FOOD AND DRUG ADMINISTRATION

+ + + + +

CENTER FOR BIOLOGICS EVALUATION AND RESEARCH

+ + + + +

VACCINES AND RELATED BIOLOGICAL PRODUCTS ADVISORY COMMITTEE

+ + + + +

OPEN SESSION

THURSDAY

NOVEMBER 29, 2001

The Advisory Committee in Versailles Room I and II in the Holiday Inn Bethesda, 8120 Wisconsin Avenue, Rockville, Maryland, at 8:30 a.m., Dr. Robert S. Daum, Chair, presiding.

PRESENT:

ROBERT S. DAUM, M.D. Chair

MICHAEL D. DECKER, M.D., M.P.H. Member

WALTER L. FAGGETT, M.D. Member

BARBARA LOE FISHER Member

JUDITH D. GOLDBERG, Sc.D. Member

DIANE E. GRIFFIN, M.D. Member

SAMUEL L. KATZ, M.D. Member

KWANG SIK KIM, M.D. Member

STEVE KOHL, M.D. Member

PETER PALESE, Ph.D. Member

DIXIE E. SNIDER, JR., M.D., M.P.J. Member

JUAN FELIX, M.D. Invited Participant

THOMAS FLEMING, Ph.D. Invited Participant MICHAEL GREENE, M.D. Invited Participant

PAMELA McINNES, DDS. Invited Participant

MARTIN MYERS, M.D. Invited Participant

DENNIS O'CONNOR, M.D. Invited Participant

SONIA PAGLIUSI, Ph.D. Invited Participant

WILLIAM REEVES, M.D. Invited Participant

ELLEN SHEETS, M.D. Invited Participant

ELIZABETH UNGER, M.D., Ph.D. Invited Participant

EDWARD WILKINSON, M.D. Invited Participant

  A-G-E-N-D-A

Call to Order, Dr. Robert S. Daum 3

Committee Discussion and Recommendations 6

Organizational Structure and Overview of Research and Regulatory Responsibilities in the Division of Bacterial, Parasitic and Allergenic Products, Dr. Richard Walker, FDA 153

Organizational Structure and Overview of Regulatory Responsibilities in the Laboratory of Bacterial Toxins, Dr. Willie Vann, FDA 162

Description of Research Activities, Dr. Willie Vann, FDA 170

Description of Research Activities, Dr. Michael Schmitt, FDA 181

  P-R-O-C-E-E-D-I-N-G-S

8:33 a.m.

DR. DAUM: Good morning. A couple of announcements before we get down to business, so to speak. First, for panel and committee members there are bins up in front for paper that you've carried her laboriously and don't wish to carry home. Please use them.

Secondly, for panel members Denise and Rosana are, as always, kind enough to help us arrange transportation to airports or other destinations. For panel members at the table, please feel free to ask them to help you should you need.

Thirdly, I would like to call on Bill Freas -- where is he? There he is -- to make the briefest of announcements.

DR. FREAS: Thank you, Dr. Daum. I would just like to announcement that at the end of the meeting, whenever that is, that will be at the end of the closed session, we will have a short retirement ceremony for Nancy Cherry.

Let me just take two words to comment quickly on Nancy's distinguished 10-year career at FDA. Committee members know that she's always working late at night which seems to be the norm. But she's also here at this meetings long before I even roll out of bed in the morning to make sure everything is set. We really are appreciative of all the hard work that she has been doing. On behalf of CBER and her colleagues, we're going to have a little cake. We invite the public. We invite everybody on the committee to share this little party with us.

This is the unofficial requirement party just because she won't officially retire until January 3rd but we wanted to have something and to celebrate her distinguished career here while the committee members were here. Thank you.

MS. CHERRY: Thank you. I was trying to keep it quiet until the end of the day but I appreciate it. Thank you, Bill, Bob, everyone.

DR. DAUM: And for committee members and temporary voting members, guests at the table, you've got about three hours to talk her out of it. We're hoping to be able to apply pressure.

I can tell you in a short time as chairman of this committee that no Nancy, no meeting. It's just as simple as that. I'm incredibly grateful for the support and constant vigilance that she provides. Jabs in the elbow notwithstanding, it's been a great collaboration.

