Blood Donor Suitability Workshop
UNITED STATES OF AMERICA
DEPARTMENT OF HEALTH AND HUMAN SERVICES
PUBLIC HEALTH SERVICE
FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION & RESEARCH
BLOOD DONOR SUITABILITY WORKSHOP
MONDAY, NOVEMBER 23, 1998
The workshop took place in Conference Rooms D and E, Parklawn Building, 5600 Fishers Lane, Rockville, Maryland, at 8: 30 a. m., Andrew Dayton, M. D., Ph. D., Chairman, presiding.
PRESENT:
ANDREW DAYTON, M. D., Ph. D. Chairman
CELSO BIANCO, M. D. Speaker
MICHAEL P. BUSCH, M. D., Ph. D. Speaker
KEN CLARK, M. D. Speaker
LYNDA DOLL, Ph. D. Speaker
SIMONE GLYNN, M. D., MPH Speaker
HAROLD JAFFE, M. D. Speaker
BERNARD POIESZ, M. D. Speaker
SUE PRESTON, Ph. D. Speaker
TOBY SIMON, M. D. Speaker
RICHARD STEKETEE, M. D., MPH Speaker
SUSAN STRAMER, Ph. D. Speaker
GEORGE SCHREIBER, D. Sc. Speaker
ALAN WILLIAMS, Ph. D. Speaker
IAN WILLIAMS, Ph. D. Speaker
THOMAS ZUCK, M. D., FRCP (Edin) Speaker
ALSO PRESENT:
DAVID FEIGAL, M. D.
Opening Remarks
--Harold Jaffe; Introduction of Retro-viruses into Human Populations: A Model for Emerging Pathogens
Prevalence and Incidence of HIV, HBV, HCV, and HTLV in Men Who Have Sex with Men (MSM), Sex Workers (SW), and Intravenous Drug Abusers (IVDU)
--Rick Steketee; Prevalence and Incidence of HIV in MSM, SW, and IVDU
--Bernie Poiesz; Prevalence and Incidence of HTLV in High-Risk Behavior Groups
Prevalence, Incidence, and Risk Factors in Blood and Plasma Donors
--Toby Simon; Prevalence and Incidence of HIV, HBV, HCV, in Plasma Donors
--Lynda Doll; Estimates of New Blood Donors if Eligibility Criteria Change
--Ken Clark; Risk Factors in Blood Donors Positive for HIV
--Simone Glynn; Risk Factors in Blood Donors Positive for HCV
--George Schreiber; Risk Factors for HTLV Positive Blood Donors
Questionnaire Behavioral Issues
--Celso Bianco; Self-Identification of Deferral Risk
--Thomas Zuck; Iterative Questionnaire
Testing Issues
--Sue Preston; PCR Testing: Narrowing of the Window Period
PROCEEDINGS
(8: 40 a. m.)
DR. FEIGAL: Good morning. Maybe we could get started. I'd like to welcome you to FDA's Workshop on Blood Donor Suitability.
I'm David Feigal. I'm the Deputy Medical Director at the Center for Biologics.
And one of the more important responsibilities --one of the responsibilities actually recognized in the last revision of the Public Health Service Act in 1944 is our responsibility for assuring the quality and the safety of the blood supply.
Today's workshop is intended to gather scientific information to assist the FDA and the Department of Health and Human Services in efforts to update and revise blood regulations on donor suitability.
It has only been about two decades since we began explicitly asking donors to self-identify or began looking at the kinds of factors that might be risk factors for transmitting infectious diseases.
And since that time, much has changed, both in our knowledge of the epidemiology, in the emergence of infections that we were unable to even test for two decades ago, and in our knowledge of the transmission.
The way that the process works is that regulations are promulgated in the Code of Federal Regulations. Fans of the CFR know these as, in the book of numbers, as Sections 21 CFR 610 and 640.
And when we propose a change in regulations, the process which we go through to do that is to first carefully, and in consultation with advisory committees and with workshops and with our partners in the public health service, including the Center for Disease Control and the National Institutes of Health, develop the scientific basis for updating the regulations.
Today is part of that process for taking a look at these specific regulations. Another part of the process which moves more quickly than changing the regulations is to use guidance documents.
These, in the past, have had various names --Points to Consider, Blood Memorandum -- although we unified all of the ways that we deliver guidance through a procedure we call Good Guidance Procedures. And so now all of these different vehicles are called guidances. And we are able, through this process, to actually also communicate important information.
Several of the exclusionary criteria we're discussing today were initially issued as guidance documents. And one of the questions, as we go through the updating process, is since many things that come out that are done initially are done through guidance and they don't have the binding force of a regulation, when is it appropriate to turn the guidance into regulation so that that standard is enforceable, since the regulations are our interpretation of the Public Health Service Act and sometimes the Food, Drug and Cosmetic Act?
More broadly, I guess, today we're looking at the very first parts of the multiple layers in the safety net of the blood supply. This step that we're looking at today first begins with providing educational material, screening donors by asking the donors questions about their health and risk factors.
This means that trained personnel need to be able to interview the donors and help determine if that's a suitable donor, and find out if potential donors should exclude themselves.
The FDA recommendations and regulations to exclude potentially infective donors have expanded over the years as we've sought to exclude for risk factors for hepatitis B and HIV, but also included such viral variants as --and looked for the donor exclusion questions that would also identify high risk for HIV Group O and the thorny issue of the theoretical risks for diseases such as Creuzfeldt-Jakob disease.
The second part, after donation, is that blood is tested for blood-borne agents, including HIV, HBV, hepatitis C and HTLV I and II. This, in fact, gives us feedback in terms of how successful we are in some of the donor exclusions and provides some of the scientific basis for identifying our success in this process.
The difficulty and the reason why testing cannot completely replace donor exclusion is because of the difficult issue of window periods -- that time when someone is infectious but you cannot yet detect it with your blood screening tests.
Today we'll hear scientific information on the risk of transmission of HIV, hepatitis B, hepatitis C, the HTLVs, and emerging infectious diseases, in categories that have been identified for exclusion in the past --men who have had sex with another man even one time since 1977, men or women who have exchanged sex for money or drugs since 1977, and men or women who have abused intravenous drugs. We will also hear presentations on the risk to partners of such individuals.
The underlying question that we're grappling with in looking at our current guidance and regulations is whether the FDA should maintain the lifetime exclusion for these individuals that have been described as being involved in these activities.
And also at issue is the rationale of deferring sexual partners for such persons for only 12 months.
We will begin and hear the epidemiology on the introduction of retroviruses into human populations. We'll hear information on the incidence and prevalence of HIV, hepatitis, and HTLV in individuals who engage in activities thought to be at high risk for infection.
We will hear a presentation on the prevalence and incidence of blood and plasma donors and the impact of the donor deferral criteria on blood safety. And we will consider the advances in donor testing and narrowing the window period by the introduction of investigational genetic tests for HCV and HIV.
We'll also consider a model to assess the impact of changes for these donors. The challenge before us is to maintain safety and availability of blood plasma and products and balance the enthusiasm, based on improvements gained by advances in test technologies, with due caution based on the past unfortunate experiences of being unable to stop disease transmission with the methods of those days.
I'd like to just conclude by welcoming you all. I think that it's a testimony to how interesting and important this topic is that we have such a good turn out and such broad representation at this time of year during a holiday week.
And let me introduce Dr. Andy Dayton, who will also make some introductory remarks. Return to Agenda
DR. DAYTON: Good morning. Thank you all for being here, and welcome to the Donor Suitability Workshop.
I think you've had a very good introduction as to what our scientific questions are today, and all I'm going to do is just remind all of us of the theoretical framework in which the FDA tends to look at deferral issues.
This is what we're trying to prevent, obviously, infection getting from potential donors into the blood supply. Our main weapon for this is, of course, tests for infectious agents.
Could I have the next overhead, Martin?
Okay. So we have tests to prevent bad things from getting into the blood supply. But tests are imperfect, and here are ways that infections get into the blood supply, bypassing tests. We essentially have prevalence issues.
In this case, it would be undetectable strains of a pathogen which the current tests don't recognize. In this general category would also, of course, come emerging pathogens which have not -- for which there aren't good tests --blood bank errors --these are very rare. And general failure rate of the test depends on the test. Certainly, for something like HIV, this is essentially zero.
And then we have incidence issues by which infectious agents can bypass the tests, and this is the window period that Dr. Feigal referred to.
Martin, can you move that up a little bit now?
So that we can actually --in an ideal world, we can actually calculate the total number of infectious slipups, total number of times we get an infected unit slipping into the blood supply.
And it merely equals the number of blood donors times the prevalence times the summation of these various errors for the prevalence issues. And for incidence issues, it's blood donors times essentially an incidence factor, which is described here.
Can I have the next overhead? Okay. That's good.
Now, how do we --is this perfect?
Well, no, this isn't perfect. Things can get around the questionnaire as well as getting around the tests. If society is well educated, we have a large number of self-deferrals, which is good.
The questionnaire which we have designed to block the people from --infected people from actually becoming potential donors can also be bypassed not only by self-deferral, but there are ways in which the questionnaire can fail.
And these are very difficult issues to pin down. For instance, ineffective risk identification. If we have not appropriately identified a certain risk category, those people will, of course, get through the questionnaire and get to the testing stage.
Test-seeking behavior --you see this in people who show up to the blood donation centers because they know they're going to be tested. And even if they know they're in a high-risk category and they're not supposed to show up, they appear because they know they can get a test.
Sometimes there is resentment. And it's a very easy thing to understand how people can feel resentful towards being told that they're not appropriate for giving blood.
Peer pressure --a group of people all decide to give blood at, let's say, some kind of community organization, like a church, and peer pressure can induce people to give inaccurate answers on the questionnaire.
Misunderstanding of questions. If it's a poorly designed questionnaire and somebody doesn't understand what's being asked, you can have people inappropriately getting through the questionnaire.
And this is a --we're not going to discuss the questionnaire issues very much today, but it's a very significant problem, because the questionnaire is getting fairly long and there's a lot of interest in subcategories in high-risk behavior, to see if we can factor out lower-risk subcategories of what we currently consider high-risk behavior. And that can give you a problem in making a very complicated questionnaire.
Could I have the next overhead, Martin?
So how do we approach this? Well, there are two ways. I really should say prospective approach, or forward approach, for the first way of doing it. And that's to determine all of the numbers that feed into that model that I just showed you, and we will be discussing that data today.
We want to know the prevalence, the incidence, and how that factors out according to risk behavior. We want to know the size of the behavior categories, and that's important in determining the number of blood donors that we have from that category.
We would like to know blood bank error rate. We don't have a lot of good data on that. We would like to be able, of course, to quantitate undetected strains. Easier said than done. We would certainly like to know accurately the assay failure rate for other reasons, and I think we probably have reasonable data on this in most cases.
And we really want to know what is the behavior of the various risk groups in terms of self-deferral and questionnaire behavior. And this is actually a very complicated question and a very major question because --and it will be addressed today in several talks.
But when you don't know how many people in a certain risk category are going to correctly answer the questionnaire, or how many are going to self-defer, it makes it very difficult to estimate what the risks are of having that particular behavior group donating blood.
