again, for being here to share our findings in a nonhuman primate model, which, as you all realize, is fairly related to humans on the evolutionary scale.

I think we have heard enough of the background of the XMRV. It was discovered by Dr. Silverman.

What is important for us is really to keep an open mind. As I said, an etiological link has yet to be really established between XMRV and prostate carcinoma or chronic fatigue syndrome. However, I think one thing to keep in mind is that a number of these patients may suffer some level of immune impairment, and that may create a milieu that allows the virus to replicate. So I really think that in order to figure out what happens early on, since we see these patients relatively late in infection, an animal is where you can get some answers.

What we were trying to do was document whether the virus is able to replicate in a monkey model. Then, if it induced an active infection, what are the viral kinetics and behavior in vivo?

The primary question we had was, first of all, does it replicate in vitro in monkey cells? Number two, is there any preexisting immunity in rhesus macaques?

This was done right here, where we tried to replicate the virus in a primary fibroblast line that was developed in my lab. You can see that not only did the virus replicate, but we also got protein. This is an example of the animals that we have been using in the study. But we tested about 25 from the Yerkes cohort, and basically we didn't see any evidence of preexisting immunity.

We enrolled five rhesus macaques. Initially we started with three, two males, of course, since we were interested in the prostate, and for the sake of balance, we had one female. We basically tried to stack the cards in our favor. We gave them a relatively high dose of XMRV intravenously, followed the animals, doing a number of different collections. Around five months post-infection, we sacrificed one animal and tried to reactivate the virus infection by reinfecting the two remaining animals IV with purified virus, followed them, and then ultimately, to boost the titers of antibody, we vaccinated them with recombinant XMRV proteins in incomplete Freund's adjuvant. We necropsied the animals two weeks later. Then we went back, took two animals, and, once we figured out where the acute infection was, we sacrificed two animals during the acute infection stage.

This is the list of techniques we used to follow the animals. I will not show all the data, for the sake of time.

These are the viral loads that we observed in these animals. Basically, monkeys different just like humans, and we found three different viral load patterns. One animal had a relatively rapid peak of viremia, which was below 10,000 viral copies -- so fairly low -- and then was undetectable at day 14. The female animal had a delayed viremia, a lower one as well. In the third animal we basically never saw any viremia, plasma viral RNA in the blood. However, when we looked at PBMCs, you can see that we were detecting proviral DNA in all three animals for about a month post-infection. Reisolation of full-length RNA from these animals at these time points, day 18 and day 21 -- we were able to get full-length RNA, which was sequenced.

Something that was brought up by Dr. Stoye in the previous talk is that there are a number of host restriction factors that will work on XMRV. It's a simple virus. It doesn't have all the accessory genes to counter these natural antiviral mechanisms.

This work was done in Bob Silverman's lab, looking at this one animal at day 21, analyzing about 1,200 bases. You can see that you have extensive mutations that occur, G-to-A hypermutation, which is probably the work of APOBEC3G, which is fairly active. That goes along with some of the questions that were brought about how wide you would find the same sequences maintained, especially in the blood.

The second monkey that we were looking at -- this is a female animal at day 18 -- you can see also that there is extensive mutation going on.

This may not be an entirely kosher comparison, but this was human PBMCs tested in vitro, where you also see a lot of mutations going on. But you do at least have some intact sequences that you see here. Again, it's a different part of the genome. I'm not trying to make a strong comparison.

If you remember, we went back sacrificed one animal, reinfected two at this time, and then followed the viral loads.

Basically we were not able to detect any plasma viral RNA at this stage. However, by proviral DNA, you get these intermittent signals that we saw for about a month. This was the same high dose of sucrose-banded virus that was given intravenously.

If you remember, at the late stage, we immunized the animals. This is one of the two that we sacrificed at nine months post-infection. Interestingly enough, in this animal we found again about 2,000 viral copies, which was completely unexpected. That suggested that the immunization and the strong adjuvant was able to actually reactivate replication-competent virus in there. I think it's fairly clear that the virus infection is established and chronic in these animals.

The next thing that we wanted to look at was what the target cells are for the virus, in the blood at least. We took the cells collected during the initial infection, days 3, 5, and 7, pooled them and sorted them in CD4/CD8 T cells, monocytes, macrophages, B cells, as well as NK cells. To make a long story short, we see that the virus is mostly in lymphocytes, T and NK cells particularly. Interestingly enough, except for the one duplicate here that was positive, we didn't see it in monocyte macrophages, which was a bit of a surprise.

In terms of phenotypic analysis, we were following these animals in the blood for different parameters, multicolor flow. As you can see, something that struck us in the beginning, as the animals -- we saw the B cells sort of spike after the initial infection, NK cells also. If you look at a proliferation marker, Ki-67, there was clearly some immune activation going on, at least in these subsets.

After the reinfection, though, although the B cells didn't move so much in terms of percent, the NK cells did, and the proliferation markers on these cells were really very marked. This is higher than what you would see in HIV infection. The T cells did show up, if you look at total T cells. When we drilled down to memory T cells, we found the same level of activation. I won't show the data.

In terms of humoral responses -- this work was done in collaboration with Abbott -- basically, we found relatively rapid humoral response to the virus. The first one was gp70, which is sort of obscured in that blot. This was redone down here with recombinant surface protein. We found responses at day 9 post-infection, followed by the transmembrane protein p15E, and then, finally, Gag responses thereafter.

Abbott has done a tremendous amount of work -- and you will hear more about it from John Hackett later on developing a high-throughput assay based on the Architect platform. This is based on microparticles bearing the different antigens that are tested both in an indirect format, where you detect the bound antibodies, or a direct format that detects both IgG and IgM.

Both formats were able to detect the response. This is just the p15E response on each of the three monkeys. The solid line is the direct one, the intermittent one is the indirect, which is shown after the initial infection. Obviously, you can see the peak of IgM early on.

If you look at the three animals kinetically, to the different proteins that were tested -- this is the p15E you can see that after the initial infection, you get a nice response. It gets boosted in the two animals that were reinfected and then comes right back down and comes back up after the immunization.

But what I would like to draw your attention to is the fact that these antibody titers come back down significantly, suggesting that, similar to other infections that don't replicate a whole lot, there doesn't seem to be a lot of antigen available to the immune system to respond.

Being pathologists, it was important to find the virus somewhere. Since we are working with monkeys, we can look at them, sacrifice them at different time points, and look for virus.

One place we focused our attention to first was lymphoid organs. We are looking at spleen, both in acute infection and chronic infection. You can see these isolated cells that were positive for XRMV gag. The same thing in the GI mucosa. This was confirmed by FISH assay against the entire genome.

The other question that we were after is, what are these cells, really? We went to looking at the jejunum as one of the examples, but it's similar in all the lymphoid organs that we looked at. We did staining for anti-CD3 T cell, XMRV.

Basically, the only cells that we see that are positive are all T cells.

The other question is -- as you all know, T cells come in many different flavors. We tested for CD4 primarily and showed again that most of the cells that did stain for XMRV were CD4 T cells.

So the virus seemed to be lymphotropic, even in lymphoid organs.
How about non-lymphoid organs? As aging males -- we have quite a few in this audience -- we're always interested in prostate.

The one thing that was really flabbergasting to us -- this is relatively low-power magnification of the prostate during the acute infection -- you can see that you have these foci of infection, which was mind-boggling, given the difficulty we had in finding the virus in the blood of these animals. There is no lack of virus elsewhere. Obviously, there was a fair amount of replication going on, or at least infection, and gag production in that organ.

At a higher magnification, you can see that all the cells we found there were these acinar epithelial cells lining the tubuli there.

In the second animal, also at the acute stage, you can see exactly the same thing. You have these acinar cells that stain.

At the high magnification, again you have these cells lining the acini of the prostate that were positive, and some of these there, which are probably just cut tangentially.

Moving on to the chronic phase, though, to our surprise, the prostate was pretty much negative by in situ histochemistry. That was unexpected.

If you look at the other male animal done at nine months, the same thing. We didn't see any signal by IHC.
However, if you drill down, not looking for protein, but looking for the viral -- really, the FISH technique that we have detects XMRV RNA -- you can see that we find cells that are still positive there. So it's not like the virus was cleared out of the prostate, but it clearly is limited. We don't seem to have any protein production. Which cells are positive -- it's very difficult to localize the type of cells by FISH.

Besides prostate, of course, a few more organs are interesting. This is the acute infection, again the lymphoid organs. We found some in the pancreas. The lung has these alveolar macrophages that are positive. In testes, we found a lot of cells, especially during the acute infection, and in the chronic infection.

Other organs, like liver and kidney, we couldn't do by IHC because of the background. So we confirmed them by FISH. There is clearly signal there.

Finally, the female animals. We should be equal rights, equal opportunities. As you see, there is a lot of virus still present in the GI. But when we looked at the cervix, you had these isolated cells that came up positive, as well as in the vaginal wall.

Cervix was potentially one organ where we saw two different types of cells being infected. This is a magnification. We might find some epithelial cells, as well as some interstitial cells there being positive, which is sort of interesting.

The other one that we looked at where we saw also a lot of cells that were infected, and some of them that might be of a different lineage, is lung, where you find some that look like epithelial cells and then you have your alveolar macrophages, which are fairly positive over time.

This is just a summary table of the detection by immunohistochemistry. These are the animals at different stages post-infection that were sacrificed. You can see that in the lymphoid organs, we find virus throughout. By that time, it was very difficult to find in the blood, except for the animal that was reactivated. The GI mucosa, of course, was fairly positive. Then we had a whole series of tissues where we couldn't find it, but then the reproductive organs, as you can see, were positive, both in male and female animals, which tends to suggest maybe some hint as to how that virus may be transmitted.

In conclusion, I hope I have been able to convince you that rhesus macaques are susceptible to infection. It is a chronic, persistent infection that can be reactivated, given the proper context. But again, we don't see very much of the virus in the blood past the acute infection or some stage of reactivation.

Again, the PBMCs or the lymphoid organs that are primarily of the lymphocytic lineage are being infected. The tremendous infection in the prostate during the acute infection is something that we are following at this point also.

The other thing that we are looking at -- and we're really sort of zeroing in on some of the reasons for that -- is this transient activation of memory T cells, B cells, and NK cells, which might be another way by which -- even if the virus is not oncogenic per se, if you have constant viral replication, you may have chronic immune activation, which by itself may cause oncogenesis.

But again, in the monkeys, otherwise XMRV is clinically silent. I can vouch that they were not fatigued during the entire experiment.

So where are we going with that? One question we would like to address is, does the virus cross the mucosa to induce infection? Can it be transmitted as a model of sexual transmission? What happens in the prostate? Why is the virus controlled and not replicating constantly?

I have a "Draft" sign because this is data we are working on at this point. I got some of the data a couple of days ago. We went back, with a small grant from the Geyer Foundation, and got four macaques that we infected. We just deposited the virus into the urethra, right where you have the prostate canaliculi, and then followed the animals. Since we didn't see a whole lot, we infected back the rhesus macaques into the prostate.

I will just show you the antibody responses in these two animals. You can see that in one of the animals here, after day 50, the antibody to gp70 really shot through the roof. This was before the animal was reinfected into the prostate, suggesting that, clearly, the virus had crossed over and was able to infect the mucosa. This was the mucosa. We see that also for the other proteins.

The other thing is that there was a clear delay by the time we could find antibodies.

My last slide is just to acknowledge the people who have done the work, as well as the organizations that have funded the study. Thank you. DR. HOLLINGER: Thank you, Dr. Villinger. Questions? Yes, Dr. Nelson.

DR. NELSON: Two things. One, did you study any central nervous tissue? If there is a problem with activity, fatigue, et cetera, that would be a tissue that might be of interest.

Second, did anybody modify the activity of these animals, how long they slept or how much they ran around? You showed us tissues, but you didn't show us what they did after they got exposed.

DR. VILLINGER: The answer to your first question is yes. I didn't put that on. You can find an occasional FISH-positive cell in the CNS of these monkeys, but no protein. In my take, this is really not a very good milieu for the virus to replicate, for whatever reason. The monkeys are fully immunocompetent.

So your second question, during the day we follow the animals. They are certainly as active as the rest of the members in that room. At night, generally monkeys are even more active, but we don't have cameras in these rooms.

DR. NELSON: One of the last papers had how many hours the people slept when they had severe fatigue. That's one thing somebody could probably monitor. Do you know how long these animals sleep?

DR. VILLINGER: They don't sleep more than -- well, it's hard to say. The animal handlers come in in the morning. There was no evidence that these animals were -- what you have with monkeys, especially the ones that have SIV late-stage, is that they become sluggish. You come in the room; they stop jumping around. These were no different, by any means. I really don't think that it had anything to do -- they don't lose weight or whatever.
DR. HOLLINGER: Dr. Hanson?
DR. HANSON: The question has been raised here about how much variation occurs in XMRV after infection. I was wondering if you have any sequence to compare to the XMRV that you used to infect these macaques. In other words, in the prostate afterwards, is the XMRV sequence a lot different than what you started with or not?

DR. VILLINGER: That is work we still have to do. I agree. That's a very good point. We have shown that in the blood it is quite different. But again, monkey APOBEC's regime may be more active than human APOBEC's regime. That's maybe one of the limitations of our model.
DR. KLIMAS: The degree of immune activation -- chronic fatigue patients have a very marked level of T cell and immune activation, CD4 and CD8. You said you saw more immune activation. How much more?
DR. VILLINGER: I don't know if you saw the level. It jumps from maybe 2 percent to anywhere from 10 to 20 percent of CD4 and CD8 memory T cells. Certainly in the NK cells, which are also lymphocytes, you see marked activation. That's during the time when you do see evidence of virus in the blood. Yes, we believe there is a link there.

I think if we could induce the virus to replicate on a more chronic basis -- that's one of the next experiments that we are looking at.
DR. HOLLINGER: Thank you, Dr. Villinger.
DR. KLIMAS: The point is just that you have a very low viral load, and still a marked immune activation. There is more persistence --
DR. VILLINGER: But very transient.
DR. KLIMAS: So the activation fell.
DR. VILLINGER: Yes. You just have this peak.
DR. HOLLINGER: Okay, thank you.
We have three more talks in this session. We are going to do one more and then we'll take a break and do the last two afterwards and then go into the open public hearing.

The next talk is by Graham Simmons, and he is going to give us an update of Blood XMRV Working Group activities.

Table of Contents

261

Agenda Item: Update of Blood XMRV Working Group Activities

DR. SIMMONS: The first few slides I'm going to present are very much a summary of what I presented at the last BPAC, so I'm going to try and go pretty quickly through those.
This is the panel of the research working group, including the five or six laboratories which are actively involved in the studies.

The mission of the group is really to design and coordinate research studies to evaluate whether XMRV poses a threat to blood safety, initially trying to evaluate some of the existing assays present in the participating laboratories, and then going on to establish a preliminary prevalence of XMRV in blood donors, and then hopefully determine whether XMRV is transfusion-transmitted.

We split our initial studies into four distinct phases, starting, as I said, with evaluating the performance of the existing nucleic acid test assays in the participating laboratories.