The strategy for this morning that I would like to propose to the panel is to have some free discussion first to look at issues that people felt were hanging from yesterday to raise issues that either we need more clarification or that you would just like to hear some committee discussion with regard to the questions only one of which currently fits on the screen but is up there for your viewing pleasure.

Once we get a sense of the fact that we are sort of starting to be repetitive and not raising crisp new issues, then I would like to take stock to address the question directly. At that point we may have heard from half or two-thirds of the panel on the issue but we will ask every member, regular and temporary, to comment directly on the question.

So with that sort of introduction, I've rigged things a little bit with an issue that was on my mind and would like to ask Marty Myers to initiate the first issue. It doesn't mean that we have to stay fixated on this issued. We can wander around on anything the committee's pleasure. Then we will eventually reach a point where we start focusing directly on the question.

Marty, you were kind enough to accept this gauntlet from me and would you start us off.

DR. MYERS: I thought we would talk around the very important issue that I remain a bit confused about. I would like to ask a question to the people who are experts in this.

When we were talking about the contextual issues yesterday, the specific issue really didn't get laid flatly on the table so I would like to put it flatly on the table.

Specifically in Dr. Schiffman's presentation, at least as I understood it, he implied that persistent infection of a year's duration would, in fact, imply that there was a standard of care that would be implied. Somebody with persistent infection might, in fact, require therapy.

As I look at the data, it seemed to me that places a woman at very high risk of high-grade disease and might require long-term close supervision. If, in fact, it implies a standard of care of treatment, then, in fact, the experts in the field have already defined this as a surrogate. It makes it very difficult to consider using CIN 2, CIN 3, for example, a high-grade disease, as an endpoint because everybody will have had intervention before.

My question is really to those people who understanding the management of these individuals. If a person has a persistent infection, does that imply a specific therapeutic intervention or is that a supervision? I think that's a critical issue.

DR. DAUM: I think so, too, and I'm glad people want to respond. Let's start with Dixie and then Drs. Wilkinson and Felix.

DR. SNIDER: Actually, I want to elaborate because I had an opportunity to talk with Dr. Schiffman more about that particular issue which was troubling me greatly as well.

If I understood him correctly, during his presentation he was telling us that the optimal time wasn't really known but that, in his opinion, it was somewhere between one and two years.

The reason he -- if he's in the audience, perhaps he should speak. The reason he came up with one year was not because of the data that he has in hand, but because he has been receiving lots of pressure from organizations who feel compelled given the current body of knowledge to come up with some definition of what recurrent infection is. Persistent. I'm sorry. What persistent infection is. For lack of the more extensive data from his study not being available, not yet being analyzed, the one year is somewhat arbitrary in terms of his personal recommendation. There is some concern on his part that a number of organizations, standard settings, professional organizations may take that number and do exactly what Marty is implying.

It's just a little elaboration, I think accurate, from Mark about how this transpired. Then I, too, would like to hear what some experts think about that particular situation.

DR. DAUM: Thank you very much. Let's continue with Dr. Wilkinson, then Dr. Felix, and Dr. Sheets.

DR. WILKINSON: I would just like to address the issue of persistent viral shedding. The ASCCP guidelines that were developed, these are guidelines not standard of care that were developed in September of this year, had access to National Cancer Institute data that is yet unpublished relevant to persistent viral shedding which Dr. Schiffman alluded to yesterday.

First, let me say that viral shedding in and of itself would not be an indication for treatment but it may be an indication for reevaluation of the patient by colposcopy.

In that setting under the guidelines, and these are submitted at this point, but basically an acceptable -- not recommended but an acceptable statement from the guidelines is that an option in follow-up of women with LSIL, where an option has been chosen to follow the patient rather than treat the patient, the recommendation is colposcopy first, biopsy any visible lesions, with mild dysplasia the option would be that you could follow the patient.

There's only a couple of exceptions. Adolescents and elderly women are some exceptions. The point being that at the end of a year or at some point, possibly two years, your option would be that as an acceptable option to do HPV testing for high-risk HPV type.