Could I have the next overhead?
And then there's the retrospective approach in which we look at failures and determine their sources. And we'll see data on this today, too. The typical example of this would be to take case histories of post-transfusion episodes. Now, this is particularly important for emerging pathogens in early stages of epidemics when there aren't good tests.
More relevant to what we're doing today is identifying and categorizing the risks associated with the units that test positive. You can consider this basically as a reality test for the calculations that we would have made with the data I just --the kind of data I just indicated we look for. Or you can consider it as the truest assessment of direct threats to the blood supply --in other words, residual risk. This is what the blood testing centers and the tests actually see.
The next overhead, please, Martin.
Now, I'm not going to dwell at this on length, but in certain special cases pertaining to changes in policy, which is often what we're faced with, sometimes a modified approach can be pursued to calculate the effects of policy changes. For those of you who, about a year ago, came to the MSM presentation of the BPAC about a year ago, this is what we approached that issue with.
If a deferral policy is already considered adequately safe, and that's a big if -- but if it is considered adequately safe and there is perhaps a desire to change the policy, for example, from a highly restrictive policy such as lifetime deferral to perhaps a less restrictive policy such as a one-year deferral, one can ignore the bypassing of the questionnaire issues, which, as I said, is a very difficult thing to calculate.
And the reason you can do this is because anybody who's already bypassing the questionnaire would be unaffected by the enlarged inclusion categories --in other words, the narrowed exclusion categories.
Or another way of saying that is they're already getting to the testing stage under the current policy, and they'll still get through the questionnaire after the new policy, whether they intend to be newly included or not.
So in situations like this, one can then assess the effects of changes in policy simply by appropriately multiplying the prevalence and incidence rates in those first equations I showed you by the expected donation rates in a high-risk behavior category, and the size of the newly- included category, to estimate new challenges to the testing step. And I won't dwell on this further.
Let me just sum up in the last --so there are many different aspects of a transfusion- transmitted disease that must be understood in order to understand its risk to the blood supply and to appropriately formulate a policy deferral --or deferral policies.
There are different approaches to estimating risks, and they're often complementary and are rarely mutually exclusive. We seldom have all the numbers we need to perfectly estimate risk. It's not a perfect world.
Because of these indeterminacies, we must build redundancy into the system. And you can see that, and that's basically why we have both test and questionnaires.
We must always consider --and here I haven't discussed this at all, but it will be the subject of our first talk. We must always consider the great unknown of emerging, poorly understood, poorly characterized pathogens, because we can't pick them up with the questionnaires always and we can't pick them up with the tests always.
And we feel --the FDA feels very strongly that the public and Congress have made it clear that they desire a zero error tolerance policy with respect to the blood supply. In other words, the health of the recipient of the blood or the blood products must always be our primary concern. So with this very brief overview, I want to thank you for your attention.
And let me introduce our first speaker, Harold Jaffe, who will talk on the introduction of retroviruses into human populations, a model for emerging pathogens. Return to Agenda
DR. JAFFE: Good morning. I'd like to thank Dr. Dayton and the FDA for inviting me to be with you.
While I'm not entirely sure how my topic is going to relate to the rest of the meeting, I was very pleased to see the term "emerging pathogens" used because it makes my bosses at CDC smile.
What I'm going to try to do is examine the epidemiology of retroviruses that are known to infect humans as a model for emerging blood-borne pathogens. And while, clearly, HIV 1 is the most important and well understood of these, I want to compare and contrast HIV 1 with some of its retroviral relatives.
Could somebody turn the first slide on?
Okay. Well these are the subfamilies of the retroviruses that we're concerned about, and they include, of course, the lentiviruses. Now I can't see. I don't need to see, do I?
The lentiviruses, HIV 1 and 2, and their simian counterpart --SIV --which, as I'll point out, has actually infected humans; the oncoviruses, HTLV I and II; and the spumaviruses, which are also known as foamy viruses.
For each of these subfamilies, we can really ask the same questions. We can ask: where did the virus come from? When was it introduced into humans? Once it was introduced, did it spread? If it did spread, what were its transmission routes?
And once it spread, did it establish itself in any particular populations?
Then, based on the answers to these questions, I want to see if we can draw any general conclusions about the likely epidemiology of a new blood-borne agent.
Let's just start with some very basic information that I'm sure you mostly know about HIV 1.
As I'll illustrate in a moment, the closest relative of HIV 1 among the non-human primates is the chimpanzee simian immunodeficiency virus, SIVcpz. As I'll also try to illustrate in a moment, although we don't know exactly when HIV 1 first occurred in humans, it was probably on the order of about 50 years ago, although the global spread clearly didn't occur until later than that. We all know that it causes AIDS, and we all know its basic routes of transmission. A major question for this meeting, of course, is: which group should be considered at highest risk for these various infections?
For purposes of AIDS and HIV surveillance in the United States, these are the categories that CDC considers to be exposure groups for HIV 1 and, of course, they include homo and bisexual men, injecting drug users, persons with hemophilia, transfusion recipients, and the group reporting specific heterosexual contact with an HIV infected person, or someone known to be at increased risk for HIV.
Now I want to look at some of these points in a little bit more detail, the first question being: where did HIV 1 come from?
And how does this thing work?
Okay. This is a phylogenetic tree, which you probably can't see. And if you could see it, you probably wouldn't be able to figure it out. But I'll try to point out some of the important points.
This is a tree that was published just a few months ago by Simon & Associates. And what it does is compare the genetic sequences in the envelope region of HIV 1 and the chimpanzee lentiviruses. You can disregard this part down here which deals with HIV 2.
The point it makes is that, first of all, we can see three groups of HIV 1 viruses --the Group M, O, and N. Group M is, of course, the major group. It's the one that's responsible for the global pandemic. And it includes a number of subgroups lettered A to J, some of which are shown here.
This appearance is called a star phylogeny by the people who work in this area. And what they say is the star phylogeny suggests a single introduction of an ancestral virus that then evolved into these many subtypes.
Now, of course, Group M is the predominant subgroup of HIV 1 in the world, and it's the one that's really responsible for the global pandemic, but there are several other groups as well. The Group O viruses, which are shown over here, were first reported in 1990.
They're genetically quite distinct from the Group M viruses. They're found mainly in Cameroon and adjacent countries in Africa, although two African patients have been reported with Group O infections in the United States.
Finally, in the article that I just mentioned by Simon, the authors describe a new subgroup, Group N, which is represented by a single isolate again obtained in Cameroon from a person with an AIDS-like illness. And this is thought to be a prototype for this new group.
There are also two chimpanzee viruses shown up here, CPZant and CPZgab, which represent viruses from Zaire and Gabon, respectively. The genetic distances on this tree are indicated by the branch lengths. And you can see that the Group M and the Group O viruses are not particularly close to these chimpanzee viruses. But the Group N actually is quite close to this virus from a chimpanzee in Gabon, and it appears likely that these two are highly related.
I think most people in this field believe that there were separate introductions of ancestral viruses, most likely from chimpanzees, that resulted in these three groups of HIV 1.
Now, if it's true that each of these HIV 1 groups has its own ancestor, when were these ancestors introduced into human populations? The only group that we really have much information for is the Group M, the predominant virus in the world. And this comes from a study that was done by David Ho & Associates in which they were able to look at a plasma sample that had been collected in 1959 from what was then known as the Belgian Congo and were able to obtain at least a fragmentary genetic sequence of a virus in that sample.
It's shown here in yellow. And the point is that this sequence seems to be very close to the hypothetical ancestral strain from which the subtypes D, B, and F viruses were derived.
Based on what's known about the evolutionary rate of HIV 1, these authors suggest that the Group M viruses probably shared a common ancestor, perhaps in the 1940s or the early 1950s.
Now what happened after these viruses were introduced into the human population isn't really known. I think most likely the viruses did spread relatively slowly in parts of sub-Saharan Africa for a number of years, and it's possible that the spread then accelerated with the social disruption and population movements that occurred following the end of colonial rule in many of these countries in the 1960s. In retrospect, there probably were clinical cases of AIDS in some African cities by the mid 1970s.
How and when the virus entered the United States is also not known. In collaborative studies that CDC conducted in San Francisco, we found that, looking at serum samples that had been collected from gay male STD patients in 1978, about five percent were seropositive.
It would be nice if we had comparable data from injecting drug users and other groups at that time. Unfortunately, we don't. One way we might gain some insight, though, into the very early spread of HIV in the United States and the resulting AIDS cases is simply to look at this chronology of the first reported AIDS diagnosis in these various exposure groups.
This is based on CDC's surveillance data. I've excluded a couple of very early cases that seem pretty questionable. But it's interesting to see that, in retrospect, the first case of what we now call AIDS that was diagnosed in a gay man actually was in 1977, which was four years before the epidemic was recognized.
Two years later, we had the first case in an infant born to an at-risk mother and in a transfusion recipient; in 1980, the first case in an injecting drug user; and, in 1981, the first case in a hemophilic and in a heterosexual contact.
Again, I wouldn't take this chronology too literally, but I think it would at least give us some idea, or a rough idea, of how the virus was spreading in the early years in the United States.
The story of HIV 2, I think, bears many similarities to HIV 1, but there are some important differences that I want to try to emphasize. Like HIV 1, we think that HIV 2 as derived from a non- human primate --in this case, the simian immunodeficiency viruses that affect sooty mangabeys.
We don't know when this crossover happened. The first documented infections, in retrospect, in humans were in specimens collected in West Africa in the 1960s, but the virus certainly could have been there before then.
The geographic distribution of HIV 2 -- we know that it's by far the most common in West African countries and in several of the former Portuguese colonies in Angola and in Mozambique, but certainly has not had the same kind of worldwide spread that we've seen for HIV 1.
We know that HIV 2 causes AIDS, but the rate of disease progression is certainly lower than what we see for HIV 1. And while the roots are the same, as I'll point out in a moment, the rates of HIV transmission by these routes are substantially lower.
Within the United States, the only group that can be considered to be at increased risk for HIV 1, at least right now, would be persons born in certain West African countries.
Now, trying to look at these points in a little bit more detail --again, this is a phylogenetic tree, which is certainly confusing.
But it's also kind of interesting, and I'll just try to point out the main points it's trying to make.
This comes from Beatrice Hahn & Associates, and it looks at a series of subtypes of HIV 2 virus, AID F, shown here. The HIV 2 strains are all shown in white. And simian strains, particularly from sooty mangabeys, are all shown here in yellow.
The important point here is the genetic relationship between the human virus HIV 2 and the simian viruses is very, very close. It's much closer than what I showed you previously for HIV 1 and the chimpanzee viruses. In fact, the relationship is so close that we can use HIV 2 antibody tests to detect these simian infections. Beatrice Hahn has suggested that each of the HIV 2 subtypes that are shown on this slide probably represent a separate introduction of an ancestral SIV strain into a human population.
The differences in the rates of HIV 2 transmission compared to HIV 1 are really very striking. This slide, for example, looks at the rates of perinatal transmission of the two viruses in three studies, two of them from West Africa and one in France.