To do this, we made a number of analytical panels by serially diluting chronically infected cells or supernatant from these cells into either whole blood or plasma. These media were really chosen as they are more amenable to high-throughput screening of blood donors.

We then moved on to a Phase II study, which was pilot studies. They did give the first opportunity for multiple labs, some of whom had not detected virus before, to screen clinical samples that were found to be XMRV-positive in some labs. But the main purpose of the study was really to look at what sample type was really amenable to these assays. Most of the tests have been done on PBMCs in the published data, but we wanted to look and see whether whole blood and/or plasma were equally amenable or better for detection of XMRV, just because our repositories are based on those sorts of samples rather than PBMCs.

We also wanted to look at the timing of processing of the collected blood into these different blood components to see if that affected sensitivity of the assays.

Phase III is really an extension of that. We have a lot more positives, both from the WPI, from the Lombardi paper, and also from Harvard, from the Lo et al. studies, and also an increasing number of negatives. So we can really start to look at the clinical sensitivity and specificity of the assays involved in the studies.

Finally, our Phase IV study is blinded panels of over 300 blood donors, in order to start to get a preliminary prevalence of XMRV, both by nucleic acid and also hopefully by serological testing.

To go back to the Phase I study, as I said, it's serial dilution of chronically infected cells with a supernatant in whole blood or plasma, which were obtained from negative donors who were pedigreed by all of the participating labs, by PCR, serology, and virus culture, to be negative on at least two different occasions.

These are the results. The labs involved were the CDC, two different labs from the FDA, including the Lo lab, Gen-Probe, NCI Drug Resistance Program, and also Judy's lab at WPI. You can see from both the whole blood and the plasma panels that there really wasn't that much difference in the overall sensitivity of detection of the different assays from the different laboratories, the one exception maybe being that Gen-Probe for plasma did seem to be fairly significantly more sensitive.

assays for plasma were found to have no real substantial differences in terms of sensitivity. The overall similarity of the results suggests that the sensitivity of assays cannot explain the differences that these different participating labs found in clinical samples.

That is said with this one large caveat: 22Rv1 cells are obviously infected with a single isolate of XMRV, so it may be that this really doesn't adequately represent the diversity of all the XMRV and other MLV-related sequences that may be found in clinical isolates.

Moving on to the Phase IIa study, the WPI collected blood, using an independent phlebotomist, from four subjects who were found to be positive, either by PCR, serology, and/or culture, in the Lombardi et al. study. These were either processed immediately or put in 4 degrees for two or four days and then processed. Each sample was processed into PBMCs, whole blood, and plasma.

Unblinded panels were distributed to the WPI and CDC and a blinded panel, including some XMRV-negatives, were sent to the DRP lab at NCI for testing.

The four patients were all female. They ranged in age from 25 to 50 years old. Three of them had been diagnosed with CFS over 20 years ago. One was a family member of a longtime CFS patient. Three out of four of the patients tested PCR-positive at least once for XMRV. All four were seropositive at at least one time point; several, multiple time points. And probably the gold-standard assay: All four were culture-positive on multiple occasions.

One of the patients had initiated antiretroviral drug therapy around the time that these initial samples were taken.

Looking at the CDC results, for whole blood and PBMCs, they performed nested PCR for polymerase and gag. All of these samples were negative. They also performed quantitative RT-PCR for protease, and again all of these samples were negative. Controls for beta-actin were positive in both cases.

However, when they looked at plasma, initially using a nested gag PCR assay, two out of the four subjects became positive -- interestingly, only at the day 2 and the day 4 delayed-processing samples. The day 0 immediately processed samples were negative for both of these subjects.

When they sequenced the product from these nested PCRs, it appeared to be XMRV-related.

They also performed quantitative RT-PCR for protease and for integrase. In this instance all four subjects came up positive at at least one time point -- and again, this strange phenomenon, that the day 0 was negative, but the day 2 and the day 4 were all positive. In this instance, you are unable from this product to tell the difference between XMRV and generic MLV sequence.

Importantly, for all PCR-positive samples, they tested for mouse mitochondrial DNA contamination.

Similarly, the WPI results -- this is using their nested PCR described in the Lombardi et al. paper for gag, followed by sequencing of the bands for positive identification. In this instance again whole blood samples all tested negative. Unfortunately, they were unable to test the PBMCs. But on plasma, they saw very similar results to the CDC. Only one out of four of the subjects was positive at day 0, while at day 2 and day 4, all of the subjects that were tested turned out positive.

Moving on to the NCI/DRP lab results, they used a single-copy quantitative PCR assay for XMRV gag. Similarly to the CDC, they ultracentrifuged the plasma to pellet virus, whereas, in contrast, the WPI extracted directly from plasma. They also spiked an internal virus control into the plasma, a control for the pelleting step. In this instance NCI found that all of the samples by plasma Western blot and whole blood and PBMCs, and at all time points, were all negative. The plasma internal controls and the genomic DNA controls were all within range.

So two out of three labs detected XMRV in clinical samples. Plasma outperformed whole blood in both laboratories and outperformed PBMCs in the one lab that tested them, and this strange phenomenon, that day 2 and day 4 delayed-processing samples outperformed day 0 for plasma samples in both labs. We really have no good explanation for what's going on there, other than, possibly, cells may be dying and releasing either viral particles or viral nucleic acid into the plasma.

Caveats: Obviously, this panel was distributed in a mostly unblinded fashion. It's a small sample size. The third laboratory failed to detect virus, despite a phase-sensitive assay that was equally sensitive to the other two assays involved in the study. It was mainly because of this third point that we decided to repeat these studies with a more structured collection of the samples.

The samples were collected by the same independent phlebotomist at the patients' home or work. At this point, the individual who had initiated antiretroviral therapy had been on therapy for a couple of months. We also included a pedigreed negative. This is the same negative that was used in the analytical panel, so it had come up negative in multiple assays and at least two certifications.

The blood was supplied directly to BSLI. Again, we either processed the samples the same day or, in this instance, just left them at 4 degrees for two days. We skipped the day-4 processing.

The panels were then blinded and coded and distributed to the same three labs, and also to Gen-Probe, which we included in this round of testing.

To the data. NCI performed the same assay as in the first round, except that, as well as ultracentrifugation, they also directly extracted nucleic acid from the plasma. So they did it both ways.

Unfortunately, one of the labs has yet to report the whole blood assays, so we are not presenting the data from any of the whole blood studies in this. We are just presenting plasma and PBMCs.

In this instance, NCI found again that both time points for both sample types in all of the subjects were negative. Again, the internal controls were in range.

In this round, the CDC, unlike the first round, found all the samples to be negative. Again, they ultracentrifuged prior to extraction from the plasma. They used the same assays, in addition to a couple new assays, which included a nested RT-PCR for envelope and a quantitative RT-PCR for MLV gag. However, in all of these samples, the samples were negative. Again, the controls were within range.

Gen-Probe used a high-throughput TIGRIS platform and performed target capture using a duplex assay that targets conserved sequences in two separate regions of the XMRV genome and then used transcription-mediated amplification for detection. As I showed you in the analytical panels, this is a highly sensitive assay. But again in both the plasma and the PBMCs, they were unable to detect virus in any subject at any time point.

One caveat is that it hasn't been fully validated with other MRV-like sequences. Gen-Probe is currently developing a next-generation assay that should see generic MVLs, in addition to definitely seeing XMRVs. I think Jeff is going to talk about some of this work later.

Finally, moving on to the WPI results, using the same nested RT-PCR, followed by sequencing for confirmation, they did not find the virus in any of the plasma samples. So none of the labs involved in this round of studies found virus in the plasma. They did find virus in PBMC samples in three of the four subjects. However, virus was also detected at one of the time points in the pedigreed negative. Investigation following decoding of the results determined that there had been a procedural error during the PBMC sample extraction which may have introduced contamination into this portion of the study. So while we can say that PBMCs may have been positive, we really can't draw any conclusions from the time points that we see from these studies, as there may have been a contamination in this instance.

This is just a summary of all those results, again highlighting that three of the four labs were negative, including the CDC, in this round of testing. We also did serology assays in this round. The CDC performed Western blot for multiple MLV antigens and found all of the plasma to be negative. Frank Ruscetti at NCI performed the flow cytometry assay that Judy described from the Lombardi et al. paper and found that three of the four subjects were serologically positive, some at multiple time points. However, again, the pedigreed negative was also serologically positive at one time point.

In this round of testing, only one out of the four labs detected XMRV in clinical samples, and it was only on PBMCs. Ultracentrifugation or direct extraction of plasma did not seem to make any difference in detection by the NCI lab that performed it. A sensitive NAT assay from a diagnostic company was also unable to detect XMRV.

So based on the Phase II findings, if you look at just the Phase IIa, you could say that there was a clear advantage to delayed processing. Based on the IIb samples, we really can't say that. There doesn't really seem to be any direct, clear advantage. So rather than do this again, we are going to continue with collection of the Phase III panel and revert back to a standard next-day processing protocol -- kind of taking the middle road, if you would, and hedging our bets that way. Also I think this study highlights that we really do need to include the serology in parallel with the nucleic acid testing in the future studies.
Thank you.
DR. HOLLINGER: Thank you, Dr. Simmons.
Questions for Dr. Simmons?
DR. BOWER: I probably missed this. The subjects in Phase IIa and Phase IIb were the same subjects?
 

DR. SIMMONS: Yes, they were the same subjects, plus the addition of an additional --
DR. BOWER: An additional one, okay.
As far as this delayed processing, to the gentleman from the U.K. who seemed to know a lot about these viruses, have you ever experienced any advantage or do you know of any reason for delayed processing in these viruses?
DR. STOYE: No experience. No Comment.
DR. BOWER: Okay, thank you.
DR. HOLLINGER: Dr. Klimas?
DR. KLIMAS: Just on that day 4 thing. I sort of lost the reason why you went from a day 0, 2, 4, to two, to one.
 

DR. SIMMONS: In the Phase IIa, there was no real difference between day 2 and day 4. If it was positive on 2, it was positive on 4. So we decided it wasn't necessary for the second round, just for practical reasons. We are really going down to day 1 because we are trying to hedge our bets -- and the fact that there didn't seem to be any difference in IIb between day 0 and day 2.

DR. KLIMAS: It just seems like such a small sample number to make that decision, particularly when you have positive data on day 4. Why wouldn't you do day 1 and day 4 or something in Phase III? DR. SIMMONS: But day 1 is standard, basically. That's why we are doing that. We really don't feel that we can go around in circles doing this Phase II study for any longer. We really want to move on. And we do know from other publications that people have done standard processing and found positive samples. DR. BOWER: At least to me, Phase II did not answer the question of whether delayed processing helps. I think that's still unknown.

DR. SIMMONS: We are initiating a separate study that is not part of the research working group that hopefully will address that.
DR. HOLLINGER: Do you have a question?
DR. HENDRY: Michael Hendry, CDC.
Just a couple of points on the working group. One of them is kind of to add to Jonathan's list about the biological plausibility. In cases whether prostate cancer or CFS, except for the few anecdotal reports from Lo's lab, we don't have any longitudinal data. Whether we are looking at PCR, serology, or anything else, we don't know whether these are maintained or whether they are transient responses.

From the Phase IIa study, I also wanted to point out the fact that the WPI and the Lo lab were the only two labs that reported positive signals in the negative control.

Finally, in comparing the positive serology results in the Phase IIb from WPI and the positive results from NCI, there was absolutely no concordance. That is, there were no cases where they were positive both by serology and by PCR, which is, I think, another important point.

DR. HOLLINGER: Yes, John, last question.
DR. COFFIN: I was going to comment a little bit further on the issue about delayed processing. I also know of no evidence in retrovirology where delayed processing, presumably leading to lysis of the cells and release of intracellular nucleic acids, which are then easier to detect, has been reported in the case of retrovirus infections. But my understanding, particularly from Dr. Busch, is that there are many other cases where that, in fact, is a very strong booster of responses.
You might want to comment on that.

DR. BUSCH: We have seen this with PTV, sort of broad-variant category, where delayed processing results in a significant increase in plasma nucleic acids, also mitochondrial DNA sequences with delayed processing. We do have one project that is not published yet, but was presented at a meeting, where HERV-K sequences, which have been implicated in breast cancer, et cetera -- when you delay the processing, the levels of plasma nucleic acids go up, even in healthy people.

The other reason why we got into this was because, as Graham pointed out, all of the repositories that I'll talk about later that we have access to, to further study this, were processed on day 1 or 2 post-phlebotomy, and we just wanted to verify that that didn't diminish detection. We are actually interested in the preliminary data from IIa that it increased it.

DR. COFFIN: Can I make one follow-up comment on that? In the case of XMRV, when we were trying to set up the assay for direct analysis of plasma, we tried quite hard to recover pure RNA from in vitro transcripts that was spiked into plasma. At best, we were able to recover 1 percent by sort of instantaneous processing. Free RNA does not survive in plasma more than seconds, I would say. So if there is material that's released that is detectable in a PCR assay, it has to be in some kind of a protected particle. Plasma really eats it up very fast. We rediscovered that.

DR. MIKOVITS: I also want to correct Dr. Hendry. There was complete concordance between Frank Ruscetti's serology results and our work with those patient samples. We don't do direct PCR. We were asked to do that for the purpose of this study. We isolate virus from these people all the time. We have done longitudinal samples over decades and isolated virus from these patients and have concordance with the immune responses. So that's not correct.

DR. HOLLINGER: Thank you, Dr. Mikovits.
It is now about 4:30. We're going to take a break until 4:45. Then we will come back and complete the last two and then go to the open session. In the open session, we will run through the talks and then ask questions later.
Thank you.
(Brief recess)
DR. HOLLINGER: I think we'll move forward. The next talk is going to be by Michael Busch, who will discuss "Prospective and Retrospective U.S. Donor Surveillance Studies."

Table of Contents

276

Agenda Item: Prospective and Retrospective U.S.
Donor Surveillance Studies

DR. BUSCH: Thanks, Blaine. Really, carrying on after Graham's work, as he alluded to, the working group's mandate is to both execute the studies in terms of evaluating assays, hopefully identifying assays and establishing their performance characteristics, and to actually apply those tests to address the transfusion-transmission question. I want to present the design of some studies -- again, these are not yet funded, but submitted for funding -- to begin to move forward to assess the prevalence over time, as well as transfusion-transmission rates. The initial studies that I'll describe would go back to archived samples dating back almost four decades that NHLBI has put away. I'll walk through those repositories in a minute.

The concept that we have proposed is to initially test 10,000 samples, 2,000 from each of these time-defined donation repositories. For three of these repositories that are linked donor-recipient repositories, the plan would be the test a larger number of the donation specimens, if we determine prevalence is high enough, to be able to then look at the recipients and assess the rates of transmission from either seropositive or nucleic acid-positive donations to recipients, correlate transmission with viral and serologic findings in the positive donations.

These repositories are quite unique in several of the cases, in that two of them flank the period of introduction of leukoreduction. So half of these recipients received blood that had the white cells removed, which is current standard of practice in the majority of U.S. transfusions, and half did not. Also one of these repositories, the VATS, is an HIV-infected cohort, so we could look at disease penetrance in the context of HIV preexisting infection in these recipients. There is data on mortality and morbidity in the recipients of these studies.