If the HPV is positive at that point, you then go to colposcopy, an examination of the patient. If we have persistent viral shedding, there is very good evidence that NCI presented that your patient probably has a persistent lesion.

I would emphasize this is an acceptable option and it's not the standard of care that these guidelines -- ASCCP does not establish standard of care. American College of OB/GYN does so that is something that can looked at at that point.

DR. DAUM: Thank you, Dr. Wilkinson.

Dr. Felix, then Dr. Sheets, and Dr. O'Connor.

DR. FELIX: I'll be brief because Dr. Wilkinson basically stated all the facts. I'll just add that I know of no organization, nor of any expert panel that will recommend therapy based on viral information. They will recommend examination of the patient but never therapy just based on viral shedding.

Clearly not only not the standard of care of, in fact, it's never been recommended officially to actually perform therapy due to viral shedding in itself. Just evaluation and diagnosis.

DR. DAUM: Thank you very much.

Dr. Sheets.

DR. SHEETS: I think there are actually two issues on the table when Dr. Schiffman was talking. I think they have been somewhat blurred in terms of their overlap here. One is the issue of what represents viral persistence in and of itself separate from a side logic abnormality.

I think there is fairly good data to show that persistence of viral shedding six months apart for a year, or maybe two years, is certainly a person who cytologically normal at that time has great risk for the development of a lesion in the future.

That's a separate issue from people who have -- women who have a cytologic abnormality and are concurrently a high-risk viral type. Then we go and subsequently a year later look for presence of that viral type as a surrogate of the lesion being still present on the cervix that gave rise to that cytologic abnormality.

That is a different scenario. That is not what is being discussed as a surrogate marker for failure of the vaccine in this mortality or this current discussion.

Those women who had cytologic abnormalities who had high-risk viral shedding at the incipient visit for vaccine therapy would not accrue in a trial. Correct? So that is a different scenario than someone who is shedding the virus at some point in the future with or without a cytologic abnormality.

I think that when Dr. Schiffman was talking about using viral shedding, as Dr. Felix pointed out, as a point for therapy or further evaluation by colposcopy, that was in the context of cytologic abnormality as the American Society of Colposcopy and Cervical Pathology guidelines indicate for LSIL at this point so there are two different categories.

DR. DAUM: Can we press you a little bit, or can I press you a little bit because that's very helpful. The only circumstances is it true -- is what I'm saying true that the only circumstances that someone would seek viral shedding in a totally asymptomatic woman with no lesions is for research purposes or documentation purposes? There's no medical care issue there at all.

DR. SHEETS: Currently in 2001 there is no medical indication for a cytologically normal woman to be tested for HPV from a medical point of view. There are no guidelines that indicate to do that. This would be a research setting at this point in time.

DR. DAUM: Do I hear in the first thing you said, though, is there talk or plans of incorporating routine screening?

DR. SHEETS: I think there certainly are a body of people in this country who think that HPV could be a surrogate for cytologic evaluation of woman, but that data is not mature for the United States.

DR. DAUM: Not about to happen.

DR. SHEETS: Not about to happen.

DR. DAUM: Thank you very much. That's very helpful.

Dr. O'Connor.

DR. O'CONNOR: I thought about this yesterday and had some discussion with a number of people and what I will give you are what I gleaned from discussions and basically my opinions.

Most papilloma virus infections regress over time. Those that don't are the infections that can result in high-grade dysplasia or worse. The interval before persistence become clinically significant is unknown but it is probably one to two years.

We do not know what factors are necessary for persistence but why only certain HPV DNA types are associated with significant disease. Although persistence carries an increased risk of significant disease, there's no evidence that these woman should be prophylactically treated because what are you treating?

I don't think there is enough evidence to suggest to me that woman with persistent unexplained oncogenic HPV have an inordinately high risk of finding underlying high-grade CIN being defined as CIN 2 and 3.

I feel that based on what I've heard there is, however, enough evidence to suggest that persistent oncogenic HPV has enough of a risk for eventually finding an underlying CIN of any grade that you can it a vaccine failure. That's as far as I would take it

DR. DAUM: Go ahead, Dr. Sheets.