You can see, as you would expect, the HIV 1 transmission rates, in instances where the mother has not been treated, between about 20 and 25 percent; but, for HIV 2, between about zero and one percent.
We can also see differences in the sexual transmission of HIV 2 versus HIV 1 in this slide which comes from a study done by my colleague, Kevin DeCock, while he was working in Abidjan in Côte D'Ivoire.
This study looks at the infection rates of HIV 1 and 2 in childbearing women. You can see, for HIV 1, in the blue bars, that over the period observed --I think from 1988 to '92 --HIV 2 seroprevalence increased from about five percent to about nine to ten percent. But during that same time, the HIV 2 prevalence actually decreased from about two and a half to one and a half percent. So in the same populations, the two viruses are actually behaving rather differently.
The reason for the lower transmission rate of HIV 2 is not entirely clear, but Kevin DeCock has suggested that a major factor explaining this might be the lower concentrations of virus found in the blood of HIV 2 infected people, especially during the early phases of infection.
This slide examines virus isolation rate from peripheral blood mononuclear cells stratified by CD4 count. You can see for HIV 1 high rates of virus isolation from anywhere from high to low CD4 counts, but that's not the case with HIV 2.
In the relatively immunocompetent HIV 2 infected patient, the virus isolation rate is quite low. It would be nice to be able to confirm these findings with plasma HIV 2 measurements, but reagents for these tests are just now being developed.
The lower transmission rate of HIV 2 I think can help us understand why the sexual spread of HIV 2 has been much more limited than HIV 1. The spread of any infection can be described by a term which is called the "basic reproductive rate," or BRR, of an infectious disease, which is simply the average number of secondary cases generated by a primary case.
If this rate falls below one, an epidemic cannot be sustained. For a sexually transmitted infection, BRR depends on three factors: the rate of partner change, the duration of infectiousness, and the transmissibility of the agent.
So even if HIV 2 infected persons have just as many sex partners as an HIV 1 infected person, and even if they remain infectious for their lives, the lower transmissibility of HIV 2 will limit its spread.
I think we get a good example by looking at the CDC surveillance data for HIV 2 infections in the United States through June of 1988, at which point we knew of 79 HIV 2 infected people in this country. Of these, 52 were persons known to be born in West Africa. There were another 15 whose birthplace was unknown, but four of these had malaria serology profiles, suggesting a West African residence.
So unlike HIV 1, there has not been a major HIV 2 epidemic in this country. And groups identified to be at increased risk for HIV 1 have not necessarily been at increased risk for HIV 2.
Finally, before leaving the subject of HIV 2 entirely, I just want to mention a case report by Rema Khabbaz and her associates at CDC of SIVhu, "hu" standing for human infection. The index case here that was published a couple of years ago was a laboratory worker who handled clinical specimens from SIV infected macaques.
The worker was found to be seropositive using HIV 2 antibody tests, but sequencing the virus infecting this worker revealed that the virus was actually an SIV which appeared to be highly related to the virus of sooty mangabeys that was being studied in this laboratory.
To date, this worker has not become ill, and the worker's steady sexual partner is not infected. This occupationally acquired infection may, therefore, be a contemporary model for what happened in the past when sooty mangabey viruses were introduced into humans, and subsequently adapted and evolved into what we now recognize as HIV 2.
Let's now shift to the second subfamily, the oncoviruses, and begin with HTLV I. Like the viruses that we've already described, it, too, has a relative among the viruses of non-human primates -- in this, case STLV I --which is widely distributed among these animals.
Unlike HIV 1, it's believed that HTLV I entered the human population many thousands of years ago, and since then spread to most parts of the world. And, of course, unlike the lentiviruses, these viruses do not cause immunodeficiency diseases; rather, cause a malignancy --adult T-cell leukemia, lymphoma, and a neurologic disease known as HIV 1 associated myelopathy or tropical spastic paraparesis. Again, the same transmission route --sexual, parenteral, and perinatal. But, as I'll point out, the transmission rates are certainly lower than what we've described for HIV 1.
The highest prevalences of HTLV I in this country are seen in persons born in Japan and in the Caribbean and in injecting drug users, although most HTLV infections in injecting drug users in this country turn out to be HTLV II.
As I just mentioned, HTLV I is clearly less transmissible than HIV, and one can see that from a number of studies. Some of them I've tried to summarize for you here.
For example, in looking at children born to infected mothers in the absence of breast- feeding, we see the transmission rate again for HIV 1 above 20 percent, and about five percent for HTLV I; for transfused blood, about 90 percent for HIV 1; and a number of studies for HTLV I, rates between 13 and 64 percent, which seem to depend on the concentration of lymphocytes in different blood products and the storage conditions.
The importance of the very strong cell association of HTLV I is seen even more dramatically when we look at studies of recipients of non-viral inactivated clotting factor concentrates.
In a study that was done in 1988 of about 200 U. S. hemophilic patients, we can see that almost 80 percent of them were infected with HIV 1, which, of course, is present both in plasma and in infected cells, versus zero percent for HTLV I, reflecting the lack of infectious virus in the source plasma used to manufacture these clotting factors.
While the studies that have been done in the endemic parts of the world, particularly the Caribbean and Japan, do demonstrate the sexual transmission of HTLV I, again, the transmission rates are considerably lower than what we know about for HIV 1.
For example, U. S. studies of HTLV I have shown a striking lack of infection in homosexual men. The example shown here was a study done in the late 1980s by investigators from the National Cancer Institute looking at HTLV I infection rates in homosexual men in major U. S. cities in which HIV 1 infection rates were very high.
But yet, for HTLV I, we see the virtual absence of infection --one out of 1,200 in Los Angeles; zero out of 300 in these other parts of the United States.
Now, why this is the case is not entirely clear. Perhaps there's been relatively little interaction between these men and others at high risk for infection such as injecting drug users. But thinking back to our discussion of the basic reproductive rate, it may be that the lower transmissibility of HTLV I through sexual contact has not allowed an epidemic to be generated in this particular population group.
Whatever the reason, the important point is that groups at increased risk for one retroviral infection are not necessarily at risk for all retroviral infections, despite the similar transmission routes.
HTLV II has been studied less extensively, but also appears to have derived from a simian virus, STLV II. Again, it was thought to have been introduced into humans thousands of years ago, and it's found mainly in this part of the world in Indian tribes for both North and South America. It has also been reported to be endemic in certain pygmy tribes in Central Africa.
Although the virus was first isolated from a patient with hairy cell leukemia, the disease associations in humans are not well established. Similar transmission routes, as we've talked about before, in the highest prevalence in the United States for HTLV II in injecting drug users and some North American Indian tribes.
Finally, I just want to mention something that you may not have heard so much about, a more recent infection introduced into humans, and that is simian foamy virus infections in human populations. These viruses are known to be quite common in a wide variety of non-human primates, but there really is not good evidence for an endemic human foamy virus.
The virus infections in humans that we know about are largely the result of cases in which workers have been occupationally exposed through their work with non-human primates, their viruses, or other laboratory specimens.
This slide summarizes a CDC study in which about 230 persons who worked with non-human primates were tested for antibody to foamy virus. And four, or about two percent, were found to be seropositive. Subsequent genetic sequence analysis showed that one of these workers was infected with a foamy virus from an African green monkey, and three others with baboon viruses.
All of these workers appear to be well. And the three spouses that were studied, all of them are seronegative. It's tempting to speculate that these represent dead-end infections. That is, infections that, although they were transmitted from primates to humans, will not be transmitted from one human to another.
However, we know one of these individuals did donate blood, and we know of a more recent case who was also a regular blood donor, and we're hoping to initiate look-back investigations of their recipients.
To try to conclude, then, let's look at some of the lessons that might be learned by examining the introduction and the spread, or lack of spread, of retroviral infections into humans. First of all, these infections appear to have originated in non-human primates. Second, cross species transmission of the oncoviruses probably occurred thousands of years ago, while the lentiviruses were introduced much more recently.
The nature of the contact between human and non-human primates that resulted in these transmissions is not known, but the contemporary examples illustrate how occupational exposure has introduced SIV and foamy virus infection into humans.
Third, once these viruses were introduced into the human population, even though they spread through the same routes, their rates of transmission are substantially different, probably related to biologic differences in the virus, such as the degree to which they're cell associated and their ability to grow or not grow to high concentrations in human tissues, which presumably reflects how well they've adapted to the human host.
And finally, looking in the United States, so-called risk groups for these infections vary considerably depending on the virus that we're talking about, and not all risk groups are the same.
Presumably, the spread of viruses into these groups resulted from some combination of factors, including the geographic and temporal proximity of these groups to the source of the virus, the interaction between persons in these groups and other infected people, and risk behaviors in these groups.
Now, what I've told you about these retroviruses may or may not apply to other emerging blood-borne infections, but I think there is one lesson that does apply overall, which is, it's a jungle and we need to be careful out there.
(Laughter.)
And I want to thank my son for downloading that from the Internet.
(Applause.)
DR. DAYTON: At this point, we'd be very happy to welcome questions on any of the talks so far. If anybody has any questions or comments, please raise a hand, go to a microphone.
Jay?
DR. EPSTEIN: Harold, I think you raised the most intriguing question, which is that risk groups for one infection may not be risk groups for another infection. And I wonder if you could turn it around and just comment on what one can do as opposed to what one can't do.
Are there commonalities that we should worry about --for example, STDs?
DR. JAFFE: Well, if we look at all the viruses that we do know about, all the retroviruses --I mean, one common theme clearly is blood exposure --that both the oncoviruses and lentiviruses have established themselves who are exposed to blood, for example, by needle sharing.
For sexual transmission, I don't actually see that the link has been made. I don't know who's talking about HTLV I, but, as far as I can tell from what I reviewed, HTLV I has really not established itself, for example, in gay men in the United States, which I find quite odd since it is sexually transmitted in endemic areas.
It has certainly been around a long time. There certainly is some interaction between injecting drug users and gay men, and yet we just don't see that gay men in this country have increased prevalence of HTLV I.
At least I'm not --if that's wrong, I'd like to be corrected.
MR. DODD: Thanks. Roger Dodd from the Red Cross.
Actually, Jay, my favorite example is an infection which may or may not be transmissible by transfusion, but it's human granulocytic Ehrlichosis. And in The New England Journal, in a particular study, the greatest risk group that was identified for being infected with this agent was having a lousy golf score because people went into the woods to collect their balls.
This didn't apply to women who were too smart to go chasing after lost balls.
(Laughter.)
But I raise the point because it speaks directly to the issue that you raised, Jay, that we don't necessarily have to use retrovirus as a model for all future potentially transfusion transmissible agents. And I know that muddies the water, but I think it's an interesting point.
DR. RUTA: Hi. Martin Ruta, FDA.
Dr. Jaffe, I was wondering if you could describe some of the surveillance mechanisms that exist within PHS and our ability to detect either variants or emerging agents that might pose potential threats to the blood supply.