At least at this stage, we have commitments and have funded through the application support to Abbott and Gen-Probe to help with the testing, so we could push through the kinds of numbers we are talking about here, which would be over 20,000 donations.

This is probably a little bit hard to see. Basically, the repositories that we have selected from the larger group of what is called the BioLINCC system that NHLBI has supported over decades and now has made available through public-use access -- the repositories that we have focused on start with the transfusion-transmitted viruses study, which was collected in the late 1970s. This is a linked donor-recipient repository, with about 6,000 donations that all went into 1,500 recipients. This is just a serum repository, so we don't have the ability to look at cell-associated virus.

We also have the transfusion safety study repository, collected in the early years of the HIV epidemic, prior to the availability of HIV antibody testing, but just before 200,000 donations were frozen down, later tested, recipients followed to look at HIV and HTLV transmission rates. The REDS repositories -- we have two of them. One of them we call the GLPR, the General Leukocyte Plasma Repository. This is just a donor repository, about 150,000 donations from 1994-1995. Here for the first time, we began to freeze down these frozen whole blood preps that were prepared one to two days post-collection by simply direct-freezing portions of the anticoagulated whole blood. So we have plasma and companion frozen whole blood.

The VATS is the viral activation by transfusion study, conducted in the mid-1990s, again, a linked donor-recipient repository. These were all HIV-infected patients who were randomized to get leukoreduced, filtered blood or non-filtered blood. They have been characterized extensively for virus prevalence and reactivation and transmission of viruses.

Finally, the RADAR repository, the most contemporary repository, collected in the early 2000s, is a donor-recipient linked repository, with plasma, as well, as frozen whole blood. There are over 100,000 donations in this repository, but 13,000 of these were transfused into 3,500 recipients where we have pre- and post-transfusion samples.

So these are the resources that we envision studying to evaluate the issues of XMRV. Again, the aim, very briefly: The idea is to test 2,000 representative donations from the five repositories that I alluded to, dating back to the 1970s. These would actually be donations that did not get transfused into the recipients. We want to first evaluate whether we have the power within the repositories to go to the most precious linked donor-recipient specimens.

The basic hypothesis -- and I do have some backup tables with the power calculations -- is that we will find prevalence in the .2 to 8 percent range and that using a combination of sensitive nucleic acid and serological assays, we will identify donors who have active viremia or evidence of past infection, and we will be able to then look at seroprevalence, viremia prevalence, over time in these donations and also examine the relationships between demographics, like age and gender, region of the country. All of that information is available.

The most critical piece, though, is the transfusion-transmission question. Our hope is that we will be able to demonstrate prevalence using established assays that we are working to validate, and then we will be able to go to the three repositories I summarized -- the TTVS, the VATS, and the RADAR -- that have linked donor-recipient specimens. Overall, there are about 11,000 donations that we propose to study that were transfused into the recipients for whom we have pre- and, in most cases, serial frequent post-transfusion samples. The goal will be to test all the donation specimens and then, for recipients who are exposed to either antibody or nucleic acid-positive donor units, as well as matched controls, we would test the recipient post-transfusion and pre-transfusion specimens.

Again, the goals would be to correlate the transmission events, assuming that is observed, with the levels of virus in the plasma and cell compartments and the presence or absence of viral antibodies. Because in two of these studies half of the recipients were transfused with leukoreduced and half not, we can look at the effect of white-cell removal on transmission. We also have storage duration, so the time from collection of the blood to transfusion, to see if that is a variable in transmission rate.

This includes essentially healthy cardiac surgery patients in RADAR -- mostly cardiac surgery -- and then HIV-infected patients in VATS. So we can really also begin to tease out whether there is a susceptibility or disease-penetrance effect of underlying disease in the recipients.

Using a hypothesis of prevalence that ranges from .2 to 7 percent, and with 11,000 donor exposures, we should identify fairly large numbers of exposed recipients. If one assumes a 50 percent transmission rate, we'll have fairly large numbers of recipients in which to study the early dynamics of infection and look at disease outcome.

In terms of outcomes, from the aim-2 recipients, we do have clinical outcome data that was measured as part of the early studies. To the extent those data exist and the sequential bleeds that we have in the freezers are characterized, we can look at the viral loads, the sequence evolution. We can attempt to isolate virus, look at immune responses, and then correlate those with the survival and the clinical outcome data that exist in these historical repositories.

So through these studies, we believe we can contribute to the characterization of viral dynamics, immune responses, and possibly to disease associations.

Finally, a slide that we have sort of debated in the context of the industry itself: When would there be a role for a prospective screening trial? To a certain extent, you have a catch-22 here. If you haven't established that this is transmissible or is causing diseases and you move forward with prospective screening, kind of analogous to the Puerto Rico dengue issue, you pick up infected donors, but you no longer are transfusing those units, so you no longer have the ability to investigate transmission rates, et cetera. To my mind, it's premature to consider launching a prospective study.

But we have begun to think about the scope of such a study and discuss it with commercial collaborators. What we would be looking at would be at least a study of 20,000 donations. I think, given current practices, we would need to interdict the seropositive donations, so this would be a real-time study. We could enroll the donors into follow-up and study them in terms of immunology and virology. Because this is a chronic infection, unlike dengue, these donors presumably have been infected for quite a while, so we could do what we call look-back, which is to trace prior recipients from these donors and evaluate the positivity rate in a case-control design in terms of a look-back study.

But I think the critical question that this committee is being asked to consider is, are we there yet? Is there any role at this point to undertake a prospective trial?
Thank you.
DR. HOLLINGER: Thank you, Mike.
Dr. Nelson?

DR. NELSON: Mike, as you know, we have a repository, the FACS study. We don't have donors, but we have a much larger population of transfused recipients and exposures. We have about 11,000 transfused cardiac surgery patients exposed to 120,000 units. We have 3,000 with essentially the same disease who weren't transfused. So there is a control.

I understand the importance of linking donors and recipients, but if the transfusion risk is lower than, let's say, one in 10,000 or something like that, this repository might be large enough to pick it up.

DR. BUSCH: I agree with you on the FACS study that you built. I think that makes sense, probably particularly for a serologic test. You have, similarly, pre-transfusion and six-month post-transfusion bleeds?

DR. NELSON: Yes. The economy of the testing is that you can test 10,000 or so pre and post, and get data on 120,000 exposures. In the early days, as you know, cardiac surgery patients got a lot of blood and blood units. We have also mixtures of whole blood, plasma, clotting factors. So that could all be possibly analyzed.
DR. BUSCH: Good.
DR. HOLLINGER: Yes?
PARTICIPANT: Mike, I noticed that one of the repositories you were using was from the 1970s, in the initial study. Does it concern you at all that we don't know when this virus entered man?

DR. BUSCH: That is partly what we are hopeful of contributing to. If those samples from TTVS are negative and we see accruing prevalence over time, that suggest that this is a new agent. If it has been around, if we find comparable prevalence -- 5 percent -- 40 years ago to now, and we don't have a lot of disease in transfusion recipients or transmissions, then that would reassure us that even though the virus is real and prevalent, it may not be a particularly penetrant infection.
DR. HOLLINGER: I think Dr. Stoye said 50 million years ago.
Yes, Dr. Coffin.

DR. COFFIN: Just a follow-up on that train a little bit, of the sequences that have been reported, their similarity to one another and to MLVs is most consistent with very short transmission chains from human to human. Assuming the virus is, in fact, getting into people, it's not impossible to consider that every little localized outbreak starts with a single mouse somewhere, because this virus clearly has very recently come out of the -- all of the sequences must have fairly recently come out of the mouse germline and might be analogous to the hantavirus outbreak, for example, where conditions, for some reason, allow this virus to replicate in some wild mouse and then spread around to humans. The same kinds of things could happen occasionally. One could have these little foci of infection of a virus that worldwide could be exactly identical from one to another, because the continuity has been carried in the mouse germline.

One has to be very careful not to -- although, as you know, I still remain quite skeptical about a lot of the issues, one has to be very careful not to think of this virus in terms of a virus like HIV. You have to sort of put what you think you know about HIV to one side, as far as things like genetic variation, epidemiology, and so on. This could be a completely different situation. We have to keep that in mind.

DR. BUSCH: I think the one prerequisite for any of these large epidemiologic studies, though, is having throughput assays that are validated and sensitive. Right now these panels that Graham described -- we send them out, 50, 60 samples, and two or three months later we're begging for results. We need these high-throughput commercial assays in order to launch these studies.

DR. COFFIN: There's no doubt about that. But, of course, there is a lot of working back and forth. Until you have assays that you know are telling you something on samples that are validated and everybody agrees with the interpretation, putting these into enormously-high-throughput platforms does not seem exactly warranted.
DR. HOLLINGER: Other questions?
(No response)
Okay, thank you, Mike.
The last presentation on Topic II, by Dr. Bagni, on "Assay Development Efforts on MLV-Related Human Retroviruses."

287

Agenda Item: Assay Development Efforts on MLV-
related Human Retroviruses

DR. BAGNI: Today I am going to be speaking about our ongoing efforts to characterize reagents and the transition to developing a sero-assay that could be used, most likely, in research labs.
Just an outline of what I'm going to cover this evening:

  • A little bit about our reagent development.
  • The availability of reagents to the research community.
  • Our assay development strategy.
  • Some of our preliminary observations.
  • Some suggestions for the path forward.

 

In terms of the development of reagents, working with the Protein Expression Lab in Frederick, Maryland, nine XMRV gene products from VP62 were cloned, expressed, and purified in multiple systems. The schematic shows the different protein products and their relative location on and in the virion.

For the reagent development, specifically, we based the clone setup on gateway entry clones, which facilitates a lot of flexibility for researchers. We have four types of protein expression clones. We also cloned some of the gene products for protein secretion.

For the recombinant antigen production, we began with initial screening in prototype sero-assays. We have moved to small-scale production and final production on some of the antigens, now that we have believe we have an understanding of which ones might be useful in a serological assay.

These are the XMRV antigen results after purification. This is a Coomassie stain gel. What you can see are the different gene products and their relative size. All of these are running at the expected size, with the exception of TM, which in this case is expressed as a fusion protein.

To take a page from some of Ilo Singh's (phonetic) work, we attempted to verify whether the recombinant antigens that we had produced would cross-react with MLV antibodies that had been generated for MLV-related studies. Kind gifts from Sandra Ruscetti and Monica Roth to capsid RT and SU, a subset of the data that I'm showing here, show that we have cross-reactivity and would expect that these proteins could be useful in a serological assay platform.

These reagents that I'm speaking of -- specifically the clones -- are now available to researchers. They can contact the NIH AIDS Research and Reference Reagent Program. We deposited the DNAs for 64 clones. In parallel effort, we are working to generate monoclonal antibodies, which would also be available to the community -- both the antibodies and the cell lines. We are doing this with the Antibody Characterization Lab. Some of these are actually ready to be deposited and others are still in the pipeline.

Just to speak a moment about the technical aspects of the platform that we chose to use, we decided to use the Meso Scale Discovery platform, which is an ELISA-based platform. Like an ELISA, you would coat an antigen onto the surface of the plate. If there is something to react with the antigen, such as antibodies in the serum, post-incubation, those would react with the antigen. You would come in with a labeled secondary. This is labeled with a proprietary molecule from MSD. With the addition of electricity and co-reactants from their proprietary buffer, you will have the emission of light. In the absence of electrical current, your background is really zero. Each of these individual spots can be quantified using software, and the light emitted is captured by a CCD camera.

Here is a discussion of the platform and the reagents that we have developed. This slide is really the transition: How do we go from having reagents which we believe would be useful in a sero-assay to actually having a developed sero-assay that would be useful to the research community? There are several hurdles, I suppose, that need to be addressed and overcome. The top three, in my mind, are listed on this slide.

To begin with, we don't know what the prevalence of the virus is in the general population. Depending on the strategy that you use to develop these assays, positive subjects in the normal donor population could confound your results and/or you may not be able to discriminate between cross-reactivity or lack of specificity in your assay, if that were to exist.

We also are still trying to characterize positive and negative samples. Not to make light of the situation, but in the case of the sero-assays that we are looking to develop, we have the situation of a little bit of a chicken and egg. We need a clinical control to validate the assay and we need an assay to validate the clinical control. That is definitely something that, if it were addressed, would make things a little bit easier.

Finally, we do not know the levels of antibodies in XMRV-positive subjects. We don't know when subjects are being infected. We don't know what the immune profile is. This makes it difficult, when data is discordant, to know whether it is the assay or whether it is something to do with the actual pathogenesis of this virus.

Some of the limitations, which I have touched on in the previous slide, but to emphasize moving forward in this talk:

  • The preliminary assay characteristics for the assay that I'll be describing are calculated from a very small sample number. We obviously are continually working to increase this number, but we are still working with a fairly small n.
  • We are also making an assumption of serostatus. We are assuming that most donors are negative, and if a subject has been classified as XMRV-positive by another test, they will be serologically positive as well.
  • Again, the lack of immune profile information is confounding.
  • I'm not the first to say this this afternoon, but if there were bona fide, pedigreed, in this case not nucleic acid, but antibody clinical controls, this would certainly would help the field move forward.

     

    To show that, besides on a Western blot, we could actually use our platform and detect monoclonal antibody reactivity to our antigen, we diluted an SFFV Env monoclonal antibody manyfold. Knowing the initial concentration of this monoclonal antibody, we are in the low nanogram-picogram range of detection using our platform and this monoclonal antibody.

    We also titrated a polyclonal capsid. Although it appears to be slightly less sensitive, based on some of our results, we assume that this still has some utility in a serological assay.

    The qualification of XMRV recombinant antigens -- the strategy that we chose was that we obtained 77 donors from the NCI Frederick Research-Donor Program, as well as donor plasma that had been archived in our freezer since the early 1990s. Those were obtained from BBI Diagnostics. We also included 39 XMRV-positive subjects from the initial Lombardi paper that were CFS patients. Without bias to which proteins might be antigenic in subjects, we assayed all nine gene products, recombinant products, to this training set. Then we used statistical analyses for each to determine if any of them had utility in a screening assay.

    The statistics that we used were the receiver operator characteristic curves. There are a few parameters that I just need to touch on so you can appreciate where we are with the data at this point. Sensitivity is the proportion of patients with the virus that will be reactive on the test, and specificity is the proportion of the subjects without the virus that will be non-reactive on the test. What you want to see when you plot these two values of sensitivity and 100 minus the specificity is, you would like to see your curve approaching the upper left-hand corner. This would be a true reactive rate, 100 percent sensitive and relatively 100 percent specific.

    If we look at the ROC curves for integrase and reverse transcriptase, we can see that this is not approaching the upper left-hand corner at all. In fact, it is along the diagonal, which suggests, actually, that if you flipped a coin, you would have a better chance of predicting this.

    However, if we move to the ROC curves for capsid, TM, and the surface unit, SU, you can see that we're doing a little bit better. In the case of SU, this data is encouraging. I need to caution that this data set we have been able to replicate with one lot of SUU protein that we produced; however, subsequent lots have not performed as well. We have identified the production issue, and we are going back and addressing that, moving forward.