DR. SHEETS: I think to elaborate on what Dr. O'Connor is saying is that when one thinks of surrogate endpoints for this vaccine therapy using approximate surrogate such as HPV, high-risk oncogenic type positivity, or persistence of that presence as a point of failure for the vaccine will slightly artificially increase the efficacy or apparent efficacy of the vaccine since some of these HPV infections by high-risk oncogenic types are transient and clinically irrelevant, not important.

Even some in the face of cytologic slight abnormalities we know will regress over time. Using a marker that is more approximate rather than more distal from the actual invasive cervical cancer rather than more approximate will make the vaccine efficacy appear to be higher.

That is neither here nor there to a certain extent, but if one thinks about the scenario of what we're tying to treat which is either high-risk precursors or invasive cancer, we have to remember that the clientele that we are treating with the vaccine are at least a decade younger than the average age of onset of high-risk precursor lesions and certainly much younger than the incident age of invasive disease.

The question arises here as to what the efficacy of the vaccine will be for those lesions later on a decade or so later. Problems with this that aren't part of the discussion today in the background that one has to keep in mind are that we know very little about the induction of mucosal immunity as compared to serologic markers of immunity induced by a vaccine.

We don't know whether memory in the mucosal immune system will be the same as the surrogate markers and serum for systemic infections. Ten to 15 years later when that 18 and 20-year-old is at greatest risk for the development of precancer or invasive disease, will this immunotherapeutic still apply? We don't know. It's outside of the discussion of this.

But if we use a more distal marker as a surrogate marker of efficacy, or even farther away from the endpoint that we ultimately want to prevent, I would think that is something that we have to think about in terms of discussing the surrogate endpoint.

DR. DAUM: Does the rigor of the definition of persistence matter with regard to your comments? In other words, if we take a one-year period and want four cultures or a two-year period and want six cultures, does that matter or the comment still stands?

DR. SHEETS: I think all it will do is enhance the apparent efficacy of the vaccine to a certain extent because you will be picking up on evidence of HPV positivity that may not be clinically relevant in the long run.

DR. DAUM: Dr. Reeves.

DR. REEVES: I think just one of the things that we are mixing some words and some concepts and you're hitting on it that we are still mixing these. As I understand it, CIN 2 and CIN 3, most of them, or some portion of them, are part of the actual natural history of the development of cervical cancer. CIN 2 leads to or results in CIN 3 in some proportion leads to or results in cervical cancer.

The same is not true for HPV. We're mixing the terms. We're often saying persistent HPV results in or leads to CIN. That's, in fact, not true. It's associated with it.

There's a rather major difference of being associated within a small number of, as are all studies, flawed epidemiologic studies and selected groups, some in populations and some not. But an association is not causal and an association does not imply anything on a path or leading to or resulting in things.

DR. DAUM: That's helpful.

Dr. Felix and then Dr. Sheets. I'm sorry, Dr. Kohl was first. Dr. Kohl, Felix, and Sheets. Excuse me.

DR. KOHL: I just want to emphasize what Dr. Sheets said in terms of something we really haven't talked much about, although it was mentioned in the modeling -- sorry, it was mentioned yesterday -- restoration of protection.

We're talking about almost a life-long risk and, in certain situations, an increasing risk over time, although that seems to be possibly controversial. We heard very, very little about hypotheses of duration or protection.

I don't remember data on duration or protection and that's really a critical issue which I think could accrue in a lodge or a long-term study but I'm concerned whether we would see much of that in a short study with a surrogate that is closer to infection versus closer to CIN 2/3.

DR. DAUM: Thank you for raising that point. We haven't talked it for a bit.

Dr. Felix, Dr. Sheets, Dr. Griffin.

DR. FELIX: I'll actually address two points quickly. Dr. Kohl brings obviously a very important point, duration of protection. But we have to remember that if you protect a woman very early on, that may be in itself even if the protection wanes an extraordinarily important protection because age at first coitus is an extremely important risk factor for the development of cervical cancer.

We don't know what it is about the transformation zone of a very young woman, but clearly woman who start sexual activity at the age of 16 or perhaps earlier have a relative risk that is much higher than woman who start first coitus after 18.