DR. JAFFE: I can at least describe some of the things that we're doing at CDC. I can't speak for the rest of the PHS. I guess the simplest thing we do, and it has actually been fairly productive, is that when clinicians are aware of oddball cases --people who appear to have AIDS or an AIDS-like illness and have either "funny serologies" or are seronegative --we often get calls and we often receive those samples. So we do have a chance to look at them.
We also do look at persons reported with AIDS who were born in Africa and residing in this country, just thinking that so many of the subtypes are present in Africa that, if something unusual were to pop up, maybe we would find it that way.
We have more formal surveillance going on in a number of countries overseas, again emphasizing Africa, where we're trying to use testing algorithms that are not necessarily subtype specific.
For example, we've used more generic techniques --for example, the AMP RT method --to look at persons with AIDS-like illnesses who test negative using conventional serologies but with a test that would detect really any retrovirus.
So we do have a number of systems in place. At the same time, I would be the last one to believe that that system is foolproof and that, if new viruses were introduced into this country and were not causing obvious disease, or were not causing it for a number of years, I don't think we have a system in place that would find it.
DR. BIANCO: Celso Bianco, New York Blood Center.
Harold, what is very interesting in your presentation is that you showed that the variants that you see in retroviruses, in general, are less virulent or less transmissible than the predominant forms. Make you almost suspect that, by selection,that the most virulent are the ones that are succeeding in the pandemic.
But is it applicable --should we assume that, because we were using before always the model of the resistant bacterium, that in a certain way we would not have the means here to diagnose there to treat with antibiotic?
Is that the model that we should use?
That is, that the variant will be the most virulent, or the least virulent, or you can't make --
DR. JAFFE: I think it would be hard to generalize. I mean, clearly, among the retroviruses that are established in humans, HIV 1 is the most virulent and probably was the most recently introduced.
So, you could look at that and say, well, that's the one that maybe is the least well adapted to humans, or the human host has not been able to develop an immune response that's protective.
On the other hand, the foamy viruses that we know about that have just been --presumably have not been introduced into humans in the past -- at least we have no evidence for it --in the small number of people who have been studied, don't seem to cause any disease at all.
So again, I think it would be hard to generalize.
DR. DAYTON: Okay. If there are no more questions, let's proceed to the next speaker.
We're now going to have a talk from Ian Williams on prevalence and incidence of HBV and HCV in various high-risk groups. Return to Agenda
DR. IAN WILLIAMS: Thank you very much.
There's a lot more people here than I expected. I brought some handouts, but they're definitely not going to go all the way to the back. So I guess I'll start in the front, and we'll run out about a third of the way back.
It's my pleasure to be here this morning. I probably have one of the more difficult talks to give this morning due to, really, the paucity of data. So I'm going to do what I can to present the data that's out there and suggest limitations, where appropriate, and hazard some guesses where I think those are also appropriate. I thought it would be important, to sort of put this all in context, to start from the general and work to the specific. What do we know about the general U. S. population in terms of hepatitis B?
And actually, this is a very nice study that's going to be published this January in the American Journal of Public Health by Geri McQuillan and her colleagues at the National Center for Health Statistics, in conjunction with the folks at CDC.
And basically, this data comes from the Third National Health and Nutrition Survey. And essentially, this is a population-based cluster sample that seeks to make estimates about a number of health and nutrition outcomes for the entire U. S. population as a whole.
And I'll get right to the bottom line. What did they find? The bottom line is they found that roughly five percent of the general U. S. population has ever been infected with hepatitis B. And when they broke it down and looked at its certain population subgroups --and again, this is a study that's not set up to look specifically at blood-borne pathogens, but to look at other health and nutrition outcomes.
When they looked at it and stratified it by the ways they were able to, they basically found that rates of hepatitis B virus infection varied quite a bit depending on what population subgroup you looked at. If you looked among non-Hispanic whites, they found rates of about two and a half percent.
If you look among non-Hispanic blacks, you saw rates of about 12 percent. And if you looked among Mexican-Americans, you saw rates of about four and a half percent. So there's quite a bit of variability based on who you look at.
And on this slide, I don't present data on those that are chronically infected, but if you look --and the numbers start to get pretty small -- overall, the rates of chronic infection are about four-tenths of a percent.
That translates into about one million Americans. So roughly 12 million Americans out there are infected with hepatitis B --have ever been infected, and about one million are chronically infected.
So what do we know about the incidence of hepatitis B as a whole? Well, basically, the incidence has been declining in recent years. Back in the mid to late '80s, we think the incidence peaked at roughly around 300,000 new cases per year. But since then, there's been a tremendous decline in the number of cases, and we think we're now down to in the ball park of 150,000 to 200,000 new cases.
So in the last 10 to 15 years, the incidence of disease has been half, and this is due to a number of different factors.
We noticed a tremendous decline among homosexual men and health care workers beginning in the mid to late '80s. Some of that is due to changes in risk factor behavior, as well as introduction of a very good, very safe, effective vaccine back here in the early 1980s, although it took a number of years to percolate into those groups at highest risk.
And since the vaccines being out there and people have been getting the message about, namely, HIV, we reap the benefits of HIV education because hepatitis B is spread in many of the same ways. So we also saw a decline, basically, predominantly among injecting drug users starting in the mid '90s.
Whether that's actually due to those prevention messages getting out there, we're not clear. But regardless, the incidence is dropping -- has dropped quite dramatically in the United States over the past decade.
So what are the risk factors for hepatitis B in the general U. S. population? And basically, they hit all the risk groups I'm going to talk about here later this morning.
Basically, roughly half of acute hepatitis B over the last decade is due to a sexual route. That is, either a heterosexual, which accounts for about 35 percent of everything, 39 percent, and homosexual transmission, which accounts for roughly 13 percent.
This data actually comes from our sentinel county surveillance study which has been done in four counties dating back to 1982. And essentially, what we do is we look at acute cases of viral hepatitis of all types and interview them and draw sera, and actually, for some subselected groups, follow them over a period of time.
So this is a very good way for us to track emerging infections. And actually, hepatitis C, which I'll talk about in a minute, was actually discovered in the serum that gave rise to --some of the antibody tests actually came from the sentinel county --was a case of non-A/ non-B hepatitis.
But regardless, this study basically interviews people who are acutely ill and then they admit to risk factors. This will become more important when we talk about hepatitis C. But basically, there's a group of people who admit to a whole, broad range of risk factors who actually don't admit to traditional risk factors such as a heterosexual partner or having a homosexual partner.
And basically, we think that all of these other people down here are essentially those that are a little truth challenged, as one of our nurses say. A lot of these people probably have all these other risk factors up here, but basically aren't admitting to them on interview.
So we think roughly in the ball park of maybe up to 50 to 60 percent of hepatitis B is sexually transmitted, and maybe up to 15 to 20 percent is through injection drug use.
Okay. So let's talk a little bit about the specific risk groups we're interested in this morning.
I thought I would present this data in the following fashion. It's important not to, when we talk about these risk factors for the population at large, talk about what is the prevalence of these characteristics in the population at large.
Basically, even though there's quite a bit of variability, when you look in the general population and look through the literature, you basically find that in the ball park of between one- half and five percent of the U. S. population has ever used injecting drugs. And this is quite a big range and really depends on who you ask and what studies you look at.
I think most people think it tends to be towards the lower end of this range than the upper end of the range. But in the published literature you see ranges of between one-half and five percent.
If you look among men who have had sex with men, this may represent up to ten percent of the general population. I could find no good data on how many people have ever been a commercial sex worker. I'm sure that data exists someplace; I just couldn't dredge it out of the literature.
There's no data on how many infected sex partners of hepatitis B are out there, or hepatitis C, but there is some good data that looks at lifetime sex partners. Again, this comes from the National Health and Nutrition Survey.
And basically, you find that roughly 20 percent of the U. S. population has had only zero or one lifetime sex partner. Fifty percent of people had between two and nine lifetime sex partners. Twenty percent have had between 10 and 49. And four percent of the U. S. population has more than 50 lifetime sex partners.
So even though we don't have a prevalence for commercial sex workers, some would think that, if you had more than 50 lifetime partners, you're probably a commercial sex worker or likely to be a commercial sex worker. So this number is probably much less than four percent, to hazard a guess.
So let's talk about the specific risk groups one by one. Let's talk about with injection drug use. Basically, hepatitis B is found in very high prevalence among injecting drug users. Roughly 60, 80 percent of people who have used injection drugs have hepatitis B.
What do we know about hepatitis B in these populations? Well, the seroprevalence varies quite a bit by age. It's strongly associated with age. The older you are, the more likely you are to become infected. And this is actually shown very clearly in the National Health and Nutrition Survey.
However, you do see some variation in prevalence by geographic region and risk factors within the injecting population. That is, different injectors use different drugs, some snort, some shoot, some shoot in different ways. So you have to think about when you look at prevalence of hepatitis B exactly what's going on in the population you're studying.
However, risk seems to increase quite dramatically with number of years of drug use. And if you look at people who have injected within at least five years, you find upwards of 90 percent of people who have injected at least five are infected with hepatitis B.
It's tough to come up with measures of incidence because injecting drug users are a very difficult group of people to follow, to get them to come back. But the ball park sort of estimate out there is probably around four percent per year of injectors become infected with hepatitis B.
However, these studies always need to be interpreted with a grain of caution because not only is hepatitis B spread through injection, but it's also spread through a sexual route. So you need to be very careful to separate out sex from the drugs when you look at these studies, and not all studies are very careful to do that. So you have to interpret the incidence figures with some caution.
Well, speaking of sex, what do we know about the prevalence of hepatitis B in various sexual characteristics. Well, as I mentioned earlier, hepatitis B seems to be spread fairly efficiently through sex. If you look among men who have had sex with men, you see seroprevalences of 20 to 40 percent. And basically, you find about the same seroprevalences among commercial sex workers, in the ball park of 10 to 40 percent. And these are also the same you see among STD clinic patients.
If you look among infected partners, you see seroprevalences of about 40 percent as well. You also see an increasing prevalence, based on number of lifetime sex partners, which peaks out about 12 percent among those who have had more than 50 lifetime sex partners.
So I hope you're convinced now that hepatitis B is transmitted fairly efficiently through sex. STDs play an important role in the transmission of hepatitis B, we believe. When you look at people with hepatitis B, at least 40 percent of these people have had an STD previously. And whether this is a marker for high-risk sexual behavior or may facilitate transmission was a little up in the air, but at least 40 percent of people have had a previous STD. Men who have had sex with men are at an extremely high risk of hepatitis B.
Risk factors include those of other sexual transmitted diseases, including multiple partners, receptive anal intercourse, and history of other STDs as well.
It is very difficult to get estimates of incidence for hepatitis B among men who have sex with men today. But if you look back in the pre- vaccine era --this is, again, sort of the pre-HIV era as well, back in the late '70s and early '80s, you see incidences up to 13 percent per year. I think we feel that the incidence is tremendously lower than that, basically due to use of hepatitis B vaccine in this population and exchanges in risk behavior.
However, it's important to remember that the seroprevalence among men who have sex with men, as well as these other risk groups, varies quite a bit by age, geographic region, risk factors, and, since there's a good vaccine, vaccine coverage within these populations.