    If we look at the training set, understanding that this data is taken directly from the training set -- so it's not blinded and we are just classifying how well we did -- for the subjects, of which we had 39, we see that 10 of these actually have reactivity above background for capsid, transmembrane, and surface unit. Then if we look at the donors, there is actually one out of the 77.

    Something that we still need to work on is, what is this background where we see that some of the antigens actually have reactivity that isn't concordant with the other antigens? This is certainly something that we need to follow up on, to understand what that really does mean.

    I would like to take a minute, with the caveat that these are preliminary findings and raw, non-calibrated data, meaning that we do not have a calibrated control -- when we looked at 1,000 donor samples that were provided to us by Dr. Harvey Alter -- 500 of these are NIH donors and 500 come from the Children's National Medical Center -- and we assayed for capsid, TM, and SU -- the SU, unfortunately, I'm not able to report at this time -- we see that the concordance between the two assays is around 5.5 percent. Again, we have this background, which we definitely need to address, in addition to, which proportion of this 5.5 percent are false positives or in the reminder, what might be actually false negatives. Until we actually have a better understanding of the prevalence in the human population, these are harder numbers to get to.

    In summary, multiple XMRV recombinant antigens produced by the Protein Expression Lab have been used in sero-assay development. We are able to detect reactivity to capsid, TM, and/or SU. I did not show this data. Some subjects, as well as donors, are reactive to p12, matrix, and nucleocapsid, but in the absence of clinical controls, the statistics are not conclusive. We suggest that we can include antigens that are reactive in human sera into a positivity algorithm and suggest that the requirement to be positive to two out of three or three out of three might lend some insight into what the serostatus for XMRV subjects might be. Our ongoing efforts:

    • We are continuing development of this. It's a daily activity in the laboratory.
    • Of course, we would like to refine our cut-points. I think clinical controls will help with that.
    • We will work to see if we can come up with a positivity algorithm or a reactivity algorithm.
    • Secondary assays, such as Western blots, nucleic acid tests, some of the tests that John Coffin has mentioned, in addition to single-genome sequencing. These are all complementary, and until we understand what is necessary or not necessary to say that a subject is XMRV-positive or antibody-positive, we need to consider that all these are definitely part of the game still.
    • Not to belabor the point, but pedigreed clinical controls will help with this.
    • Samples from experimentally infected animal models are an option. However, they aren't a substitute for clinical controls.

     

    But in the absence of not coming across this much more quickly than we probably expected to, the NCI has moved to generating animal models. Jeff Lifson, who leads the AIDS and cancer survivors program at NCI Frederick, has purified XMRV virions. This is an HLPC fractionation. You can see that's an incredibly pure prep. This is SDS-PAGE and immunological analysis using other MLV antibodies. We can see cross-reactivity to the native purified XMRV virions.

    To finalize, I'll say that this is where we are today. Our preliminary assay characteristics, yes, were calculated from a small sample number, but it gives us an idea of where we need to be going. We look forward to analytical performance panels and clinical controls to help with this. In the meantime, we will obviously be using the macaque infected models that the NCI is currently working on. But we will still hold out that this will be the ultimate control for the assay.

    I need to acknowledge the people in the Protein Expression Lab, led by James Hartley, who cloned, expressed, and purified all the proteins, the Molecular Detection Group, which is my laboratory, which has been working on the development of the assay, Jeff Lifson, who generated and purified the virus, and Denise Whitby and Nazarena Labo, who helped with the statistics. Funding was from the CCF office of Bob Wiltrout and Stuart Le Grice. And thank you to Frank Ruscetti and Kathy Jones for support. Harvey Alter provided the 1,000 donor samples, as well as Judy Mikovits for the WPI XMRV-positive subjects.
    Thank you.
    DR. HOLLINGER: Thank you.

    On the samples that you had -- maybe I missed this -- on the samples that were antibody-positive for the capsid, transmembrane, surface, et cetera, were these also nucleic acid-positive? Which ones, as it relates to the antibodies?

    DR. BAGNI: Yes. If they came from the WPI cohort, then they had tested nucleic acid-positive in their hands. We ourselves did not do any subsequent testing. We have not followed up on any of the donor results from the 1,000 donor samples.
    DR. HOLLINGER: So all the ones that you got from WPI were positive.
    DR. BAGNI: Yes.
    DR. HOLLINGER: Was there a distinction, then, in what kind of antibody you found in those?
    DR. BAGNI: Not that I am aware of, no.
    DR. HOLLINGER: Okay. You might think that maybe if it's surface antibody, it might be a lot different than it would if it's one of the other antibodies.
    DR. BAGNI: We know there are 20-some subjects that were surface unit-positive. Only 10 of those were positive also for capsid and TM.
    DR. HOLLINGER: Thank you.
    Dr. Klimas?
    DR. KLIMAS: So these were serum versus plasma serum?
    DR. BAGNI: These were actually mixed. We mostly are using serum, but we had some plasma in there.
    DR. KLIMAS: You sort of morphed from XMRV to MLV in your language. I wasn't sure if it was across all cases.
    DR. BAGNI: These are antigens that were cloned from VP62, which is XMRV. However, we used reagents, because that's all we have access to, that are MLV reagents the antibodies.
    DR. KLIMAS: And they worked?
    DR. BAGNI: And they worked reasonably well.
    DR. HOLLINGER: What is "reasonably well"?
    DR. KLIMAS: .62, .72, and .86. Those are good ROC curves. I like that.
    DR. BAGNI: The ROC curves are subjects. I do not know -- they are from the WPI study, so that is XMRV.
    DR. KLIMAS: The first two are lousy, but --
    DR. BAGNI: But those aren't included in the analysis. That was just to show that not every XMRV antigen performs equally.
    DR. KLIMAS: But PSA has a ROC curve of .62, and we use it every day. And that was your worst of those last three.
    DR. BAGNI: Sure. So we could say that we would see cross-reactivity to MVLs. But because it's a serological assay, there is always the question of specificity, in general.
    DR. HOLLINGER: Any questions from the committee?
    (No response)
    Thank you very much.
    We are going to move on to the open public hearing. We have told the speakers how much time they have. We'll see whether we are going to ask questions of them right after their talk or go to the end.

    Table of Contents

    300

    Agenda Item: Open Public Hearing


    Again, I must read the announcement for the open public hearing, so bear with me. Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making. To ensure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is important to understand the context of an individual's presentation. For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the committee of any financial relationship that you may have with any company or any group that is likely to be impacted by the topic of this meeting. For example, the financial information may include the company's or a group's payment of your travel, lodging, or other expenses in connection with your attendance at the meeting. Likewise, FDA encourages you, at the beginning of your statement, to advise the committee if you do not have any such financial relationships.

    If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

    With that in mind, Dr. Harvey Klein will be giving a statement from the AABB. Dr. Klein?

    DR. KLEIN: Thank you, Mr. Chairman. My name is Harvey Klein. I'm the chairman of the AABB's Interorganizational Task Force on XMRV. I have no conflicts of interest.

    AABB is an international, not-for-profit association representing individuals and institutions involved in the field of transfusion medicine and cellular therapies. AABB membership consists of nearly 2,000 institutions and 8,000 individuals. Members are located in more than 80 countries.

    AABB thanks you for the opportunity to participate in these discussions today by offering our perspective on the questions that have been posed to the committee. For more than a year, AABB has been analyzing the findings of various research groups studying XMRV/MLV as they seek to understand whether infection by these agents causes human disease. In December 2009, AABB formed an interorganizational task force that includes representatives from the blood community, a patient advocacy representative, XMRV subject-matter experts, and liaisons from a number of government agencies. The task force was charged to review available data, make recommendations for action to assess and, if necessary, mitigate risk, and develop appropriate educational materials for donors, medical personnel, and the public on the risk of XMRV transmission through blood. AABB has assembled a second task force that is focused on these same issues as they relate to cellular therapies.

    As has been already noted here today, AABB issued an association bulletin, 10-03, in June of 2010 recommending that blood collectors, through the use of donor information materials available at the donation site, actively discourage potential donors who have been diagnosed by a physician with chronic fatigue syndrome. I'll refer to that as CFS, since it has many other designations. Educational materials were provided with the association bulletin. This proactive step was taken on the advice of the task force as an interim measure following the article in Science in October of 2009 that reported the association between XMRV infection and CFS. At the time the association bulletin was published, the task force was also aware of the Lo study which was subsequently published in the Proceedings of the National Academy of Sciences in September of 2010, and considered these findings in making its recommendations.

    A policy of the active provision of educational materials to discourage potential blood donors from donating when they have particular illnesses or symptoms is a tool previously approved by the FDA which has been used successfully for many years by blood collectors. AABB believes this interim measure is appropriate and sufficient, based on the available scientific evidence. The recommended process is, to our understanding, the most active approach to deferral of potential donors with CFS being used by major blood services around the world and preferable to introducing unvalidated screening questions.

    Implied in the AABB recommendation is that potential donors who identify themselves during the donor interview as having received a diagnosis of CFS from a physician would be indefinitely deferred. Information about the number, the rate, and the demographics of these donors who self-identify is accumulating and at this point in time the numbers are not large.

    AABB standards already require that prospective donors be in good health and be free of cancer. In general, donors who present with a history of prostate cancer are deferred unless they are disease-free and no longer receiving therapy. These donor eligibility criteria for prostate cancer are supported by several epidemiologic studies that have shown no association between prostate cancer and a history of blood transfusion. Therefore, the association bulletin recommendations did not include a change in the deferral criteria for potential blood donors with a history of prostate cancer.

    The Blood Products Advisory Committee has been asked to consider the issue of donor testing for MLV-related retroviruses even in the absence of confirmed disease causation. Absent evidence that these viruses have a causal role in any human disease, it seems reasonable that the following criteria be met prior to the implementation of donor screening:

    • Evidence of transfusion transmission of these viruses.
    • Consistent evidence of association of these viruses with disease.
    • Development of validated assays for these agents that detect infected individuals but do not implicate non-infected individuals.
    Since no causal association of XMRV with human disease has been demonstrated, a decision to introduce a blood donor screening assay, were one to become available, would appear to be premature. Many commensal viruses -- for example, Anellovirus -- are known to be transmitted by transfusion, but despite extensive study have not been associated with disease. In the absence of direct evidence of causation, a decision to implement testing should be based on the assessment of recipient risk that includes the prevalence of the infection in the donor population, the transmission rate to recipients, and the current best assessment of the risk of recipient harm, compared to currently accepted risks of transfusion.

    Members of the blood community are concerned about the potential threat to the blood supply posed by XMRV/MLV and are actively involved in efforts to validate quality-control panels and develop tests for the detection of MLV-related retroviruses. However, we believe that current evidence does not support introducing any test methodology at this time. Furthermore, there are as yet no firm data to compare the efficacy of NAT versus antibody testing methods. AABB expresses its gratitude to the patients with CFS who are participating in these research projects and providing invaluable specimens for conducting the studies.

    AABB remains committed to monitoring activity and supporting research associated with XMRV, taking steps as appropriate to ensure the safety of transfusion recipients and of patients receiving cellular therapies, and participating in a dialogue with the FDA as necessary to further this goal.
    Thank you.
    DR. HOLLINGER: Thank you.
    Any questions for Dr. Klein? Dr. Demetriades?
    DR. DEMETRIADES: Dr. Klein, do you perhaps have any statistics about the number of patients with chronic fatigue syndrome who want to donate blood?

    DR. KLEIN: We have some data -- and I think there are others in the room who have gathered this -- on the number of individuals who have self-deferred since the AABB policy was introduced.

    Perhaps Dr. Stramer, if she is still here, wants to talk about the Red Cross data. I think perhaps Dr. Bianco may have some other data.

    DR. HOLLINGER: Sue, I think you data from the others, so maybe you could just pool them together or at least mention them.

    DR. STRAMER: Let me first talk about the Red Cross data. We implemented a procedure, based on the educational materials released from the AABB, on August 31. It was fully implemented in our regions by October 11. Over the next two months, we have queried our database to find out how many donors have stated at preregistration or post-registration that they have had a medical diagnosis of CFS.

    In the two months that we have been doing this, or for over 1 million donations, which is about 1 million donors, we have now had 34 individuals who self-deferred. That's 34 in 1 million, or .003 percent. So it's about 1 percent of the general population that we believe may have CFS, assuming .3 percent of the general population does have CFS. Just like other deferrals we see, based on, "Are you feeling healthy and well at the time of donation," donors who present are a preselected population and would be expected to have lower rates than observed in the general population.

    Of the 34 that I mentioned, the mean age is 52, 25 are females, 30 are Caucasian, and nationwide they represent a mix from widespread geographic areas.

    Regarding the ABC data, Lou, why don't you just mention your experience?

    DR. KATZ: Implementation among the independents is proceeding apace. I think we are at about 75 percent now. I expect over the next several weeks that we will begin to approach 100 percent.

    At my center, which is slightly smaller than the Red Cross system, where we started on August 2, we have had 10 donors self-defer out of 50,000, and 75, the last time that I queried our computers -- so the point estimate rate, a hair higher than the Red Cross, but the confidence intervals, completely overlapping.

    We are using a little bit different educational materials than the Red Cross. We are asking people who self-defer to provide us with contact information so we can do a more complete interview regarding where their diagnosis was made. We're doing that on the fly.

    The demographics of our 10 look just like the demographics of her 30-something.
    DR. KLEIN: I can also add to that Canadian Blood Services, which I don't think is represented here today, instituted the same self-deferral policy at about the same time. Their numbers are also comparable -- small numbers, but definite deferrals.

    DR. RENTAS: Dr. Klein, this morning we heard a joint statement from the AABB, ABC, NARC. This afternoon we did not. It was just pretty much an AABB statement. Is it safe to assume that not everyone that is part of this organizational task force agrees with your position?

    DR. KLEIN: It was the agreement of the task force. But as you know, with so many different members representing different organizations, we don't have an official position of every organization represented on the task force -- the government agencies, the Canadian agencies, the non-government agencies. It is the position that was recommended by the task force and adopted by the AABB. This is the AABB's position.

    DR. DEMETRIADES: Since the number of self-deferred donors is so small, why don't you use a stronger term -- instead of saying "actively discourage," exclude them?

    DR. KLEIN: I think with the currently available data, this is a prudent and sufficient approach. I think that's what the task force agreed with as well. I'm not sure that stronger terms, certainly in our statement, are necessary. Certainly we recommend strongly that people who have a medical diagnosis self-defer. I'm not sure there's much more strongly that I can say.

    DR. STRAMER: The procedure that the donors see say, if you have a medical diagnosis of CFS, please let us know. Then they are indefinitely deferred. It's just a little bit of semantics when you talk about what is in the AABB's association bulletin versus translating that into operational procedures.

    In the actual blood center operations, the donors are presented clearly with information that says, if you have had a medical diagnosis of CFS -- it's a full-page information sheet -- please let us know, and you will not be allowed to donate today, or until further information is available.

    DR. COFFIN: Do you have any information or any way to find out how many patients who have been diagnosed with CFS fail to self-defer?