Obviously they are sexually naive. The initial age represents a tremendous increase in the relative risk. If you protect these woman at that age, even if immunity wanes, there is at least theoretical benefit of even those first two or three years of protection in lowering the relative risk of the population for acquiring basic carcinoma.

Immunity even of a transient, I think, is something we ought to seek. Obviously it would be better if it persisted but that is maybe a very important parameter.

In response to Dr. Reeves, I think that the data suggesting that CIN 2 will progress to CIN 3 will progress to cervical cancer is robust. I think that currently there's almost as much data in the literature suggesting that persistence of high-risk viral types if you do it properly will result in the same effect.

Perhaps not at the same rates although very, very close because the rate of progression of CIN 2 is about 20 some odd percent. The rate of acquisition of the high-grade dysplasia from persistent HPV is around 26 to 28 percent also. The data is pretty robust. Both of them are associations but I think they are very equivalent.

DR. DAUM: Thank you very much.

Dr. Sheets, then Dr. Griffin and Katz.

DR. SHEETS: I think there are multiple issues on the table at this point in time for which we have no solid data to make statements one way or the other. I guess I would respectfully disagree with Dr. Felix in saying that stopping an apparent infection by the parameters that we have to test for that infection today ultimately will definitely result in decrease invasive disease in the future. I don't know that. My concern is that when we look at the epidemiology of invasive disease in America, we know that in the late teens, early 20's that these women are at great risk for oncogenic viral infection with subsequent cytologic abnormalities, perhaps even CIN 2/3 which may or may not be caught and treated at that point in time, but there is a large amount of regression through that decade.

We know that in the 30's and 40's slightly more mature individuals are the ones at risk for the reoccurrence or reestablishment of a high-risk lesion histologically that are at risk for the invasive disease that we're talking about.

We don't know what happens in that window. We don't know if the resolution spontaneously of a recursor lesion in their 20's leaves them at great risk for those women, those specific women for invasive disease.

We know epidemiologically that HPV infection is the greatest risk factor for preinvasive high-risk lesions and invasive disease sans sexual partners or age of first intercourse, but we don't have documentation of long epidemiologic studies over a long period of time with no intervention what that biology might be.

If we add on top of that a vaccine which apparently decreases the "insipid shedding of HPV infection," is that the same as never being exposed or having a latent state?

We don't know because certainly there is a great deal of discussion right now in this country in mucosal immunology and HPV research that indicates there may be a latent phase for women who apparently were either treated or regressed their lesion in their 20's redeveloping that lesion later on. We just don't know that data.

In regards to HPV persistence in the development of high-risk histology later, that is well documented that may occur but, again, subject in the 20's, late teens and early 20's, to the same problems associated with spontaneous regression and clinical relevance of those lesions at that point in time. We just don't know what that translates into later than the 30's and 40's.

Some would say that CIN 2 is variable in regression rate whether it exist or not. Listening to Mark Schiffman talk about it maybe it's not even a lesion according to him. Some of us certainly deal with it on a daily basis. That's for sure. It does regress at a fairly high rate compared to documented CIN 3 but that's outside the venue of this discussion.

DR. DAUM: Thank you very much.

Drs. Griffin, Katz, and Fleming.

DR. GRIFFIN: I guess I just wanted to reiterate one point, and that is that I think the data are excellent and nobody has really challenged them, that infection is a precursor -- becoming infected with one of these high-risk HPV types is a precursor to developing cervical carcinoma.

Granted we don't understand everything that's happening during those 20 years before you actually diagnose the disease. Therefore, it seems to me a priori that if you prevent that infection, you're going to prevent the cervical carcinoma.

Now, that doesn't mean that -- then duration becomes important, for how long you're protected. I don't think it means that if you use virus or infection with virus as a marker for the efficacy of the vaccine that you have overestimated.

If you prevent infection that you've overestimated the efficacy of the vaccine, what you've overestimated perhaps more likely is the efficacy for preventing cervical carcinoma but not the efficacy for preventing infection which is usually what we're looking for in a vaccine.

So to me the big question then becomes it would be nice to understand all these things but also whether HPV types will come in and now may play a more prominent role, etc., in the cervical carcinoma that develops in those individuals. I think preventing infection is a very important goal and readout for these vaccines.