Okay. So let's move on and talk about hepatitis C. This is, again, data from the National Health and Nutrition Survey, and this is the source of the oft-quoted number that roughly 1.8 percent of the general U. S. population is infected with hepatitis C or has antibodies for hepatitis C. And this translates into roughly four million Americans.
When you look at this data again, you find that it varies quite a bit by population subgroups. You find that roughly one and a half percent of non-Hispanic whites are infected with hepatitis C, roughly 3.2 percent of non-Hispanic blacks, and two percent of Mexican-Americans.
And actually, an interesting finding with this data is this is a cross-sectional study. That is, you take people at one time, over a short period of time in many different ages. If you take this data and actually plot it out by the age of the person interviewed versus how many are anti-HCV positive, you see a very interesting shaped curve.
You basically note that there is a big hump among the sort of middle-age groups here. And it reflects the increasing prevalence that we saw among the different population groups in the previous slide. That is, intensity lower among whites, somewhat higher among Mexican-Americans, and highest among blacks.
And if you bear with me for a second, I just drew some arbitrary lines here on this graph, and basically selected those between 30 and 50 years of age. And basically, if you look among those 30 to 50, and average the proportion that each of these groups accounts for in the general U. S. population, you basically find rates of about three and a half percent among those 30-to 50-year olds, and much lower rates among those older than 50.
This also gives rise to a number of interesting hypotheses that are often quoted in the literature --that the seroprevalence is much higher among to 30-to 50-year olds. We may be on the edge of an epidemic of chronic liver disease in this country. That is, as these cohorts start to age and move this way, we may be starting to see more and more chronic liver disease caused by hepatitis C. But that's the topic of another talk.
So let's talk a little bit about incidence. The prevalence is extremely high -- roughly two percent of the U. S. population. The incidence seems to have declined quite dramatically over the last decade or so. Basically, back in sort of the mid to late '80s, we think we saw in the ball park of about 150-to 200,000 new cases every year in the United States.
Basically, since then, due to a number of issues I'm not going to really talk about today, we saw a tremendous decline among transfusion recipients, starting in the mid '80s. And basically, that sort of started some of this decline. But we've also noticed a tremendous decline among injecting drug users in the last decade or so. And why this is happening is a little unclear, but it may have to do with saturation of the population at large, which I'll talk about here in a slide or two.
So what are risk factors for hepatitis C in the United States? Again, this is data from our sentinel county study, which basically interviews patients with acute hepatitis C and seeks to find the risk factor. The bottom line is: injection drug use today is the number one leading source of hepatitis C in the United States.
It's pretty remarkable to me that 40 percent of people will admit to using injecting drugs within the last six months upon interview. Roughly 16 percent of people admit to either having more than two sex partners in the last six months or have sex or are having sex with a person who we believe they know is anti-HCV positive.
If you'll look at this piece of the pie, roughly two-thirds of these people have an anti-HCV positive sex partner. Two of them have had more than two sex partners in the last six months and deny all of these other percutaneous exposures.
Again, since these people are interviewed and some of them tend to be a little truth challenged, when you look at people who really report none of these exposures here, basically you find they have a whole broad range of other risk factors. We think that probably another 14 percent of this pie, or accounting for about 60 percent of the total, are drug related. That is, these people are probably failing to admit to injection drug use that are actually injecting.
And we think that some of these people with a history of STD may be denying multiple sex partners. So we think that roughly about 60 percent of acute hepatitis C in the U. S. is due to injection or illegal drug use, predominantly injection, and roughly about 20 percent is due to sexual transmission.
This is a little controversial, as we'll talk about later on. However, we don't have any data on concurrent STDs in these people, which may explain why we see a higher rate of sexual transmission in this study than other people have seen. But I'll talk about that at the end.
An important point for this group is that only four percent of people report a transfusion or transfusion-associated. And interestingly, if you look at the data --we've seen no transfusion-associated cases since 19 --there have been no cases in 1995 and 1996.
And actually, we've only seen one case since 1992 when better screening became available. So this four percent is somewhat misleading because it's heavily weighted towards the 1991 end of this spectrum. So transfusion association cases seem to be declining quite dramatically in the U. S.
Okay. So let's talk about injection drug use. I've told you that injection drug use is the number one leading risk factor, and it also shows up in the prevalence data. Roughly 50 to 90 percent of people who use injection drugs are infected with hepatitis C.
Again, caveats apply. The seroprevalence tend to vary quite a bit by age, geographic region, and risk factors in the injecting population. And that explains that somewhat big spread between 50 and 90 percent. So it depends on who you look at, where you look at, and what the injectors are actually doing in that population.
However, we do know that the risk increases quite strongly based on the number of years injecting drug use. And we find that basically upwards of 90 percent of injectors areinfected within one to two years of the time they start injecting.
And depending on the studies you're reading, again, these have to be taken with a note of caution. You see incidences of up to 10 to 20 percent per year. That's right --10 to 20 percent per year.
However, there is some caveat that needs to be thrown in. A lot of these studies were actually done back in the late '80s and early '90s. There has been some studies today that seem to suggest that this incidence may actually be declining quite a bit. And why that is happening is a little unclear and may have to do with needle exchange programs, messages about HIV prevention that are getting out to the new injectors out there.
It's a topic that needs studied a little bit more. But regardless, the incidence rates tend to be tremendously high. And a lot of people have trouble buying into that the incidences are really actually that high.
And this actually is a very good study that was done by the folks in Baltimore, the ALIVE study, and what they basically did is looked at a group of injectors and asked them, "How long have you been injecting?" And then tested them for HIV, hepatitis B, and hepatitis C.
And here is what they found. Everybody thinks about HIV and injectors, and roughly 20 percent of the people were infected with HIV, but that came in number three in terms of blood-borne pathogens. HBV came in number two, with roughly 40 percent of people, by the time they started injecting, infected with hepatitis B. And this tended to increase very steady over the next two years. And actually, if you follow these people out for six years or so, it tends to plateau. So right around 60 to 70 percent.
But if you look among those with hepatitis C, basically 50 percent of people, by the time they got enrolled in a study, already had hepatitis C. And it very quickly went to 80 percent, within basically the first six months of the time they started injecting. And then it slowly worked its way up to 90 percent. And if you follow these people over the next five or six years, it sort of peaks out around 90 percent or so.
So basically, hepatitis C is acquired very, very rapidly through injection drug use, which makes it very difficult to do prevention strategy, since, again, roughly everybody is infected by the time they started injecting or very quickly have become injecting.
One other thing it's important to appreciate is is that the incidence of hepatitis C varies quite a bit, depending on when you're looking and who you're looking at. And these are the four primary sentinel counties we look at. And, again, this is our surveillance system that looks at acute cases of all types of viral hepatitis.
And just to give you a feel for where these are, Pinellas County is Tampa/ St. Pete, Jefferson County is Birmingham, this is the city/ county of Denver, and this is Tacoma, which is about 40 miles south of Seattle.
And basically, what do you see?
Basically, you can see from this graph --and again, these are on the same scale --that the incidence varies quite a bit depending on where you look. And it also indicates that you could have very large outbreaks of hepatitis C among injectors in the community. We saw a tremendous outbreak here sort of through the late '80s and early '90s among injectors in Pierce County.
So you have to sort of look at these data --look at incidence data with a little bit of --a grain of salt, I guess.
So let's talk about hepatitis C and sex.
Basically, when you look at men who have sex with men, it doesn't seem to be nearly as high as you would expect. Basically, the seroprevalences of around four percent are out there. I didn't present a range for this because the range is a little misleading. Depending on what study you look at, you see ranges from one percent to 15 percent.
The 15 percent is only in one study and seems to be a little bit of an outlier, but I'll talk about more of that in a second. But overall, I think the feeling is you see seroprevalence of around four percent among men who have sex with men.
You see seroprevalences, again, between one and 20 percent among commercial sex workers, although it tends to be more towards the lower range than the upper range in the majority of studies. Among infected sex partners, you see seroprevalences of roughly one and a half percent, although there needs to be a lot more work to look at this group. But that's probably a reasonable estimate.
You also see that the seroprevalence increases by number of lifetime sex partners, with those who have more than 50 lifetime sex partners have seroprevalences approaching 10 percent.
Again, this should be taken with a grain of salt as well, because this comes from the National Health and Nutrition survey, which didn't ask about injection drug use. So basically, we don't have any idea how many of these people actually acquired it through sex and how many got it through injection use.
And it's probably a reasonable assumption that people who have more than 50 lifetime sex partners --at least some proportion of these people are participating in injection drug use activities. So the seroprevalence is probably much lower than is actually presented in these slides once you take out history of injection drug use.
So let's talk about sex. This is one of the more controversial areas of hepatitis C research now, and it's an area that needs a lot of work. Basically, the overall opinion is the efficiency of transmission of HCV through sex is relatively low. What does that actually mean? Well, it basically means transmission can occur, transmission is probably rare between long-term steady sex partners at least, although the actual risk of transmission is unknown.
We're in the process of trying to set up a study to look at this. The general feeling is it's probably going to be less than one percent per year, which, again, makes studies very difficult to do because the incidence is relatively low. But nobody is really ready to hazard a guess among infected sex partners, among long-term steady sex partners at this point, other than to say that the incidence seems to be relatively low.
However, on the other hand, when you look at hepatitis C as a traditional sexually transmitted disease, basically you find it more frequently among people with high-risk sexual behaviors. And the risk factors, when studies have looked at it, seem to be --for hepatitis C infections, seem to be pretty much the same you see for other STDs. That is, multiple partners, histories of STD, and failure to use a condom seem to be associated with HCV infection. So it sort of looks like it could be a sexually transmitted disease.
However, when you look among men who have sex with men, they have about the same risk as basically --as heterosexuals do for this. So it seems to be a little confounding. And why this is true is unclear, and it's, again, an area for future research. But it seems to sort of fly in the face of reason that it seems to appear to act like an STD, but it doesn't appear to act like an STD.
However, one of the major limitations of a lot of these studies is they haven't really looked at other risk factors associated with transmission. There may be other factors that may promote transmission of HCV in a sexual arena, such as viral titer and other concurrent STDs. Again, it's unknown whether other alterative STDs may facilitate HCV transmission, and this is an area of important research.
Another important thing is that a lot of these studies, especially done among commercial sex workers and STD clinics, failed to do a good job of separating sex from injection. We know that hepatitis C is very, very efficiently spread through injection drug use. And if you don't do a good job of teasing out those that are injectors from those that have a pure sexual route, you can very easily contaminate your data and get to wrong results.
And finally, although there is sort of developing data on this, again, the seroprevalence for HCV seems to vary a little bit or seems to vary by age, geographic region, as well as risk factors in the population --namely, injection drug use and sexual activity.
So if you have to summarize everything onto one slide, which I tried to do here, basically, you find that hepatitis B is a relatively --occurs in about five percent in the U. S. population, is spread relatively efficiently through sex, and spread very efficiently through injection drug use. Hepatitis C occurs in about two percent of the population, is probably spread less efficiently through sex, but very, very efficiently through injection drug use.
Thank you very much.
(Applause.)