    DR. KLEIN: How many subjects who have a medical diagnosis fail to self-defer? No, at this point there's no way of knowing that.

    DR. HOLLINGER: I think it is important, just for others in the audience who may not know, that there is a difference -- indefinite deferral is not permanent deferral. I think that's always important, because some people feel that when you say indefinite deferral, you never could donate again. It's a sort of interim type of deferral.

    DR. KLEIN: Thank you. It is both an interim measure and an indefinite deferral until more data are available.
    DR. HOLLINGER: Any other questions of Dr. Klein?
    (No response)
    I think we'll go on. Another person who has asked to speak today is Kimberly McCleary.
    MS. MCCLEARLY: This actually follows nicely the previous discussion.
    My name is Kim McCleary. I am president and CEO of the CFIDS Association of America. The CFIDS Association of America is the largest and most active organization working to make CFS widely understood, diagnosable, curable and preventable. For 23 years the association has supported research through more than $5 million in direct grants, sponsored scientific symposia and research think-tanks, sought to effect more responsive public policy and has widely informed the patient community, the media, the medical community, and researchers about the severity of CFS and the individual and collective toll it exacts.

    I appreciate the opportunity to address BPAC on the topic of blood safety.

    Earlier this year -- and I'll note that it was prior to implementation of the AABB guidelines for CFS -- our organization used a Web-based survey tool to administer a 50-item questionnaire about possible risk factors for CFS. Questions were designed to analyze information based on population norms from larger national surveys like NHANES. Four items related to blood donation and transfusion experience. Those four items are:

    • Have you ever received a blood transfusion
    • Have you ever donated blood or blood products?
    • How many times have you donated blood or blood products in the past 10 years?
    • How many times have you donated blood or blood products since being diagnosed with CFS?
    One thousand seven hundred forty-seven individuals responded to the survey. Ninety percent of respondents had been diagnosed with CFS by a physician. Eighty-six percent were women, and the average age was 57 years. Of the 1,529 people who answered the question about blood transfusion, 124, or 8 percent, indicated that they had a blood transfusion prior to becoming ill with CFS and 50, or 3 percent, reported that they received a transfusion after being diagnosed with CFS. The figure for blood transfusion prior to CFS, the 8 percent, is not significantly different from the general population as a whole, especially when adjusted for age and sex. However, retrospective donor-linked studies should investigate this issue with more rigor.

    One thousand five hundred thirty-one people answered the question about having ever donated blood. Forty-two percent reported ever having donated blood. Those who reported no history of blood donation were not asked to respond to additional questions about the timing or frequency of blood donations.

    Thirty of 650 respondents indicated that they had donated blood in the past 12 months and 225 responded that they had donated blood one or more times over the past 10 years. Of perhaps greatest interest to this committee -- and somewhat a response to Dr. Coffin's question -- is that 115 of 640 people who answered the question indicated that they had donated blood one or more times since being diagnosed with CFS.

    There are obvious limitations to Web-based surveys, but these results reinforce the need to expand efforts to educate potential donors about CFS. We commend the American Red Cross and independent centers that have already taken this step and now indefinitely defer individuals who indicate a past or present CFS diagnosis.

    Many in the CFS community anticipated that the studies being led by the Blood XMRV Scientific Working Group by now would have yielded more definitive information about risks posed to the blood supply by MLVs and the feasibility of wide-scale testing of blood donations. Today's presentations indicate that these issues cannot yet be resolved. However, the lack of conclusive data does not impede the FDA's opportunity to take action that will further safeguard health without an injurious impact on the availability of blood for those who need it.

    According to the FDA website, a person's suitability to donate blood depends on two general considerations: that the donation will not be injurious to the donor, and that the donated blood will not be unnecessarily hazardous to the recipient.

    It has long been the association's guidance to CFS patients that they not donate blood or organs out of concern for the safety of both the donor and the recipient. Research has demonstrated that orthostatic intolerance, low blood volume, and infections with a variety of agents are common in CFS. While more information may be needed to assess the potential threat posed by MLVs and the prevalence of this family of retroviruses in CFS patients, there already exists sufficient evidence in the 5,000 peer-reviewed articles about CFS to support an FDA policy of indefinite deferral of individuals diagnosed with CFS. This policy would be consistent with the practice being followed by most blood collection centers in the U.S. now. Based on preliminary data on deferral rates, such a policy is reasonably achievable without unduly decreasing the availability of this lifesaving resource, a policy requirement stated on the FDA website.

    The CFIDS Association of America urges the BPAC to respond affirmatively to the FDA's first question about CFS, and we restate our strong support of a policy to indefinitely defer individuals diagnosed with CFS.
    DR. HOLLINGER: Thank you. Any questions?
    DR. NELSON: Could we get a copy? Are there copies outside? That was an excellent presentation. I would like to have a copy of it.
    MS. MCCLEARY: I can email it. I had copies to bring with me, and I left them behind, unfortunately. But I can make that available to the committee.
    DR. NELSON: These are the best data I've heard so far on this.
    DR. HOLLINGER: It's good, but let's remember that this is a Web survey. If you look at the Web survey, I think one of the things, for example, says that 90 percent of the respondents had college or higher education, which I think is probably a lot higher than you would find in the CFS population at large.
    Am I correct in that?
    MS. MCCLEARY: Actually, among diagnosed patients, probably not. Dr. Klimas can respond to that.

    DR. KLIMAS: A very important point. Less than 84 percent of patients with this illness are undiagnosed. Only 16 percent are aware of their diagnosis. So it's a very big concern. And, of course, they are the ones that can really push through barriers of health care to get to their diagnosis.
    DR. NELSON: Despite that, these are interesting data.
    DR. HOLLINGER: It is. But I think you have to take this in the context of the type of survey.
    DR. NELSON: You have to take all data in the context of --
    DR. HOLLINGER: Can I quote you on that?
    Any other questions for Kim. Dr. Key?
     

    DR. KEY: I'm interested in the position of your organization and perhaps a core group of individuals. Is it pretty uniform that a person who is deferred with this diagnosis is surprised when they are? What is the reaction? Is it homogenous or is it polarized in terms of being deferred? Do you hear feedback on that?

    MS. MCCLEARY: When we remind our readers and constituents of the policy, some people are surprised that they hadn't heard about it before, because their physician or they have heard otherwise that there's no reason they shouldn't give blood. Of course, this would all predate the beginnings of donor education activities. For most people, they say, "I'm in no shape to give blood anyway. Why would I do that? I'm dizzy and feel awful most of the time. Of course I wouldn't give blood. Why would you even have to remind us of that?"

    So there are some mixed reactions to that. Nancy can probably speak to that in her practice. The remitting/relapsing nature of the condition -- and I think maybe the data that Dr. Hanson shared about the periods of fewer symptoms and relative wellness compared to the more severe periods -- do make this perhaps more of an issue for people who don't have a severe case and are bedbound and homebound. There are people that are fairly functional who also carry this diagnosis. Maybe Nancy wants to say more about that.

    DR. KLIMAS: Just that I think you will find in most cases that the patients and the advocacy community find that the risk of this infection -- actually, it's the first time they are being taken very seriously in 25 years. So it's almost a validation. That's why you might see some of the language in the open testimony things here that is so passionate. It's because they have been soundly dismissed. When I say that only 16 percent of the cases have been diagnosed -- and that's based on some very important epidemiologic work that has been done -- that describes how poorly we have been educating physicians and how terribly these patients -- if you could hear the stories, it's just pitiful, the way these patients have been treated.

    This infection, the other infections they have -- there is a lot of objective data on how ill these patients are. But for the first time, it feels to them like there is a governmental response saying, "Oh, my God, you're really sick." I think that's what you are hearing here.

    DR. HOLLINGER: Just to clarify, Dr. Klimas, there are 600,000 patients diagnosed with CFS in this country, confirmed?

    DR. KLIMAS: They are not diagnosed. There are roughly a million-ish, between 800,000 -- there is this wide range, depending on the case definition, but say a million. In that, there are about 16 percent that are diagnosed. So that's 160,000. When they self-defer, how are they going to do that. If 84 percent don't know they have this illness, you would almost have to ask them if they -- you almost need the little checklist: Do you have five of eight symptoms? Have you been profoundly fatigued.
    DR. NELSON: Would you suggest that the donor exclusion criteria use the diagnostic criteria, then?
    DR. KLIMAS: It could well be. It's a wonderful opportunity, actually, to educate the public about the illness, because when they go in to donate blood, they might learn a little something. A little handout that says, "This is what the illness is. Do you think you might have it? If so, self-defer," might be an easier way than to say, "Have you been diagnosed or haven't you?"
    DR. HOLLINGER: Dr. Demetriades.
    DR. DEMETRIADES: At least theoretically, patients with CFS should feel significantly worse after donating blood. Do you have any information on this issue in the population.
    DR. KLIMAS: Oh, yes, they do.

    MS. MCCLEARY: The orthostatic intolerance and problems with autonomic nervous system function or dysfunction are profound and have been documented by a number of centers, beginning with Johns Hopkins and the University of Miami, and David Streeten's work also showing lower-than-normal blood volume. They do have relapses, even after giving blood for normal laboratory studies or if they are involved in a research study and they have to give many tubes to participate in a certain study.
    DR. DEMETRIADES: Are these published studies?
    MS. MCCLEARY: Yes.
    DR. NELSON: Should blood banks then not transfuse blood from a donor that collapses after donating? That might be a diagnostic criterion.
    DR. HOLLINGER: Ms. Baker, do you have a question?
    MS. BAKER: Yes. I was wondering to what extent you have been working with any of the prostate cancer advocacy groups on this issue.
    MS. MCCLEARY: We have not found a lot of interest among the prostate cancer groups on this topic, believe it or not. I have wondered about that myself, why there is not greater engagement from that community in meetings like this one.
     

    DR. BOWER: I just wanted to follow up on what Dr. Klimas said. If only 16 percent of people know they have been diagnosed, then actually both the educational material and this proposed question, "Have you ever been diagnosed with chronic fatigue syndrome," probably wouldn't screen out the majority anyway. "Do you feel well today?" may actually be the best screening question. Just a thought.

    DR. KLIMAS: Think of it sort of MS-like. There is sort of a remit/relapse course. There are periods of wellness, sometimes even years of wellness. So I don't know. I don't think that would totally solve the problem. It would help a lot, but it wouldn't necessarily solve the problem.

    DR. GLYNN: I just wanted to ask for a little bit more clarification on what exactly is on the donor education materials. Do you just have the question, "Do you have chronic fatigue syndrome?" Or is it more detailed, giving a little bit about the symptoms?

    DR. STRAMER: I don't have the AABB association bulletin with me. It has two attachments. One is a placard that is very general information about chronic fatigue syndrome. That is what we include in every donor's pre-read materials. They see a picture of an individual. It is just very preliminary information -- if you have a medical diagnosis of CFS, please let us know. It's very generic. Then, for any donor that does let us know, we provide them an information sheet. That was the second attachment in the association bulletin. The information sheet provides links to the CFIDS website and to NCI, so that donors can get more information, rather than a general information sheet that the donors may not read. We provide basic information and then the links where the donors can get much more specific information.

    DR. KLEIN: If I could just add to that, there is no question. The question is, "Do you feel well today?" The informational materials provided are as described by Dr. Stramer. But they are examples. Many blood centers have elected to give more information. Some have decided to give less information, with the feeling that you can overwhelm people with details. That's really up to them to decide.
    DR. HOLLINGER: Go ahead.
    PARTICIPANT: Excuse me, I know I am not signed up to speak, but may I have 60 seconds, as a CFS patient?
    DR. HOLLINGER: (Off-mic)
    If there are no other questions, I think we'll move on to the third person who has asked to speak. That will be Dr. Jeff Linnen, from Gen-Probe.

    DR. LINNEN: I just want to give a brief update. I'm from Gen-Probe, in San Diego. We're a molecular diagnostic company. We are in a partnership with Novartis for blood screening. A large proportion of the blood in the U.S. that is NAT-screened is screened with our assays. So we really feel an obligation to be involved, in whatever way we can, in terms of research to try to understand what the threat is to the blood supply for XMRV or MLV-related viruses.

    These are the objectives of the XMRV work at Gen-Probe. The first objective was to develop a prototype assay that would work on our high-throughput system. We really wanted to make the assay as sensitive as possible.

    I'm going to actually talk about two versions of an assay. For the first assay we developed, we're not really sure how broad the specificity of this assay for detection of related viruses. With the second one, we have really opened up the detection capabilities for all murine leukemia viruses. We want to look at a lot of different specimen types, including urine, which is one thing that hasn't been mentioned here. We do have prostate tests that work in urine. We also, obviously, want to get our hands on many chronic fatigue syndrome samples. We have a number of studies planned for that. Then, in collaboration with Mike Busch and Sue Stramer, we want to look at the possibility of transfusion transmission.

    Another interest to Gen-Probe is to look into detection in prostate cancer samples a possible way to diagnose prostate cancer. This is the high-throughput system that you have heard something about.

    But before I talk about this, just a few reminders about the assay. This is a TMA assay. The amplification -- this is not PCR. It's based on transcription-mediated amplification. The assay that was used in the studies that were discussed earlier is actually a duplex assay that targets two different regions of the XMRV genome. We tried to make some guesses about what would be conserved, because there weren't a lot of sequences to work from when we initially got involved in the work.

    It runs on the TIGRIS system. By high-throughput, we mean about 1,000 results in 14 hours. The time to first result is about three and a half hours, and you see 100 results about every hour. The great thing about the system is that it has process controls for all assay steps. All of our assays have internal controls to validate each individual reaction.

    Here's a little bit of the analytical sensitivity data. This is really consistent with the data that Graham Simmons showed. It's actually two different experiments. On the left we are looking at transcript that's synthesized in vitro, and on the right we are looking at a virus. If you look at the table at the bottom, those are the results from doing what's called a probit analysis. It's a type of regression analysis from a dilution series. You can calculate at what copy level the assay would detect 95 percent of the time. With the transcript, it's around 17 copies. With the virus, based on an estimated value, it was about 2.5 copies of the virus.

    I would just conclude from this that we think the assay is pretty sensitive.

    The first screening that we did after we put together the assay was to look at a collection of normal donors that we obtained from Susan Stramer at the American Red Cross. These were all from the Charlotte, North Carolina testing site. We screened them on the TIGRIS system. We looked at a total of 1,435. The results were all non-reactive. One of the conclusions we make from this is that the assay is not prone to false positives. We try to get as much as we can out of all the data.

    We also had some samples from a previous study, where we were looking at HIV in plasma versus whole blood. We had remainders of the whole blood samples. We had a total of 44 HIV-infected blood donors. These were identified in blood screening. These were samples that we received from Mike Busch. I'm presenting the data a little bit differently here, but the same result: Out of the 44 whole blood samples that were tested in our assay, we saw no samples that were reactive for XMRV.

    Based on those negative results, we decided that it would probably be useful to make some modifications to the early prototype assay. So, as I mentioned in the beginning of my talk, we have modified the assay, where it will detect a very wide range of murine leukemia virus-like sequences. The testing scheme may end up being a little bit more complicated, because we're not quite sure what the results will mean if they are positive. It may require a lot more additional testing to understand exactly what we are detecting.