DR. DAUM: Thank you.

Drs. Katz and then Fleming.

DR. KATZ: I think Dr. Sheets and Dr. Griffin have helped me in my thinking. We're dealing with two different worlds. One of the gynecologic oncologist and then those of us who think of ourselves as vaccinologists. The terms have been used back and forth inappropriate perhaps.

We're not talking about a therapeutic vaccine. I assume we're talking about a prophylactic vaccine so we're preventing. That's what Diane commented on. Not that we're treating and applying a therapeutic intervention.

Although Dr. Wilkinson showed me wonderfully slides yesterday, I don't know enough about what goes on in the cervix. Are there lymphoid cells? Are there the equivalent of M cells? What's there that provides -- Ellen was talking about mucosal immunity. I know a lot about the GI track and the respiratory track. I don't know anything about mucosal immunity and what to expect as far as local host defense is concerned.

I agree that antibodies may be fine but what goes on locally may be even more important. Are there lymphoid follicles? Is there trafficking of lymphocytes from the cervix to other areas of where we have lymphoid deposits in the body? Can you help me with that at all?

DR. DAUM: Dr. Wilkinson, Dr. Sheets, go ahead.

DR. WILKINSON: I think Dr. Sheets probably has more to say on this than I do but I would say that although the cervix is not considered a molt organ specifically, it is richly endowed with immunologic base cells.

Often these cells are rallied in the face of, say, invasive carcinoma. You can appreciate that in many settings. In certain infections such as chlamydial infection it's not unusual to see aggregates of lymphocytes occurring, a condition referred to as pellicular cervicitis for example.

The cervix also has secretory IgA and so forth. It's quite a complex organ and probably should be ranked among the molt organs but, in fact, is not.

DR. KATZ: So that leads to a little more optimism about preventing infection or reinfection.

DR. KATZ: I want to stay focused in this issue before we go on. When we go on, we'll go to Dr. Fleming next. Dr. Sheets and then Dr. Felix wanted to speak to this very issue.

DR. SHEETS: I think in published data that is currently available in therapeutic vaccines we know that we can give a systemic injection and have cells that were destined -- T cells that were destined for mucosal immunity in the cervix to be exposed to that therapeutic systemically and then track back to the cervix or home back.

We know T cell immunity does happen although at a much lower rate than it would happen necessarily systemically since the dose is given systemically and there's a great discussion of therapeutic vaccines, whether they should be given transmucosally much like the GI tract, etc.

In terms of IgA, IGG secretion, antibody secretion, there's no doubt that the cervix and its mucus has a fair amount of antibody occurring there. I am aware that there are efforts to create the same type of immunologic evaluation that's going on in the cirri that we've heard previously prevented in closed session to do transvaginally to look at that neutralizing antibody from the cervical mucus.

There have been assays set up to do that. The problem is we don't know a lot about the relationship between that mucosal immune system, just as we don't know about certain things in the GI tract compared to the systemic.

We don't know about durability and we don't know about level to a certain extent. This is not known. This is all very new. That's what I was pointing out.

DR. DAUM: Thank you very much.

Dr. Felix, this issue.

DR. FELIX: She presented it.

DR. DAUM: Excellent. Let's move on then. Dr. Fleming.

DR. FLEMING: I'd like to go back to Dr. Kohl, Dr. Sheet, and Dr. Griffin who have brought up a set of issues that have really been troubling me and I was delighted to see that they have pursued this.

I guess I could cast them in the broad sense of what are the durability. What is the durability of effects. What are the long-term protective effects. I think of this in at least two dimensions.

One is what is the long-term protective effect from initial HPV infection that would relate to waning of immunologic response. But the other is what is the long-term impact on rate of progressive disease in those people who are infected.

Dr. Griffin, you had mentioned that the goal of a vaccine is to prevent the infection. My understanding is that with some vaccines the actual true clinical benefit may be achieved by the impact on the immune system in being able to control infection once infection has occurred and what do we know about that in this setting, about long-term impact on the immune system.