DR. DAYTON: Thank you very much for that excellent talk. We'll have an opportunity to discuss this and the other talks in a panel discussion coming up.
The next talk will be from Rick Steketee on prevalence and incidence of HIV in high-risk groups. Return to Agenda
DR. STEKETEE: Thanks very much, and I, too, would like to thank the organizers for inviting me.
I was asked to speak on HIV prevalence and incidence in certain groups who engage in high- risk behaviors.
Specifically, I'll show some data from a variety of CDC-supported studies among men who have sex with men, or MSM; among injection drug users, or IDUs; and among women who report exchanging sex for money or drugs. I've limited it to women not because men don't exchange sex for money or drugs, but because our studies have a tendency to be more clear on that particular risk group.
The data I'll show come from a variety of sources. These include anonymous unlinked seroprevalence surveys that sampled consecutive persons attending selected STD clinics or drug treatment centers. In addition, in some STD clinics, persons who accepted counseling and testing for HIV on two or more visits were examined for incidence in the interval.
Data was also drawn from the national counseling and testing system database, and from young men's surveys, which are venue-based surveys from street outreach clubs or bars in young gay men.
All risk behavior categorization is based on self-reported or participant here. And for simplicity of categorization for the presentation, we limited the analysis, as I mentioned, just to women who are exchanging sex for money or drugs, and we'll be referring to them as commercial sex workers.
Finally, as usual, the data comes from the work of many people at state and local health departments, and some community-based projects, and investigators at CDC. And I'm pleased to present the information for them.
Let me begin with data from counseling and testing system in 1996, which is our last year of complete data collection and analysis. This slide shows the seroprevalence in various groups. Remember that they're from an amalgamation of persons who accept or seek HIV counseling and testing at publicly-funded sites, including anonymous test sites, STD sites, drug treatment centers, family planning clinics, adolescent clinics, etcetera.
There were approximately 2.5 million tests done in these settings in 1996, and the HIV prevalence was highest in MSM reporting injection drug use --around 9.5 percent --and next highest in MSM not reporting injection drug use, around 6.6 percent. And it was 4.5 percent in heterosexual injection drug users and lowest, 1.2 percent, in heterosexuals not reporting either MSM or IDU.
This map shows data from anonymous unlinked serosurveys and HIV prevalence in MSM attending STD clinics in 14 cities in 1997. Note that the bar scale is from zero to 40 percent, which is generally --and the HIV prevalence is generally high in this population of MSM and STD clinics, ranges from 3.6 percent in Seattle to about 36 percent in Atlanta. The overall median prevalence for MSM in STD clinics was 20 percent.
This map shows comparable data on women attending STD clinics, and note that the bar scale has changed from zero to seven percent, instead of zero to 40 percent. Again, prevalence is fairly consistent across the country, but ranges from approximately one percent in Denver to about five percent in Miami.
And this map shows HIV prevalence in injection drug users attending drug treatment centers in 12 cities in 1997. The scale is back again from zero to 40 percent. And as has been seen in the past, there is high prevalence generally in the east and substantially lower prevalence in the west, where prevalence overall, the median prevalence in injection drug users was 15 percent for men, and for women it was 11.6 percent.
This slide shows HIV prevalence in women who reported exchanging sex for money or drugs, or commercial sex workers, in three different settings. HIV prevalence was 9.7 percent in commercial sex workers attending drug treatment centers, 6.1 percent in those attending STD clinics, and 3.5 percent of those reporting commercial sex and attending various counseling and testing sites.
Next what I'd like to do is show some data on HIV prevalence among those reporting the risk behavior during the past year, compared to those reporting the risk behavior more than a year ago and not during the past year.
Among attendees at an STD clinic in 1997, this shows reported risk in yellow --I'm sorry, reported recent risk in yellow and past risk in blue, among men who have sex with men, among heterosexual IDUs, and among commercial sex workers.
Although HIV prevalence varied a little between the groups, and across recent versus past risk behavior, all groups have reasonably high HIV prevalence.
This slide shows similar data from drug treatment centers where HIV prevalence was high and did not differ by recent or past reported risk behavior in injection drug users or in commercial sex workers.
Finally, I'd like to show a few slides on estimates of HIV incidence in these risk populations. While prevalence is indicative of cumulative acquisition of infection, incidence tells us about recent or current transmission patterns. This slide shows incidence per hundred person years in MSM, in women and heterosexual men in STD clinics repeatedly tested during 1991 to 1996, in seven different U. S. cities.
The measured incidence varied from seven per hundred person years in MSM in Houston to very low rates in heterosexuals in Denver. That is, around one to two per thousand person years, as opposed to seven per hundred person years.
And this slide shows HIV incidence in STD clinics with heterosexuals in yellow and men who have sex with men in red. Of interest, incidence in MSM gradually declines with increasing age, and amongst heterosexuals it gradually increases slightly with increasing age.
However, in those less than 40 years old, the incidence of HIV in MSM is between three and 10 times higher than it is in heterosexuals.
Finally, with the assistance of the San Francisco Health Department, we were able to obtain incidence estimates in one population --that is, men who have sex with men --in one city in various venues in a recent year. As you can see, first of all, that the incidence over here, total in the STD clinic, is about one per hundred person years. In MSM, in that environment, it's about four-fold higher.
And in two other types of venues --that is, anonymous testing sites and out in venue-based surveys --the incidence of HIV roughly varies between two and four per hundred person years. And it is not greatly dissimilar across the different sites.
So in summary, in 1997, HIV prevalence and incidence is still high in traditional risk groups. Among men who have sex with men, this is true in a fairly wide geographic distribution, in a wide age range, across various venues of surveys, and regardless of recent versus past reported exposures. And at least for HIV prevalence that's true in this recent versus past exposure.
Similarly, HIV prevalence in injection drug users remains high, although there is greater geographic variation. And in women exchanging sex for money or drugs, they continue to have high prevalence of HIV, also across different venues and geography.
Thank you very much.
(Applause.)
DR. DAYTON: We'll move along to the next presentation now. Bernie Poiesz will give a talk on prevalence and incidence of HTLV in high- risk behavior groups. Return to Agenda
DR. POIESZ: Thank you. Dr. Jaffe has already done a nice job in introducing the topic. I'm asked to concentrate on discussions about HTLV I and HTLV II, which, as you can see, are members of an oncogenic genus of retrovirus that also contains bovine leukemia virus.
We have developed a convention of referring to this group in its toto as the primate T-cell lymphoma leukemia viruses, because, as was mentioned, the genetic overlap between simian strains of this genus is quite frequent. And you really can't separate the strains by species; you have to separate them by geography and temporal dissemination from each other.
As was mentioned, HTLV I causes a variety of diseases, most notably adult T-cell lymphoma leukemia, but also myelopathy, polymyositis, Sroegen's syndrome, and perhaps a variety of other autoimmune diseases. It can cause a low degree of immunodeficiency, and half the patients that present with HTLV present with opportunistic infections and quite often die from that, but certainly nowhere near its distant cousin, HIV.
HTLV II probably does cause some finite amount of disease in humans, but it has to be extremely rare. We've been involved now in working up in toto, in the entire history of our laboratory, eight cases of CDA positive T-cell lymphoma which we believe are caused by HTLV II, as opposed to thousands of cases of adult T-cell lymphoma leukemia.
We've also been involved with identifying approximately 20 patients who have a neurologic disorder that is quite similar to HTLV I, except that the area of greatest involvement in HTLV I seems to be the thoracic cord, whereas in HTLV II it seems to be the cerebellum and the cerebellar tracts, such that the patients present with a cerebellar ataxia.
We're involved in large studies in endemic groups in paleoAmerindians to try and really identify the true incidence and prevalence of disease.
HTLV I causes disease in about four percent of infected people over their entire lifetime. However, if one is infected perinatally, the lifetime risk for developing adult T-cell leukemia goes up to about 10 percent; hence, one of the major pushes to stop perinatal transmission.
To my knowledge, no one has developed adult T-cell lymphoma leukemia from an HTLV I infection that occurred via blood transfusion, although certainly people have developed HTLV I associated myelopathy; and, in fact, have developed it in a very quick timeframe. The earliest that I know is three months post-transfusion. So that seems to be the major risk. But, of course, if you transmit it via transfusion, then the chance of transmitting it to other people and getting that perinatal infection goes up.
I want to talk a little bit about the biology of this genus of retroviruses because it is clearly different from HIV. It replicates very slowly. It's hard to transmit it, and its efficiency of transmission is about one one- thousandth, that of HIV. And it expresses its RNA and proteins to a very low degree.
As you'll see, the point I'll make is that if you want absolute sensitive detection of this group of viruses, serology assays probably won't do it because in some people there either are defective viruses or a very slow latent period in terms of expression such that seroconversion can take a long period of time.
This is another phylogram showing you the BLV genus group here, and the HTLV I or PTLV I group here. This is the human group of HTLV II. Mixed in here are several simian strains of STLV I's, and they just overlap. I'll show that a little clearer in another slide.
These are two new members of the genus that have been identified in the past couple of years. Primate T-cell lymphoma virus long has been found in Entrean baboons whose previous geographic range was southwest Asia and northeast Africa. This is STLV II, which is found in pygmy chimps in Africa.
All of the HTLV II's identified to date fall in a very close group, no matter what human, in what part of the world, even Central Africa; they seem very close to those strains found in paleoAmerindians. The HLTV I's and simian strains are divided into two groups: those in Africa and those in Asia, Australia, and Melanesia. To date, no one has found a human counterpart to PTLV I or to STLV II, but I would submit that perhaps people haven't looked enough and they may exist.
Among the genus, divergence of one percent takes about 500 to 1,000 years of separation, so there is relative conservation making development of degenerate or generic assays somewhat easier than it is for HIV. There is very little evidence for recombination. Although I don't have time to show you, we now have evidence that modern BLV represents a recombination of something between STLV II and PTLV I, with an H and PLV.
It's important to note that because we now know that we have many intravenous drug abusers who are co-infected with both HTLV I and HTLV II, and, to my knowledge, no one has looked to see if recombination has occurred and what would be the biology of such a recombinant strain. We know that out in areas where there's different strains of HIV recombination occurs in about 10 percent of the isolates looked at.
Most of the serology strains used to look for antibodies to HTLV I or II are developed from a West African HTLV I isolate. Recently, people have developed recombinant peptides from the West African strain to add to the assays or from an HTLV IIA strain. And very recently, Abbott Laboratories has used an HTLV II strain to add to the HTLV I antigen to broaden the mix.
But you can see that there is relative divergence among these, and absolute cross- reactivity might not occur. There is approximately 40 percent divergence between HTLV II and HTLV I, and about 60 percent to BLV.
Now, I want to talk about the biology of the virus and the differences between HTLV and HIV. Again, HTLV has gotten into humans, into primates, tens and tens of thousands of years ago. And over time, evolution has probably resulted in a more symbiotic relationship than we see with HIV 1.
One of the differences between HIV 1 and HTLV is the presence of complete retroviral DNA transcripts in the virus. We now know that HIV is capable of full reverse transcription in an extracellular mode, not to the degree of the hepatitis B virus --remember, hepatitis B virus is a retrovirus in disguise.