    In addition to broadening the range of detection, we have also made some changes to optimize the assay. There are some chemistry modifications that we can make to the assay to increase the sensitivity for DNA. This assay, we think, will have equivalent sensitivity for both RNA and DNA.

    What I'm showing here is some preliminary analytical sensitivity data, similar to what I showed for the earlier version of the assay. What you can see in this experiment is that the analytical sensitivity is very similar. The 95 percent detection level is estimated to be I can't read it, but it looks like 16.5 copies.

    So we think we're ready to go with this assay. We want to do a similar kind of study with maybe about 2,000 normal donors -- we are in the process of getting those samples right now -- to get an idea of, with an assay design like this, what you detect in a normal blood donor population. We should probably be doing that in the next week or so.

    What do we have ongoing right now and planned for the future? You have heard that we are working actively with the NHLBI-sponsored working group. Some of the data was presented earlier today. We are very grateful to be included in those studies. We would like to take part in the additional phases that are planned.

    As I mentioned before, we are working on arrangements to get as many CFS patient samples as possible. Really, the way I'm looking at this, I would just like to screen as many as we can. We can easily, in a week, go through 5,000 samples, if those samples are available. The tricky thing is how well the patients are identified for having CFS.

    We also would like to go back to a much larger population of normal blood donors, at least 10,000, but possibly 20,000. This will be done in collaboration with the American Red Cross also. As you heard from Mike, we are going to be taking part in linked donor-recipient testing, both with samples from the NHLBI repositories and with some linked donor-recipient samples that we will obtain through Dr. Stramer and Dr. Dodd.

    Thank you very much. I'll take any questions if there are any.

    DR. HOLLINGER: So, Jeff, this assay can detect both modified polytropic MLV and XMRV?
    DR. LINNEN: Right. With the assay that we have just recently developed, that's the case.
    DR. HOLLINGER: Thank you.
    Other questions from the committee?
    (No response)
    Thank you, Jeff.
    We have a final person who has asked to speak. This is Dr. John Hackett, from Abbott. You have 10 minutes, John.

    DR. HACKETT: Thank you. I would like to thank the chairman for allowing me to speak to you. Again, I'm at Abbott Laboratories, a stockholder of Abbott. We do have a relationship with Cleveland Clinic related to licensing of diagnostics related to XMRV.

    A little over three years, we initiated studies, because of the concern of another retrovirus, potentially. in the human blood supply. Obviously, there are many questions that are very fundamental in terms of trying to understand causality, and even prevalence and modes of transmission. The key to this will be to have appropriate assays to be able to measure this. Ideally, for these sorts of studies, we would have some serologic tools that would allow us, in a very specific and sensitive way, to reliably detect antibodies to HIV.

    Dr. Villinger earlier today showed the studies we initiated in rhesus macaques. Our approach early on is -- we had three primary hurdles. The first is, we didn't know what proteins would be responded to by primates for this virus. Secondly, we didn't have seroconversion samples to work with. Thirdly, we didn't have any real source of bona fide positive control. So our solution to this problem was to create the macaque model.

    As he indicated, if you look on the Western blots -- this is showing time after inoculation with XMRV. This was VP62. It came from a prostate cancer patient. You can see that we had gp70 reactivity, which on the Westerns for recombinant are easy to see, p15, as well as gag p30 reactivity. All of the animals responded. I'm showing results from just one here. But this was our very first insight into what proteins antibodies would be generated to in this system.

    Similar to what Jeff said, we wanted to move to a system that would facilitate analysis, once we had assays available. In this case, we are working on the Architect platform, which can run approximately 200 tests per hour. We have developed chemiluminiscent assays targeting 15E, gag p70, as well as capsid protein p30.

    In terms of these assays, there are two primary formats that can be used. The first is the indirect format, where you have a recombinant antigen or some sort of antigen that captures specific antibodies. Then one comes back to detect those antibodies by the use of an anti-human immunoglobulin. An alternative approach is to use antigen essentially to bridge. You capture the specific antibody with antigen and also detect it with antigen. There are benefits to both systems. I have to say, these are much easier assays to develop. On the other hand, these tend to be far more specific.

    To illustrate this, I show here, for the p15E assay that we generated, a distribution on the left of 97 blood donors. These were screened for other blood-borne infections and were negative. In this case, it's relative light units on the y-axis and frequency on the x-axis. Here you had a mean of about 2,300, with an SD of 1,758. Merely changing the format, not changing any of the recombinants involved -- a dramatic difference in distribution. Now you can see in the direct format that we have a mean of 329, with an SD of 114. The impact of moving to that format is that you can discriminate positive and negative samples much more readily.

    On the top panel, what I'm showing are 100 blood donors and then 36 Western blot-positive primate bleeds. These are bona fide positives, because we have seen it, and seen it by Western blot. You can see that with the indirect format, you have some overlap here. If we cut it off to call everything that we know is positive, positive, we have only 75 percent specificity.

    This came up earlier when Dr. Stoye was speaking. There can be false reactivity in these assays.

    to the direct format, that we now have 880 blood donors, including these 100, and the 36 Western blot-positive bleeds. This is a log scale along the x-axis, so you get excellent discrimination between negative population and the positive. In this case, we had one blood donor that was above our cutoff, which is established at the mean plus 16 standard deviations -- so a very conservative cutoff. That individual wasn't positive on Western blot, so we considered it a false negative. I should say, we don't know that with certainty. But we still have very high specificity in that assay.

    For gp70, we realized early on, as studies were going on, that we were going to need as sensitive an assay as we could possibly get. I should say, that p15E assay is a very sensitive assay, certainly as sensitive as any HIV assay that's on the market.

    If we look at gp70, what we did was use a signal amplification modality in the conjugate, where we have an avidin-biotin system with the gp70. This direct format detects both IgM and IgG, but generates far more signal.

    One of the questions, of course, that will come up is, because of the issue raised with the MLVs, in particular with chronic fatigue syndrome, how would these assays perform? The only thing we can do is use other surrogates. In this case this is a goat anti-Friend murine leukemia virus antiserum. You can see, for the gp70, it titered out at 1-to-16,000, was still positive, 1-to-32,000 in the 15E assay, and 1-to-64,000 in the p30 assay. Another goat anti-envelope -- it's a Rauscher murine leukemia virus -- that titered at 1-to-10,000 for the gp70 assay. This is just showing Western blots, which we have also developed, based on XMRV proteins, as well as recombinant proteins expressed in mammalian cells. I should say that the gp70 assay uses a recombinant-produced gp70 for detection.

    I just want to turn to blood donors quickly. We don't have a tremendous amount of data. This shows the 15E assay run on nearly 3,000 blood donors. You can see that we had five of them that came out above the cutoff that we had set. Three of these are quite interesting.

    On this slide, on the left panel is XMRV lysate Western blot. These three do have evidence of gag p30 reactivity. This is the gp70 recombinant, and we do see some gp70 reactivity in these three, too. So in this case, we have some samples that do appear to have multiple reactivity against XMRV proteins.

    I want to conclude. We have identified the primary markers elicited by XMRV infection. It's interesting that Dr. Bagni essentially has resolved to the same three proteins in their efforts. We have developed three high-throughput assays. Gp70 is by far the most sensitive, 15E next, and then p30. Now we are in the position that we can begin to do other studies. Mike Busch mentioned that we'll be working with him, and Drs. Stramer and Dodd will also be involved with studies along with the Gen-Probe group there.

    Hopefully, these tools will begin to allow us to determine whether we are seeing transmission of this virus.
    I would just like to close by acknowledging my colleagues, as well as Dr. Silverman and his group and Dr. Villinger and his group at Emory.
    Thank you.
    DR. HOLLINGER: Thank you, Dr. Hackett.
    Questions? Yes, Dr. Klimas?
    DR. KLIMAS: Are you using both sera and plasma or just plasma in those assays?
    DR. HACKETT: We can use either. It doesn't seem to matter.
    PARTICIPANT: The volume?
    DR. HACKETT: We need approximately 150 µL to load. That makes it tough. But usually that's the issue; there's not enough volume in these samples. That's per assay.
    DR. HOLLINGER: Thank you. Any others?
    (No response)

    DR. HOLLINGER: This really exceeds our time for the open public hearing, but I'm going to allow the lady who asked to speak just a minute ago to come up and speak to the committee. Then we'll close the public hearing. PARTIICIPANT: Thank you very much. I know this is not protocol.

    I just wanted to please ask that -- the idea that "if you are well today, you can give blood," please reconsider that. I appreciate that the AABB and Red Cross maybe followed something a little more rigorous.

    I'm a CFS patient. I was diagnosed at Johns Hopkins with CFS and fibromyalgia in 1999. I have had it for 16 years. When you say remitting and relapsing, like Dr. Klimas said, we are not talking about just weekly or monthly; we are talking about hourly. When I first got here, I had rested four days in order to be here today. Slowly, throughout this very long meeting, I'm at a point now where my lymph nodes are killing me from head to toe. I feel like I'm about to get the flu. It's a combination of mono and the flu.

    So this morning, I would have been fine to go ahead and give blood. By this point, my cognitive issues are shot. I'm sure other people's are, too. Mine went about an hour or so ago. I couldn't put simple math together at this point.

    So please do reconsider that stipulation right now. It isn't effective, even for the small population that knows they have CFS. It's definitely up and down, even daily.

    Thank you very much. I appreciate it. DR. HOLLINGER: Just a question while you are there. Thank you for your comments. The question has to do with well or unwell when you donate blood. You would be given an educational piece of information, which would say that if you are a CFS patient or have that diagnosis, you should not donate. That would have nothing to do with whether you feel well or unwell at the time of donating. Does that make a difference for you?

    PARTICIPANT: I do believe that would make a difference. I think I like Dr. Klimas's suggestion of having a pamphlet that would be able to explain CFS to the public in general as well. The patient population has been waiting for the FDA, to be able to say that we agree with some of the things that the other blood institutes are recommending right now, and we would like to see the FDA be able to back that up and say, yes, at least for now, while we're working all the science out, let's be cautious and indefinitely defer CFS patients.
    DR. HOLLINGER: Thank you very much.
    Dr. Alter, last comment from the floor.
    DR. ALTER: I just want to make a long statement.
    DR. HOLLINGER: That's all right; we can cut you off at any time.

    DR. ALTER: Since Dr. Lo had to leave early, I felt I had to come up and do some defense of him and Judy as well. I think, when a group finds a new agent, they become biased that this agent is real. When another group doesn't find an agent, they become, I think, even more biased that the agent is not real. That leads to this kind of contentiousness.

    I think our goal should be not to bring the other side down, but to find the truth. I think the truth will out over the next year, with studies that are already planned.

    At this point I concur that we have no evidence for causality. That's going to be very difficult to come by, especially when we are detecting at the limits of detectability and when assay performance is very critical to get equal results.

    But I still want to counter by saying I think the current evidence for disease association is very strong, even though not universally confirmed. But it has been confirmed now in at least four studies, two of which were presented today, that either XMRV or a polytropic MLV is associated strongly with chronic fatigue syndrome. A point that I think was misrepresented today: In those labs who do find the agent, it is very reproducible. Judy has found the same patients to be positive by culture year after year. We have found a patient to come back after 15 years and still be positive. So this is not a single, isolated finding. It's confirmed by sequencing. It's reproducible over time.

    Dr. Hanson has shown today how critical the assays are. When she tweaked her assay, she went from no findings to findings almost identical to the Lo lab. The diversity is now being confirmed also in the original WPI group. XMRV isn't the only agent even in the WPI lab.

    Despite the very legitimate concern for contamination -- I think this is a serious issue -- there have been hundreds of negative controls in the same laboratory that are always consistently negative. An extremely sensitive mouse mitochondrial DNA has always been negative in the Lo laboratory. Lo has done the IPA assay that Dr. Coffin recommended. That is also negative. There just has been no evidence for contamination. Although you could say maybe the negatives could be negative somehow and the positives positive for contamination reasons, it really is not logical that that would be so.

    I'm not a molecular biologist. I defer to Dr. Stoye, who is world-renowned in that area. But just as a simple doctor, it seems to me that you have used single-case anecdotal evidence to knock down the various possibilities. I just want to make a case to the committee that you can't -- your conclusion is that anything can happen in assays, and therefore it probably has happened this time. I think using that kind of anecdotal probability is not valid to negate reproducible data from four different laboratories. So at least keep that in mind.

    Lastly, I'm not a chronic fatigue doctor, but I have learned a lot about chronic fatigue in the last six months and have spoken to a lot of patients. I'm absolutely convinced that when you define this disease by proper criteria, this is a very serious and significant medical disease, and not a psychological disease. It has the characteristics of a viral disease. It usually starts with a viral-like illness. If XMRV is not the causative agent -- and it may well not be -- there is still need by other groups to look for the next agent which may be the case.
    Sorry to take so much time.
    DR. HOLLINGER: Thank you, Harvey.
    We are going to close the public hearing and open up for discussion by the committee. I would like to have the questions put up, please.

    Table of Contents

    340

    Agenda Item: Questions for the Committee


    We are just being asked to vote on the first question, and then there is some discussion that will take place afterwards for the other questions that are up there.

    DR. NELSON: Is number 1 one or two questions? The deferral that the AABB talks about is a physician diagnosis, but here it talks about a medical history and/or diagnosis. One could ask symptom questions or other things in the deferral process.

    DR. EPSTEIN: To clarify, we're talking about whether the donor questionnaire should contain a question specific to CFS/CFIDS/ME. The reason we framed it this way is that what we have learned historically about other conditions, such as syphilis or gonorrhea, is that some people react if you ask if you have a history and some people react if you have a diagnosis. Given the ambiguity, we just tried to capture both.

    DR. NELSON: (Off-mic) -- medical history and/or diagnosis.
    DR. EPSTEIN: Yes, that's correct. But let me just be a little bit cautious here. If the committee advises us to go forward with a specific donor question, we might well want to do some cognitive evaluation, behavioral study, to figure out how best to ask the question. But what we are getting at is a broadly framed question that would properly communicate to the candidate donor that they have the condition that we want to defer. The heart of this question is, should there be a question on the questionnaire specific to CFS/CFIDS/ME?

    DR. COFFIN: My answer to this question will depend a little bit on what the meaning of "scientific" is here. If you are talking about the sort of scientific data that we have been discussing all day concerning XMRV, I probably answer this question one way. If you are talking about my general knowledge about this condition, I think I would answer the question differently, excluding the XMRV possibility.

    Could we get some clarification on that?
    DR. EPSTEIN: I think, again, without being too precise, we are asking the committee, given everything you have heard and everything you know, do you think there should be a specific donor question? I don't think we are trying to define science. We're all struggling with the quality of the available data. We don't know any more than you have just heard.

    DR. HOLLINGER: I think the issue, at least for me, is when you start to ask a question which is difficult to add to the questionnaire, which is very lengthy anyway, you place a little bit more emphasis on items that you think have more scientific data available. I think that's one of the issues. For me personally, I think the educational material is more than ample right now for what we know about association with disease and transmission and a variety of other things, whereas I think if you asked a question about it, it leaves the impression that we know more than we do. That's how I view these two questions.