I would also say that whereas the effect of the vaccine may be to prevent infection or, in fact, it may be to prevent the sequelae, in essence what to my way of thinking really motivates any intervention is to prevent something that is clinically tangible or meaningful.

In this sense what we've really focused on is cervical cancer. It seems to me entirely likely based on the epidemiology that large numbers of people become infected and the immune system is already capable of clearing the infection in a manner that there are no sequelae.

What I worry about is just because there is this association and it may be causal. If we provide protection in 80 percent or 90 percent, it may be that those are the very people whose immune system was already capable of clearing the infection and, hence, preventing the clinical sequelae.

I think this does become inherently very complex and I think these issues of long-term impact are important not just from the perspective of what's the ability because this is a chronic risk situation. A 20-year-old woman will be at chronic risk for infection.

Beyond that even when you do become infected, what is the overall impact of the vaccine induced immune response on progressive disease, not only over the short-term but also the long-term. These are a lot of questions that I'm very uncertain about.

DR. DAUM: Thank you, Dr. Fleming.

Dr. Snider, then Dr. Katz.

DR. SNIDER: With regard to the issue -- continuing with the issue of preventing infection, I'm still having some mixed feelings about that.

I mean, certainly with hepatitis B, for example, if we had said the most severe consequence of hepatitis B is cirrhosis and hepatic carcinoma and we want to establish a trial to show reduction in cirrhosis and hepatic carcinoma, the size of that trial would have been tremendous and it would have taken a very, very long period of time.

Then the issue of preventing infections that are trivial. We've dealt with this before. I mean, as everybody knows, what are the numbers, Sam? I mean you prevent 100 infections or is it more for every clinical case that occurs.

Right now we don't know how to pick out who is going to develop paralytic polio so we prevent a lot of infections with polio virus that are going to be trivial. It seems to me that -- I understand that the question has to do with intended to prevent cervical cancer and that this is perhaps a little bit off the mark in terms of addressing the questions.

I guess I'm still wondering with Diane if there is not enough evidence to suggest that preventing infections may be something that is quite useful, particularly when I hear that persistent infections are likely to result in maybe not therapy but in terms of additional interventions which I understand the cost of those can't be -- the dollar cost can't be weighed in this discussion but performing these procedures do inconvenience people. They result occasionally in certain morbidity. Then dealing with some of the lesions that will not apparently result in cervical cancer there is not only morbidity but some low-level of mortality from complications.

I guess all I'm trying to say is that I see some societal benefits perhaps of preventing infection which doesn't mean I would give up in any trial design in trying to get a trial design that would also show that there was a reduction in the higher grade lesions.

I don't think we're going to be able to use -- I mean reveal right now cervical carcinoma as an endpoint. I don't think ethnically that's justifiable. Nevertheless, as intermediate endpoints it seems to me those are very worthwhile. The question becomes whether that would be sufficient information to recommend a general use of the vaccine or not.

DR. DAUM: There are three people who want to commend on what Dixie said before we go to Dr. Katz next. First is Karen Goldenthal and then Tom Fleming.

DR. GOLDENTHAL: I just wanted to make a comment about endpoints for vaccine clinical trials. It seems that for most of them, in fact, there has been some type of clinical case definition associated with it. I mean, for example, in polio in the Francis trial, the Francis Field trial, it was really paralytic polio was the endpoint.

With regard to hepatitis, I keep hearing about hepatitis and infection was the endpoint. Certainly in the FDA label it says that the vaccine is indicated for the prevention of hepatitis B infection. But all this talk about hepatitis B also prompted me to go back and look at the smuness and don Francis efficacy trials. In both of those trials there was actually -- they did show a prevention of infection, but they also showed a prevention of hepatitis that was significant between the vaccine and the placebo group. I just wanted to make that point clear.

DR. DAUM: Thank you.

Dr. Fleming. We are going to stay on this very point for a minute.

DR. KATZ: It relates to exactly what Karen has said.

DR. DAUM: Go ahead. But Dr. Fleming is next.

DR. FLEMING: I yield the floor.

DR. KATZ: I hate to disagree with you but the very data you're quoting you do not prevent infection. They showed very well that with hepatitis B vaccine you could have infection as shown by the fact that individuals developed anti-core antibodies.