It replicates to an RNA intermediate, but, intracellularly, almost completely replicates its DNA into a double-stranded DNA, finishes that extracellularly. That's, in part, why your multiplicity of infection and your transmission rate for hepatitis B virus was higher than HIV. And HIV does this to some degree; HTLV does not do it well at all.
Proviral DNA may. You get one copy of DNA for every 10 3 copies of HIV RNA; whereas, for HTLV you get one copy of DNA for every 10 6 molecules of viral RNA. So there is roughly about a thousand- fold difference in transmission.
We now know in our laboratory we've at least studied this. This is the reverse transcription step, and it can kind of be broken up into three parts. Viral RNA starts as a single- stranded RNA, and there is a tRNA primer here that primes what's called strong stop DNA synthesis. That RNA then gets degraded by the viral RNA's H, and this strong stop DNA has to make a jump to this end of the viral RNA where its complimentary to the repeated sequences. And then first strand synthesis occurs, then there's more degradation, and eventually full length.
We can make primer pairs and do PCR to look for these various components. We now know that HTLV makes strong stop DNA about one-tenth the efficiency of HIV. It makes the first jump at about one one-hundredth the efficiency and full length at about one one-thousandth. Somewhere in here is the major block in HTLV replication relative to HIV.
Obviously, therapeutically, if we could identify these molecular reasons, we might be able to design attack points to make HIV behave more like HTLV and slow down its transmission. But this, in part, explains why HTLV replicates so slowly.
After the viral DNA gets integrated, in HIV there is relatively rapid transcription of the RNA, and modulation of splicing patterns that is different than what we find in HTLV.
In all of the complex retroviruses, there is regulation of splicing. Initially, when a viral RNA transcript is made, the complete primary transcript is synthesized. In both HIV and in HTLV early infection, this RNA is quickly spliced down to multiply-spliced or singly-spliced molecules.
The multiply-spliced RNAs encode for these proteins. In HTLV I, it's TAX and REX, and a variety of others, the single splice for the envelope, and the primary transcript for the GAG/ POL proteins. In your antibody tests, the major proteins are the GAG and the ENV proteins. In HIV, there is rapid progression from dominant multiply- and singly-spliced messages to making unspliced message and making infectious virions and making all of the proteins.
In vivo, it has been noted that asymptomatic patients will tend to have these dominant species, and then as they go to symptomatic make more of the primary transcript.
In HTLV, the opposite is true. Both in vitro and in vivo, the dominant species, one hundred to a thousand-fold over the primary transcript are the singly-and multiply-spliced RNAs. HTLV- infected cells simply do not make a lot of retroviral virions, and they don't make a lot of GAG protein; hence, they don't stimulate antibody production to the major protein that we have in the assay.
When you use the purified virions to make an antigen prep for the serology assay, there is also a great difference because of these problems in replication, or differences in replication, between HTLV and HIV. The antigen preps are made by purifying virions from cell culture condition media. Again, in that media, there is roughly about one one-thousandth the content of HTLV virions per cellular debris than there is for HIV. So the viral protein to cellular debris ratio is off quite a bit, and your preparation is not as pure.
The other thing is that in HTLV purified virions, the envelope proteins gp46 and gp21E are deficient. Now, we know that the cells make them to a varying degree, because if we do RIPAs we see them there. But somehow they don't get incorporated into the virion to the same degree that HIV does.
One of the reasons is that HIV has a regulatory gene called VPU. VPU's function in the golgi apparatus is to degrade the cd4 protein and message such that receptor for HIV glycoprotein is not present, allowing the envelope protein to make it to the surface. HTLV has no such gene, and it doesn't down regulate its receptor like HIV does.
So when you make an HTLV virion, it is relatively deficient to its GAG proteins in this envelope protein. And if you look at a Western Blot on some of the classical assays that are FDA approved, unless you put a recombinant envelope protein in there, you won't see any reactivity to an envelope. So it's a major difference.
In HTLV I, some of the non-specific reactivity in normals is against the p19 and the p21. We now have data that part of the reason for this is we all contain endogenous retroviral sequences in varying amounts and in varying different sequences and varying degrees of expression during our lifetime that have homology to these two proteins.
The epitopes have been identified in p19, and the epitopes for cross-reactivity have been identified in p21E. And if you make peptides that do not encompass those overlapping epitopes, you get a much better preparation.
Gene Labs made a Western Blot with an epitope called GD 21, and that's actually a very good, very specific epitope. They have another one called BA 21 that cross-reacts in about seven percent of normal humans and higher in certain diseases. But probably to make a better HTLV I antigen relative to HIV, you probably are going to have to depend more upon recombinant proteins and peptides to fill in these gaps of deficient proteins and to try to avoid some of the overlapping sequences that may be expressed by endogenous sequences.
To look for HTLV I in a sensitive manner, in my opinion, you have to do PCR for DNA. It doesn't help to do PCR for RNA, because HTLV I, HTLV II, BLV-infected animals do not express a lot of RNA. Our range of detection for RNA is such that only about 60 percent of infected individuals have detectable RNA in their plasma, and the copy numbers range from anywhere from 10 to 1,000 per ml, where the copy numbers for HIV will be in the hundreds up to ten million. So it's very rare to find high copy number viral RNA expression.
We've done --one of the problems with PCR is making it sensitive, making it multiplex so that you can look for variety assays, and making it specific. In terms of sensitivity, we have collaborated with the folks at Johnson & Johnson, and they have developed two monoclonal antibodies against the DNA polymerase, Taq polymerase. And when the antibody is added it activates the Taq polymerase.
This prevents false primer extension should your viral primers anneal to something in the human genome that has some homology, all right, and dampen your productivity. If we add these antibodies, we get approximately a thousand-fold greater yield in our PCR product after about 40 cycles.
So it enables us to do sequencing a lot easier, but it has made all of the assays robust, such that we can usually develop a PCR assay for a known human retrovirus that is sensitive down to one copy per aliquot in a Poisson distribution, i. e. 60 percent of the samples at that concentration will be positive, the maximum sensitivity.
Another problem is carryover. If you amplify the DNA in an open lab, open everything up and then try to go detect it again in another person, you'll start getting false positives from the synthetic DNA that you've made and aerosolized in your own laboratory.
In our laboratory, and with the data I'm about to show you, we have physically separated the pre-and post-PCR people, equipment, personnel. They're actually in a separate building. That helps. We also use uracil N glycosylase. We incorporate DUNP into the synthetic DNA and can presterilize that DNA by treating it with uracil N glycosylase, which hydrolyzes the synthetic DNA.
The other thing we do in our primers -- we add linker sequences on their 5 prime end, such that all of the synthetic amplicons have this non- human/ non-viral DNA at their tail. And we go back and make primers just to the yellow portion, the non-viral portion, and scan our samples to see if we have any false positive. A negative result would suggest that our positive result before on the human sample with the viral-containing primers is a true positive.
We have recently collaborated with Fred Kramer at the Rockefeller Center to develop a system and test it in human retroviruses. I think it solves a lot of these problems. They have worked with beacon probes. They can do PCR now in a single tube that doesn't have to be opened from start to finish. You can throw it away at the end and can multiplex several different assays, both for sensitivity detection and for quantification over several cycles. The capacity at the moment is up to 28 simultaneous targets, either 28 different strains of HIV, 28 different resistance molecules, or 28 different life forms at any one time.
The beauty of this is that their probe can be silenced completely. The business end of their detector sequence is shown here in the circle, and it has a tail on either end. The open circle is a fore, and the dark circle is a quench. When it doesn't see its target, kinetics are such that it wants to stay in this stem loop structure, and that brings the quencher close to the fore and completely inactivates it.
When it sees its target and hybridizes, however, the fore is now removed from the quench, and you have light. You go from dark to light. And the background noise is extraordinarily low, such that you can add all of this in the beginning and it starts to hybridize as you do each PCR cycle.
This just shows you results with HIV 1, HIV 2, HTLV I, and HTLV II. We have very sensitive detection of these. We can mix and match them. We could look for different strains and get very robust amplification. We made primer pairs to all of the known strains of HIV and all of the known strains of HTLV. And at least what was in the literature we could find all of the known variance that exists in the world, to our knowledge.
We're actually in a position now of making mutations. We're making random mutations and selecting out viable mutants to try and see if there is anything that can escape our primer pair system now. We're going to make them, rather than go out into Africa and find every strain. We're going to make them in the lab, and we've proven you can do that.
It's also linear and quantitative over a very long range. It's hard for you to see, but this is the multi log of linear range because of where you're starting at. So quantification occurs --the cycle that the background --the signal comes off the background noise determines the copy number, and the range is almost over a million-fold. So it makes it very suitable for quantification.
As for some actual results --again, I'm not an epidemiologist, so I have some prevalence slides. I'll talk about some incidents as I know them, and I'm probably not as sophisticated as some of the other speakers.
HTLV II is endemic in paleoAmerindians. We have been collaborating with Dr. George Ferrer and Eduardo Esteban, studying the Indians of the Gran Chaco plateau in South America. This is the plateau that skirts northern Argentina, Paraguay, and Bolivia, and it is made up of two major linguistic and genetic groups of Indians. They have a very high incidence of HTLV II and prevalence of it.
Now, as I show you this data, this is not the entire tribe, and it's fair to point out that this is us going and looking at family members and sex partners and children of some of the initial infected people. So the prevalence rates will be quite high.
We used a variety of screening ELISAs; again, made primarily with an HTLV I Western Africa antigen prep that we used to select ELISA which can discriminate between HTLV I and II. And at this point in time, we used the Gene Labs' 2.3 Western Blot that does not contain that GD 21. And we did PCR from a primer pair that's conserved in HTLV I and II POL gene.
This is the sensitivity and specificity on Indians that we found. You can see that the screening ELISAs, etcetera, have a relatively low sensitivity relative to PCR, but that PCR was not a hundred percent. The specificity of the screening ELISAs vary. Actually, the lowest one was removed from the market, in part because of that. The select ELISA --and the Western Blot is how we interpreted reactivity to p24 and gp46 being a positive --was quite specific and the PCR was quite specific.
Now I'll show you similar sensitivity results, if you can see them, in a variety of groups at risk for HTLV I or HTLV II. These are American IV drug users, irregardless of race, and they had a 14 percent positive rate. And the serology assay was about 89 percent, and the PCR was 98.6 percent. The people who were seronegative tended to be those who have picked up their IV drug abuse relatively recently. And when we came back to them and followed them two years later, about 10 percent of the seronegatives had seroconverted.
In sub-Saharan Africa, the prevalence rate was 11 percent. Again, the serology is 93.8 percent, and the PCR was 100 percent. In paleoAmerindians, we've studied three major groups --the seminole, the Yaruro Quahibo in Venezuela, and the Toba and Matako Mataquaqan in the Gran Chaco. And again, you can see the serology results go anywhere from 71 percent to 83 percent sensitivity, and the PCR is 97 percent to 100 percent.
The point being, that if you really want to find all people infected with HTLV II, or all people infected with HTLV I, you pretty much have to do both assays in order to pick them all up. A number of labs have done this now, and I think it's a believed truth.