    Yes, Simone?
    DR. GLYNN: I could not agree more with what you just said, Blaine. I was wondering if we could also maybe vote on a question which supports the provision of educational materials and the self-deferral of donors if they say that they have such a diagnosis or they are not feeling well, rather than -- you can also ask the other question, which is adding a question to the questionnaire.

    DR. BIANCO: I want to support this idea. That's the approach of the AABB recommendation that is being utilized. It's not part of the consideration of the question now. I would like it to be considered, yes or no, that approach. We have over 40 questions that we ask donors on every donation. Those have sub-questions. So things get diluted. Donors, after the third question, don't know where they are anymore. The value of that information that is given to the donor is much more important than simply adding more and more questions.

    DR. GLYNN: I really like what you said about maybe using the information that could be provided to each donor as a way of maybe querying some donors about some symptoms. They may not have been diagnosed, but they may recognize certain symptoms. So if there is more information that is provided, you may be able to also provide more information on the disease.

    Again, a question -- since most of them have not been diagnosed, if you ask them whether they have the disease, they are going to say no, because they haven't been diagnosed.

    DR. KLIMAS: I would say this is a tremendous public health potential here. When 84 percent of the people with illness don't know they have it, we have done a very poor job of letting people know that there is an illness out there. Here we have a way to really educate a lot of people, whether or not they have the illness. They might recognize it in a friend or relative. It's a fabulous possibility. I know it's not our primary goal here.

    Now, I'm going to say, as a clinician who has treated thousands and thousands of these patients and who has had many, many, many papers about that, we have very robust literature on viral reactivation in chronic fatigue syndrome and immune dysfunction and so on and so forth. From my own knowledge of XMRV, separate from it, there's not a chance that I would let one of my patients donate blood. I have never recommended it. I think it's a bad idea, from infectivity, but also from the patient's own health. They have a 20 percent reduction in blood volume. Why would you give another liter? It's just nuts. But they do, because they are very charitable and kind and giving, and when they are well, they want to do their bit.

    So there are a lot of good reasons to want them to defer.
    DR. HOLLINGER: Dr. Rentas?
    DR. RENTAS: I think everyone has summarized what I wanted to say here. If you have taken a look at these educational materials that we are putting out there now -- if I recall, the title is "What You Should Know about CFS and Blood Donation," which we are making available to the donors out there -- to me, that's a much better process at this point, because for 84 percent, if you ask them the question, they are going to answer no because they don't know that they have it. To me, it's a much better process to let them know, "If you have any of these signs and symptoms, you may have CFS. Why don't you go ahead and self-defer and then find out whether you have it or not," than to ask a question, most of the time on a computer, with no one there, that you most likely will answer no.

    By the way, I need to put this out. Ever since this policy came out to the AABB, we have called this policy "they don't ask, you tell."
    DR. HOLLINGER: Thank you, Dr. Rentas.
    DR. DEMETRIADES: I think we all agree that there is no evidence about the causative relationship. However, there are some significant at least theoretical concerns about the welfare of the donor and the recipient. In my mind, it makes good sense to have a specific question that excludes these patients from donation. It's not going to affect significantly blood donation. At the same time, we alleviate a lot of fears of the public

    DR. NELSON: But we're not talking about replacing the educational materials, just adding a question. After a person reads the educational materials, then I don't see anything wrong with asking a question.

    DR. BOWER: I was going to say that I do not believe that we should have this because we think that people with chronic fatigue syndrome can transmit it to others through transfusions. But sitting here and learning something that I didn't know when they talked about all the other commensal viruses they may have and transmit makes me think that it would be a good idea to keep them from donating blood, for their own good and for the good of the recipients.

    Analogous to what the best way to do that is, I would like to ask the FDA why they think that the current policy of giving educational material is inadequate and asking a donor-referral question would be better.

    DR. HOLLINGER: Dr. Dodd?
    DR. DODD: I just thought it might be worthwhile reminding the committee that in a time of considerable uncertainty about AIDS, or whatever it was called at that time, Mike Busch reported that a policy of asking would-be donors whether they had a risk factor for HIV/AIDS at that time succeeded -- and he published this in reducing the frequency by about 100-fold in a very high-prevalence area. I think it's important for the committee to recognize that the educational and self-deferral approach does have something going for it. Harvey actually said this, but I think it might have slipped by the committee.

    DR. KLEIMAN: I just want to urge the committee to -- because people are saying things that are sort of lumping things together. I think it's two separate issues. Just sticking with the educational material, should it address a diagnosis of CFS or should it, in addition, provide people with a list of signs and symptoms, and let them self-diagnose at the site and make a decision about whether they donate. I think that latter approach, for one thing, may have a big impact. We don't know how many people have had some symptoms, and it may not be people with CFS. So I think the non-specificity -- I think a diagnosis is -- I'm not a CFS expert, but I think it is supposedly difficult to make a diagnosis, that you really have to interview the person and get five out of eight symptoms or something like that. To rely on a person -- maybe that's incorrect.

    DR. KLIMAS: They have disabling fatigue. That's a very big hurdle. And exercise-induced relapse is a second really big hurdle.

    DR. KLEIN: But I do not think we are going to get disabling-fatigue people -- I mean, if it's that severe, I don't think people are going to come to donate blood.

    DR. NELSON: I don't see anything wrong with combining the question. Maybe the person hasn't had good medical care, but they just got out of bed. It seems like there is a lot of undiagnosed illness. If somebody self-diagnoses, do you want them to donate blood?

    DR. KLEIMAN: I think the point I'm making is that if the recommendation is to educate people and ask them to defer based on their own symptoms, we had better have a very clear way of presenting those symptoms and to differentiate disabling fatigue from somebody feeling that tired on the day or having felt a little rundown in the past week, but not really having -- apparently, you can do that. You can come up with those things. But if we are going to go that route, I think we need help from the clinicians to make sure that we really provide the right symptoms, and not just sort of vague discomfort.

    DR. NELSON: It's voluntary. So if they self-defer because they don't feel well -- I don't know. I don't think that's going to destroy the blood supply, frankly. I think the question should --

    DR. HOLLINGER: Well, it depends on what the material says. If it talks about fatigue and so on, that's one thing.
    Dr. Stoye, do you have a response to some of the questions that have been asked here?
    DR. STOYE: I just wanted to say one thing. In Britain, a decision was made to defer indefinitely CFS-positive donors on the basis of their own health, which seems to be what Nancy is saying is at least one reason for doing it. But, in fact, as I understand it, among certain communities, this decision met with derision, because it was said you are just trying to protect yourselves; it has nothing to do with the patients' well-being.
    You should think about how you make any decision before you make it.
    DR. GLYNN: So in England do they ask a specific question?
    DR. KLEIMAN: The answer is no.

    DR. EPSTEIN: There is no country that is asking donors a specific question at this point in time. It has been contemplated in some countries, discussed in Canada. But to my knowledge, although there are many that are asking donors to self-defer based on a history or diagnosis, no country is asking a specific donor question.

    I think it comes back to Dr. Hollinger's point, which is that we tend to see that as meeting a higher threshold of certainty that intervention is warranted. Again, this can be debated, but that has been the general philosophy. You have to know you're doing some good, in other words.
    DR. HOLLINGER: Dr. Bower, you have been patient. Do you have a question?
    DR. BOWER: You know, I forgot what I was going to ask. Let me think about it.
    DR. KATZ: In fact, there is one country that asks specifically, the Flemish Red Cross.

    How many anecdotes does it take to make data? Celso and I are on a group with the European Blood Alliance that monitors emerging infections. When they were queried with regard to the number of deferrals, it's somewhere in the range of what Sue and I are seeing. It's not obvious, unless the Flemish in Belgian have a much lower prevalence of chronic fatigue syndrome -- it hasn't been clear to me that a specific question is, in fact, markedly more effective. But one country.

    DR. KEY: I am not a CFS expert. Neither am I a blood banker. I have to ask, if five years from now the link between XRMV and CFS is shown not to be the case -- there is no link causally -- and this is an indefinite deferral, what happens at that point? We heard most of the data in terms of transmissibility, I think. But we are hearing now that, particularly in the U.K., it's to do with the donor health as well.

    What are we being asked to vote on here? Are we being asked to vote on both of those, one of those? Do they stand alone, individually?

    DR. HEWLETT: I just happen to be up here at the microphone. I was going to make a comment about Lou Katz's information about the Flemish Red Cross. It's our understanding that they already had in place a question a while back and it was really not connected to the identification of XMRV. I just wanted to make that point.

    No country has put any questions in place as yet as a result of this finding of MLV and XMRV.

    DR. BOWER: Just to answer Dr. Key's question, at least this committee member is looking at this aside from the XMRV. The data on that is too all-over-the-place to really allow me to link that to chronic fatigue syndrome. I'm just looking at it as, should someone with that diagnosis be allowed to donate?

    DR. HOLLINGER: Prerogative here. I think, unless there are some burning questions, we have talked back and forth here. I think we'll go ahead and vote on this question, unless somebody has a real burning -- yes?

    DR. BIANCO: Real burning question. The real burning question is, is there an opportunity for this committee to recommend the approach that AABB is recommending of providing educational materials to the donors, without asking a specific question or not? These questions say -- I can't vote, but if I answered no to this question, that doesn't mean that I don't agree with the AABB is recommending. That's missing there.

    DR. HOLLINGER: To me, I think it's obvious. The AABB already has educational material out there. We're assuming that that would remain the same. Yes, Jay?

    DR. EPSTEIN: FDA pays a lot of attention to voluntary standards, and if we object, we make our objections known. We do not object to the AABB voluntary standard. It's true that we could establish that as an FDA recommendation. I don't object if the committee wants to weigh in on that point. But we do see value in the AABB current practice.

    DR. GLYNN: Could we also recommend that maybe more data be collected? I personally support the AABB recommendation. Let's say that you go ahead and gather data at the blood banks -- some of those data are being collected, but maybe you could collect additional data. If you ask the question, in a particular research study, how many patients would have self-deferred, you could gather additional information to inform us about exactly what's happening with just the self-deferral and the provision of the information. Would that be possible to do?

    PARTICIPANT: (Off-mic)
    DR. GLYNN: I don't have any money -- not personally, that's for sure.
    DR. HOLLINGER: Okay, I think it's time to answer this question. The question is straightforward. It's really about asking a question of donors who come into the blood bank about a medical history and/or diagnosis of CFS as a basis for indefinite deferral. It's about a question, not educational material. "Yes" would be that you favor asking a question; "no" would be that you don't favor it, that you think -- in my opinion, what it means is that you favor the educational material which is currently being used.

    DR. BOWER: I want to make sure that we understand that, because you threw in the educational material. It sounds to me like a yes vote means that they put a question, "Do you have a diagnosis or history," on the questionnaire versus a no vote, which is the status quo, that they proactively provide information and allow them to self-defer. DR. HOLLINGER: Yes, that is how I would interpret it.

    You see your blinking lights before you. You can vote yes or no or abstain.
    (Vote)
    Before we look at these, Dr. Bianco, how would you vote?
    DR. BIANCO: I would vote no. I favor the educational material.
    LCDR EMERY: The committee has voted. We have 9 that have voted yes, we have 0 abstentions, and 4 that have voted no.
    Dr. Coffin has voted yes.
    Dr. Judith Baker has voted yes.
    Dr. Klimas has voted yes.
    Dr. Bower has voted no.
    Dr. Troxel has voted yes.
    Dr. Nelson has voted yes.
    Dr. Hollinger has voted no.
    Dawn Aldrich has voted yes.
    Dr. Trunkey has voted yes.
    Dr. Demetriades has voted yes.
    Dr. Rentas has voted no.
    Dr. Key has voted yes.
    Dr. Glynn has voted no.
    DR. HOLLINGER: Thank you. So we have a vote.

    Let's go on with the second question, which again does not require a vote, just comments. The question is, please discuss whether the scientific data support asking donors -- it's the same question as the one before, but it has to do with prostate cancer as a basis for indefinite deferral.

    Comments? We're not going to vote on this, so feel free to give your comments.
    DR. BOWER: My comment is that looking at the available data, there is no association with prostate cancer after blood transfusion. If you look at the studies that have looked at XMRV in prostate cancer patients, they have not been able to find it in the blood. I believe that having prostate cancer does not put you at risk for a transmissible disease, so I would say no.

    DR. HOLLINGER: I have seen one report from Japan that apparently found virus in PBMC of one or two donors, I think, out of four. Does anyone else had any data that they would be able to talk to us about, about finding it in the plasma or PBMCs of patients with prostate cancer?

    DR. TRUNKEY: I don't have anything in that, but the most positive things I heard today were the three studies that were presented. One was to go back and look at blood samples in a repository, repeat that, and then try to link it to the things that we are talking about here. The second one is the effort to look at other antigens within the virion, which I thought was very good. Then the macaque studies I thought were very good, particularly as it pertains to prostate cancer, because he showed us the virus is there in the prostate in these macaque monkeys.

    I think we have an opportunity here in the next year or two to be able to answer this question. And I think we should wait.

    PARTICIPANT: Infectious virus and antibody have been found in prostate cancer patients. I have found it.
    DR. HOLLINGER: Infectious virus in blood or in the tissue?
    PARTICIPANT: From the plasma and PBCs of prostate cancer patients.
    DR. HOLLINGER: And this is published?
    PARTICIPANT: No, it's not published.

    DR. COFFIN: There is a small-scale, 100-patient study in process within the NCI to try to dissect this issue in more detail, in at least the small sample of patients being treated at the Clinical Center. That should provide us with some more information on this issue within the next year.

    My yes vote in the case of CFS was not based on the XMRV connection at all. On that basis alone, I would have voted no. But it was based on the evidence that I heard that matches my intuition and that of other people that there could well be an infectious agent here. It might be XMRV; it might be something else. For that reason, that kind of caution is warranted.

    There's no such evidence in the case of prostate cancer. There's no evidence, anecdotal or otherwise, for any kind of a predisposing infection or anything that I'm aware of. I thought the discussion we heard earlier on the subject from one of the public presenters was very much on the mark on this issue.

    DR. HOLLINGER: Dr. Hewlett?
    DR. HEWLETT: I was just going to respond to your question about whether PBMCs were found to be positive in prostate cancer patients. We heard from Frank. There are some published studies. The one that I had on my slide is actually by Fischer et al., from Europe, where they looked at 40 PBMCs from prostate cancer patients. They were all negative.

    As I understand it, the CDC presented some data at the CROI meeting last year where they had looked at some prostate cancer patients, found them in the tissue, but the PBMC was negative.

    So that's what we have to date. But you're right, it might be worthwhile doing a bigger study to get more information on that.

    DR. NELSON: Where the virus was found in tissue of patients with prostate cancer, was it early and localized? Was it when the tumor was localized or was it after the patient had either metastatic or -- in other words, when is the risk?

    DR. KLIMAS: Those are mostly the biopsy-for-diagnosis samples. So it was pretreatment and, I'm sure, at all stages of illness.

    DR. COFFIN: In the study published from Ilo Singh's lab, it was found in all stages, but there was a correlation of frequency of detection with the Gleason grade of the tumor -- not necessarily with the stage of the tumor, but with the grade of the tumor.
    DR. NELSON: So the virus could have been secondary to the tumor.
    DR. COFFIN: That is always a possibility.

    DR. HOLLINGER: By the way, I forgot to mention that Dr. Stephan Monroe is in the audience. He is a guest from the CDC. He's not a voting member.

    Do you have anything to add to this discussion?
    DR. MONROE: Just to clarify the CDC data on the prostate cancer, I believe there were three samples in that study that were positive, of which two had matched PBMC, and those two PBMCs were negative for DNA. It's very small numbers.

    DR. HOLLINGER: Thank you. If you have any other comments somewhere down the line, please feel free to jump in.
    Go ahead.
    DR. SHOCKMAN: Gerry Shockman, from Abbott.
    I just want to mention to the committee that the original study by Urisman and Silverman, as well as the Danielson and Kimata paper, really were suggesting that the XMRV was not found in all prostate cancer, but was in hereditary or familial cases. The Danielson paper actually looked particularly at familial cases, where there were close relatives. So when you think about XMRV and prostate cancer, I think you have to be a little careful and not paint this with a broad brush, and think about, more specifically, the issue of familial or hereditary forms of prostate cancer.

    DR. DEMETRIADES: Coming back to Dr. Nelson's comment, although there is no published association, is it possible that perhaps in situ cancers are very different from more advanced cancers? That's my concern. I think they should be excluded.

    DR. HOLLINGER: Say that again, please.
    DR. DEMETRIADES: It is possible that blood donors with an early in situ cancer -- it's fine, it's safe. But patients with more advanced cancer maybe could transmit the disease.
    DR. HOLLINGER: Dr. Rentas? DR. RENTAS: Thank you. I just wanted to say that, as opposed to CFS patients, I don't think there is a blood collection facility in the U.S. that I know of that doesn't cover cancer as part of their questionnaire already. This is already covered. If you do have a history of cancer, to include prostate cancer, you will be deferred from donating.
    DR. NELSON: Even if successfully treated?
    DR. HOLLINGER: My understanding is that if you are asked the question and you have had prostate cancer and you have been free for a year, off therapy, then you are able to donate. You can donate at that point. If I'm wrong, would the blood banking community let me know?
     

    DR. NELSON: This says medical history. If we agreed with this deferral, that person would be deferred, if they had an in situ and had a prostatectomy.
    DR. HOLLINGER: Simone?
    DR. GLYNN: I just want to remind everyone that the data we have is that there is no association that we know of between prostate cancer and a transfusion history. Neither do we have any data -- it's very little data, but the little data that we have does not support an association between chronic fatigue and a history of transfusion. So you don't have that particular association.

    DR. NELSON: Yes, but here the concern is not so much whether the prostate cancer resulted from a virus transmitted by transfusion, but whether the prostate cancer patient maybe has a superinfection of their tumor and is infectious as a blood donor. We're talking about deferring them. I'm not sure it's the same.

    DR. BIANCO: There are huge studies done in Sweden that have analyzed years of transfusion recipients and any association with cancer, and they have failed to find that association.

    DR. HOLLINGER: I guess the question would be, though, Celso, if some of these studies, for which we'll be waiting on the results -- if you do find that it's positive in the plasma and/or in the PBMCs, then I think you are probably caught at that point in asking the question again or using educational material, whatever. I'm not sure how you get away from that. I'm not talking about causality or anything.

    DR. BIANCO: We will be back here talking about it.
    DR. KLIMAS: Just a comment. In Frank's data -- I can't speak for where the samples came from -- most of these biopsies are time at a time when someone is acutely ill, and so they are going to be more viremic, one would presume. Someone who is post-prostate cancer -- that's what you want. You need that sample, to be able to see if there is any plasma or cellular infection in that sample. I don't hear anyone saying they have that for us.

    DR. HOLLINGER: I think what we have here is that we are waiting for more data. There are some studies that are going to be pursued with that information.

    Let's go on to the third question. The third question is, please comment on the scientific evidence that would be needed to justify a policy of donor testing for infection with MLV-related human retroviruses. In particular, should donor testing be considered in the absence of confirmed disease causation?

    Comments. One thing that I have not seen that I would like to see is testing in other patient groups that are immunosuppressed or even with cancer. We have it in breast tumor. We have a small study there. We have another small study in the respiratory infections and so on. But I have not seen any other studies looking at detection of this virus in immunosuppressed groups. I personally would like to see that kind of information.
    DR. BOWER: There's a negative study in HIV.
    DR. HOLLINGER: That's right, in the HIV group.
    DR. STOYE: In one of the negative studies for CFS, the control group included some transplant patients.
    DR. HOLLINGER: Thank you. I appreciate your telling me that.

    Other comments about what kind of evidence you would need for such a policy in terms of donor testing.
    DR. BOWER: I will just say, at a minimum, it would be consistent evidence of association with the various studies that are going on, especially the collaborative studies, like the Phase III and the other studies that Mike Busch talked about.

    DR. HOLLINGER: So you would need to see transmission. You would need to see infection. So transmission, then some sort of infection or disease causation preceding that --

    DR. BOWER: At a minimum, just an association. I think I would like to see disease causation as well.
    DR. DEMETRIADES: I do not think the existing evidence can support routine testing.
    DR. COFFIN: I completely agree with that. I think what's needed here, really, is one of many different possible kinds of tests. It's hard to say exactly what the test is. But the critical thing is to get some kind of consensus out of it.

    Does you have a test that -- this is what happened with HIV, for example. Soon after it was discovered, it did not take very long before there was a consensus that this was there. As far as the more subtle issues of causality, pathogenesis, it took a much longer time to iron out. But once people knew what to look for, everybody could find the virus.

    We're not at that stage with this yet. We have to be there before we can even begin to move forward to ask questions about causality. I think, as Jonathan said at the XMRV meeting, none of us came into this trying to disprove the idea that this virus -- none of us came into these studies looking for negative associations. But that's the way the data that many people have found have turned out. So we are still lacking the kind of scientific consensus that there is really infection in people, much less what the extent of that infection is, and much less the possibility that it's causing CFS or some other disease. We simply have to start by getting a way to move forward by having consistent assays and consistent studies that just give the same results from one lab to the next. That's still not the case.
    DR. MIKOVITS: (Off-mic)
    DR. COFFIN: It has been a long time. Until your publication, Judy, the field was very, very quiet, I have to say.

    DR. MIKOVITS: In HIV/AIDS, that was five years after the first publication of the Science paper isolation to where groups were agreeing on people who were infected. We didn't have a PCR assay in 1982, at the first publication. So we isolated virus and looked for serology.

    DR. NELSON: It seems to me that one question that probably will be answered, hopefully, in the not-too-distant future is whether or not this disease is transmissible by transfusion, by looking at these repositories. That's one of the values of these repositories. I think that question can be answered. Obviously, if that is negative or if the risk is very low, then that would make it less important to worry about screening donors. But if there is clear evidence of transfusion, then the endpoint of that -- now, it seems possible that some of the people who received transfusions and got infected 20 years ago -- that they might have had an outcome by now. It's possible. They might have had CFS or some sort of outcome from the infection. That's the value of having these cohorts and having samples from the cohorts.

    DR. HOLLINGER: There are several studies which have followed patients -- for example, our study, initially -- which have looked at patients 25 years later that includes the TTV study, as well as the NIH study, as well as the VA study -- in which patients were looked at again 25 or 30 years later. Those samples are available, if you find some positives, to look at those individuals later. There are several studies, Rh studies in Germany, in which females acquire disease. They have been following them for 35 or 40 years or so. I think that would be important to look at as well.

    DR. EPSTEIN: Question 3, if I could perhaps communicate a little bit of what's troubling the FDA -- I agree with Dr. Nelson that the studies that are being launched will probably tell us in the next year or two whether there is transfusion transmission of the virus and infection. But proof of causation may take a lot longer. What we are really trying to get at -- these viruses are sort of in a bad family. When you look at what they do to the mammals that they infect, you have neurologic complications, immunosuppression, and cancers.

    The question is whether that background is sufficiently compelling that we should be precautionary or proactive about donor screening, if transmission is shown. Or should we actually wait? I think the arguments can go two ways. For example, with simian foamy virus, which we discussed with the Blood Products Advisory Committee, we had pretty much the same set of questions: Is it transmitted? Is there disease association? There was one look-back study that didn't show transmission, even though we think transmission might still be possible. But the compelling argument for the BPAC was that there was no apparent disease association.

    Here the issue, I think, is a little bit differently framed. As I say, in other species, there is clearly disease association. So should we be waiting for causation or not? I think that's the perspective we are trying to get clarified. We are trying to anticipate if studies of transmission show transmission, but we still haven't demonstrated disease causation. Do we act or don't we act?

    You know, we can wait and cross that bridge in a year or two. We don't have to answer it today.,br> DR. HOLLINGER: We won't do it tonight.
    DR. NELSON: But wouldn't some of the repository studies -- if the recipients had a transfusion-transmitted infection in 1983 and a higher proportion of them ended up with some sort of outcome, it's conceivable that we might be lucky -- or if they had no outcome. Those data would be useful.

    DR. KLIMAS: Just one clarification. Probably Judy can do this better than I, because she and Dan Peterson reported this very large case series of mantle cell lymphomas in chronic fatigue syndrome patients. It's a kind of tumor that most oncologists don't see in their entire lifetimes, and yet there is this very large series. I know Judy has done a lot of work with that.

    I myself have seen a number of lymphomas in my patient population, usually EBV-associated lymphomas. We don't know what the coinfection roles might be. So I am afraid of the oncogenic potential of this type of virus in the population.

    DR. COFFIN: With all of the models for oncogenesis by these viruses, that would be the best possible scenario, in fact, because it's really easy to show. If what is happening in a patient with a disease like that is the same as what happens in a mouse or a cat or any other model of disease that is caused by these viruses, every single cell and tumor that's going to have a provirus is going to be sitting next to a proto-oncogene. It's a real smoking gun that is easy to detect. So study of diseases like that could provide by far the best evidence in a real hurry, if that was possible.

    I would agree, though, that if, in fact, we accept that the virus is present in some people, in a fraction of people, and not in everybody, and if we have a good, validated assay that there is consensus on that shows that, and if that virus is being transmitted by blood transfusion, I would consider it highly prudent to test for and decline blood that was likely to be contaminated with this virus, even if we don't have proof of causation of this disease. I think there are too many other things it could be doing that are bad to take that chance.

    DR. MIKOVITS: I would just like to comment on Nancy's comment. We have a very large group, 30 percent of the cohort that Dan Peterson has studied across the United States, with both CLL and mantle cell. With one particular individual, we could isolate virus from 1984, before he was diagnosed with CFS in 1988 and 10 years before he was diagnosed with mantle cell in 1999. He succumbed to mantle cell in 2008. We have samples where we can isolate the virus throughout the disease and a cell line where every single cell is infected with the XMRV, and not EBV or any other virus that is known. We have a number of cases like that who have gone on. At least one -- you should have received written testimony from a gentleman who couldn't be here, who developed the disease after surgery and a blood transfusion, a JD/MD who couldn't work any longer. There is mantle cell in his family as well.

    DR. HOLLINGER: Thank you.
    Let's look at the last question, the final question: Assuming that testing is warranted, please comment on the potential utility of nucleic acid tests and/or serologic testing of blood donations to ensure safety of the blood supply from transmission of MLV-related retroviruses.

    Any comments from the group?
    DR. COFFIN: My answer is yes.
    DR. HOLLINGER: And it looks like these tests can be validated, hopefully, in time and would cover a broad spectrum.

    DR. NELSON: This question depends, really, on the sensitivity and cost and so on of these assays. Right now we don't know for sure.
    DR. HOLLINGER: That could reduce the blood supply by anywhere from 1 to 8 percent?
    DR. NELSON: Depending on the number of false positives, I guess.
    DR. HOLLINGER: Any other comments from the committee, besides being tired?
    Jay, do you have anything else that you want from us?

    DR. EPSTEIN: Of course, these technologies will continue to evolve. I think what is concerning the FDA is that when you look at the current state of the art of NAT tests, there is a lot of variability, we seem to be at some limit of detection, and the procedures would require types of sample processing that are not currently feasible in the donor-screening setting. Also there is the possibility that over the course of infection, titers in the blood may go down; the game may just be in the tissues.

    So what we are really trying to think about is, is it a better target to pursue serologic testing? What we have seen so far is that the yield or positivity rate in serologic testing with current, state-of-the-art tests is seemingly lower than NAT in these high-risk cohorts. So what we are really looking for from the committee is, is there a basis to advise us where to invest effort? Should we be trying to facilitate NAT versus serology? Should we just be neutral and try to develop both, recognizing the logistic complexity of NAT.

    DR. HOLLINGER: I think one of the problems with looking at NAT in plasma and PBMCs and so on -- in so many cases, you have to do a lot of manipulations. That's not going to be possible in the blood community -- ultracentrifugation, putting it in culture to amplify it, doing activation. Where you are collecting PBMCs, you are not going to do that, I don't think, in the blood bank. This, to me, is one of the major issues that has to be resolved, as distinct from serology, which is a simple enough test to do. I think that becomes a real issue, and how sensitive that's going to be and which one is going to pick up the most with a good throughput.

    DR. BOWER: Right. Obviously, getting a test that is consistent and one that works is what you need to do. It seems like if you had all three that were working well, serology would probably be the best, because it would be the easiest to do. At least I assume we are looking for chronic infections and we not trying to pick up incident infections for this -- although I don't know, because I don't know exactly what the incident rate is of chronic fatigue syndrome.

    DR. HOLLINGER: Yes, Dr. Coffin?
    DR. COFFIN: My answer wasn't entirely flip. In the case of HIV, if you think about that as sort of a model, both of these have their place. It is well understood that serologic tests don't detect early incident infection, as you point out, and they have high false-negative rates, particularly in the context of widespread screening, requiring repeat testing with other methods. Nucleic acid-based tests are more cumbersome and more expensive to do on a broad basis, but provide a greater level of sensitivity and detection of incident infection.

    I think we don't know how this is going to play out until we understand much more about the biology of the virus, if there is a biology of the virus. So we really can't answer that question yet. I think this has to be pushed forward on all possible fronts in these areas, and perhaps even others, if somebody can think of another kind of test, until we know we are really beginning to get somewhere and we can begin to bore down with that particular set of studies.

    The two companies that presented, Abbott and Gen-Probe, are to be applauded for the amount of resources that they are putting into this. They are taking a real risk, but they have the potential of bringing us the best possible assays to be used for this kind of screening.

    But we still don't know where we're going. We're casting around in the dark right now.

    DR. HOLLINGER: Thank you.
    I don't see any other burning comments from the committee here. I do want to thank both the audience and the committee for your patience. This has been a long day, and we are very appreciative of your staying around. I think we have learned a lot about this disease.

    Tomorrow morning, 8:00, we'll get started. Hopefully we'll finish on time tomorrow.

    (Whereupon, at 7:15 p.m., the meeting was recessed, to reconvene the following day at 8:00 a.m.)

    Table of Contents