We have also done this in animal models. Part of the variation --we find that we now know the receptors for HTLV I. We have identified that humans and animals have different alleles for this receptor. And what we don't know is whether those alleles correlate for different rates of infection, etcetera.
This is the prevalence rate in various groups that we've tested for HTLV I or II. This was done approximately about five years ago, so it doesn't reflect recent data. And here we're calling them positive if they are seropositive and PCR positive. Remember, if we probably had done PCR in all of these people, the prevalence rate would be slightly higher.
This is a volunteer blood donor group. It's predominantly blood donors in the northeast, and the prevalence rate was about .02 percent. One person was HTLV I; the other person was HTLV II. You can see in paid blood donors that the prevalence rate goes up higher and it's statistically different.
We studied caucasian IV drug abusers to eliminate the background noise of HTLV I being endemic in black people. And this is predominantly IV drug abusers in the Syracuse and New York City area. And you can see the prevalence rate there was about four and a half percent.
In studying caucasian prostitutes in New York City and Syracuse, we didn't find any of them infected.
This is --we've got homosexuals, hemophiliacs. Again, this is predominantly caucasian homosexuals and hemophiliacs in the central New York and New York metropolitan area. And only one person was positive --a homosexual.
The hemophiliac data points out what Dr. Jaffe alluded to. With Alan Williams, we have done studies in the past and looked at people who have gotten seroproducts, either Factor VIII, that were hemophiliacs, or immunoglobulin preps. We find no evidence of plasma products ever passing HTLV I or II. The transmission rate of HTLV I or II via cellular products occurs, and it depends upon the amount of blood that a person received and the timing of the blood. Those blood products that were stored for more than five days tended to have less of a transmission rate.
In family members and sex partners of HTLV positive, you can see the rate was about 13.6. The data, if you follow these people, are that in babies born to mothers who breast-feed for at least two years, the transmission rate is about 30 percent. If you cut off breast-feeding at about six months, the transmission rate drops considerably. And, in part, this has been suggested to be due to a decrease in neutralizing antibodies in the breast milk to the virus.
In Japan, where they have identified -- in southern Japan, where they have identified most pregnant women as being HTLV I positive or negative, and mandated that those women not breast-feed, the maternal transmission rate has dropped down to less than one percent. So that seems to be a significant thing in another part of the world where you can affect the use of breast-feeding in positive women.
In sex partners, the transmission rate male to female is greater than female to male. And in life partners over their time, from someone who we believe was infected perinatally, the transmission rate is about 30 percent to their sex partner, if it's male to female, and about 10 percent if it's female to male. But per year, the transmission rate is very low.
Needle stick victims --we had a contract with the NIH and a variety of other groups to look at all of their HTLV-related accidents, where people had jammed themselves with a needle or pricked themselves, etcetera. These are the first thousand people. We have not found anyone that has been infected via that route, so it must be relatively rare, if it occurs at all.
These are black people coming to medical clinics in Brooklyn. It doesn't necessarily reflect the general black population of Brooklyn, but people coming to a medical clinic, and the prevalence rate there was around four percent. When we looked at the same type of group in central New York, the prevalence rate dropped to 1.2. And if we looked at caucasians, the rate was much lower.
And then this just represents the study within the cancer acute leukemia group B, to look and see how many HTLV I related lymphomas or leukemias are occurring per unit of time. So over a six-month period, we collected a variety of patients with either CML or AML, ALL, CLL, and found none of them to be infected, even though they had gotten blood transfusions both in central New York and mostly metropolitan New York City.
Others have probably looked at an earlier population where screening for blood may not have been drawn on, and in that population that had received a lot of blood transfusions have identified infected people. This occurred after we had testing for HTLV I, and that seemed to have solved that problem.
In our lymphoma group, we found eight positive people, and they were all in the other than low-grade, non-Hodgkins lymphoma for a prevalence rate of four percent in that group, which we would suggest is probably the prevalence rate of that disease in other than low-grade lymphomas in the United States.
So I'll stop there. It's clear that there are risk groups for HTLV I. If you want to monitor them, serology and PCR seem to be required. Thank you.
(Applause.)
DR. DAYTON: Thank you very much, Dr. Poiesz.
We're going to take about a 10-minute break now, and then we'll try to fit in a panel discussion afterwards, if all of the speakers who spoke this morning could join us up at the front table here.
(Whereupon, the proceedings in the foregoing matter went off the record at 10: 39 a. m. and went back on the record at 10: 52 a. m.)
DR. DAYTON: If we could begin to get organized, settled, we'd like to begin the panel discussion. And I'd like to invite all of the previous speakers to take a seat at the table.
We thought we'd get things started with the panel discussion by just reviewing some of the general questions that we have, basically general questions which are the theme of this workshop. And I can read them, if --and I'll just read them fairly quickly.
In the face of sensitive tests for HIV, HBV, HCV, and HTLV, should men who have had sex with another man even one time since 1977 --that's A, or B, people who have had sex for money or drugs since 1977 --you can see that the language of this largely comes from the HIV epidemic --C, people who have ever abused intravenous drugs, and, D, sexual partners of the above, should these groups be deferred for life?
Another general question is: what lessons have we learned from prevalence and incidence of the diseases we have discussed today in individuals who engage in these activities with respect to blood safety. Obviously, this is very closely related to the question we just went through.
What lessons have we learned from emerging infectious diseases in individuals who engage in these activities, with regard to blood safety?
So if I can encourage any of the speakers to either volunteer to get things started, or perhaps we could start with a general discussion. As I was discussing with Harold Jaffe during the break, what do we do with unknown diseases? And, of course, that's an almost unanswerable question, but it --on the one hand, we --as Dr. Jaffe pointed out, each new pathogen or each pathogen can behave very differently in its rate of transmission through various modes, even if it share modes of transmissions with other pathogens.
And how do we handle this in terms of, do we consider certain high-risk behaviors that are high-risk behaviors for several pathogens? Do we justifiably consider them as high risk for unknown pathogens? Is there anybody who --Dr. Jaffe, would you care to comment on that? I'll put you in the hot seat.
DR. JAFFE: I think I've been set up.
DR. DAYTON: Absolutely.
DR. JAFFE: I don't know the answer. It seems to me, you know, reasonable to think, though, that injecting drug users would be at risk for any blood-borne pathogen, almost by definition, if you're injecting a contaminated syringe into your own body that you would be exposed.
So I think that's probably a safer assumption than to say that any agent which has been shown to be --or any member of a group of agents which has been shown to be sexually transmitted, that other members would be sexually transmitted as well. So I think it's a safer bet to think that injection drug users probably are going to be at risk for future emerging blood-borne infections, and that it would be harder to generalize about sexually-transmitted infections.
DR. POIESZ: I would say, number one, it's pretty evident that we keep getting pathogens introduced from some other source, other than humans, episodically over our lifetime as a species.
The other thing is that all of these phylograms that we're putting up there, the one thing we didn't have time to get into, if you actually work out the mathematics and the degree of divergence, the degree of mutation that they have per unit of time, there is things that are missing on those phylograms.
You saw my thing with the BLV. We looked at cattle across the world, dairy and beef cattle, and the total divergence is only six to eight percent. But the other side of the node, we have HTLV I and HTLV II that are 40 percent divergent. And yet every mathematical calculation we make says that the BLV side should have mutated to the same degree as the PTLV side, and yet we don't find it.
Now, nobody has gone to yaks, water buffalo, etcetera, and looked for these other strains or looked at other primates for them. But there have to be either extinct strains on that side of the phylogram or they're still out there. And what they would do to man we don't know, but there have to be a lot of other strains that can fit on those phylograms.
The same for HIV. And if you do it for hepatitis B, hepatitis C, you come to the same mathematical conclusion. So I'd say that one thing you could predict is there are other variants out there of these known groups.
DR. IAN WILLIAMS: One sort of caveat I guess I'd add from the hepatitis B and C perspective is is the hepatitis B and C have been around probably for long periods of time. Hepatitis B has probably been around --is a relatively ancient disease. And the data on hepatitis C is a little less sure, but there has been at least one study that found hepatitis C in a group of Air Force recruits as early as the late 1940s.
So when you think about putting date limits on questions, you have to consider that some of these diseases have been around much earlier than HIV. So it's somewhat artificial or something to at least consider when you think about hepatitis B and C.
DR. DAYTON: We had some questions from the floor, I think. Did you --
DR. IAN WILLIAMS: We have at least one question, and the question is: can HCV be transmitted through close contact within households? The answer is yes, probably, but it occurs very, very rarely. Basically, our current recommendations say that household members shouldn't share anything that could potentially become blood contaminated, such as toothbrushes and razors or anything that could become blood contaminated. And if you have open cuts and sores, you should keep them loosely covered.
This is more of a response to the fact that --a theoretical risk rather than we actually see transmission occurring by these means. And the bottom line really is is that yeah, transmission could occur, but we really don't see it. So, you know, hugging, sneezing, kissing, all those sort of things that cause general public concern do not transmit HCV. And probably a little common sense about exposure to blood is warranted in the household setting.
DR. DAYTON: Thank you.
I'd be very willing to open this up to questions from the floor. If anybody has any questions, just go to the microphone.
Jay?
DR. EPSTEIN: I believe it was Dr. Steketee who showed us a graph of prevalence of various markers split out by whether there was history in the last year, or I guess it was lifetime history. And although the slide went by quickly, it looked as if there were no significant differences. And I wonder whether that observation has any implication in your own mind about the question of lifetime versus temporary deferrals.
DR. STEKETEE: Yeah. I think basically the answer is that picking a specific year for when risk began was used for HIV largely because we had a time when we thought HIV was introduced in the population. And as you just pointed out, HCV and HBV have been around for a lot longer than that. So our data right now suggest that there is no clear year to pick when somebody had recent behavior versus long-since-past behavior that would help us.
And let me --I'll make an additional comment. Going back to Andy's --one of his opening slides about using tests to intervene between the selected blood donors and eliminating any possible infections in that pool, and then using questionnaires to get us a suitable donor population. And what I would suggest is that given the prevalence of HIV and incidence of HIV are high, what we have done in the past is that we initially set out donor suitability criteria as the first gate, and then used tests and have spent an enormous amount of time trying to use tests to get --to find those incident infections because we had a fairly good test and we have used donor suitability criteria in order to reduce the prevalence in the population of acceptable donors to such a level that we could account for all of the prevalent cases with our test and then spent a lot of time on the incident cases.
If we change the prevalence in the population by relaxing criteria and hoping for the test to pick up all of the prevalent cases, then we double the indemnity on the test. And just --I think that's what our data would suggest. There's not a year to go back to, and the prevalence is still high in those traditional risk groups that we've accepted and have requested that they self- identify and self-select out of the donor pool.
DR. DAYTON: It's interesting that you mention that you would double the indemnity, because when we did the MSM calculations about a year ago for BPAC, and actually assembled a group that did a great job putting together the numbers, that's about what we came up with is that you essentially double the indemnity, even though you don't know what it is.
DR. BIANCO: I would like to hear ideas from the panel. I'm not sure th