DR. GOLDENTHAL: And certainly some of them did.

DR. KATZ: So you prevent hepatitis but you don't prevent infection and that's why what Tom was saying I think is to me -- again, I apologize. I'm the vaccine person. I'm not the gynecologist. With most vaccines you do not prevent infection. You abort infection and you use polio as an example.

If you take individuals who have been immunized and don't get paralytic polio, they will shed virus. If they are exposed to enough virus, they'll shed virus for an abbreviated period of time in contrast to the naive individual who has never seen it before. I think the concept that you prevent infection is looking for too stringent a criteria. You abort infection and prevent persistent infection.

DR. DAUM: Thank you, Dr. Katz.

DR. KATZ: Sorry.

DR. DAUM: No problem. There was light shed on issues.

DR. FLEMING: I'm delighted to hear it. I concur.

Dixie, I want to just follow up on your thought about if you prevent the infections. We have considerable evidence. It is association evidence but there's considerable evidence that there is a necessity here. HPV infection is a necessity in the overall causal process that leads to cervical cancer.

What I've been struggling with all along is this issue of sufficiency. You had used as an example, and it's probably a very reasonable approximation, maybe for every 100 infections that you would prevent, you would prevent one case of cervical cancer.

If I knew that if I prevented those 100, I would prevent the one case of cervical cancer, I would be persuaded that I'm achieving something very important. I'm not of the perspective that I have to know if I prevent 100 infections that I'm preventing 100 bad things.

My big concern is that I may prevent -- if, in fact, I have 100 percent efficacy as a result, then I can be confident that when I'm preventing all 100 of the infections with my 100 percent efficacy that I am preventing those 1 percent of the cases of cervical cancer that will follow.

My concern is if I prevent 80 of the 100, I may well be missing the one, in fact, that would have resulted in cervical cancer. If I have 80 percent efficacy or 90 percent efficacy even, I may have almost no efficacy against what I really care about. It's the sufficiency issue that I keep saying. It seems to me that because this is a setting where the numbers suggest that it's something on the order of 50 or 100 people who will have HPV infection for everyone one that eventually will over their lifetime have cervical cancer, this is clearly a situation where it's far more complex than simply saying is the vaccine going to prevent the initial infection.

What I'm struggling with here is what is it that we have to achieve in order to be confident that we are actually having a meaningful impact on what we really care about which is reducing the rate of occurrence of clinical events.

Now, we focused on those clinical events being primarily cervical cancer. I would, however, accept a broader sense of clinical events, i.e., if we believe that we are also achieving a reducing in the need for invasive surgical interventions, etc., that's part of the overall benefit as well, although I think our highest priority clinical event here is

preventing cervical cancer

So the bottom line is we acknowledge we're preventing many more cases of infection than we are clinical sequelae, important clinical sequelae. I just want to know that when we do get this reduction it translates into a meaningful reduction in cervical cancer.

DR. SNIDER: Could I just quickly respond and say, Tom, I think you and I are in agreement. All I'm saying is that if we don't look at persistent infection as one of the endpoints, it seems to me that would be a shame because we're not preventing persistent infection. I'm not optimistic that CIN 2 and 3 are going to be prevented.

DR. FLEMING: So you're saying that's in your vision of what the markers would have to be. That's one of the necessary components that has to be impacted.

DR. SNIDER: Right.

DR. FLEMING: I'm very willing to accept that. I'm struggling with what are the other elements that give me a sufficiency condition here such that when I see persistent infection and what else, is this going to be adequate to be reasonably confident I'm having an impact on cervical cancer rates.

DR. DAUM: Okay. We're going to stay on this issue before we go on so people who want to talk to this very issue. Dr. Reeves, Dr. Felix, Dr. Kohl.

DR. REEVES: Just a quickie of something that I would have liked to have heard and I don't see any of the NCI people that can give the answer. I think this vaccine to prevent cervical cancer is going to be unique among vaccines. Diphtheria, influenza, and many of the vaccines I'm aware of work very quickly.

Rolando Herrera, I believe, two years ago presented some very elegant modeling studies of the effect of vaccination on the rates of cervical cancer world wide which, again, is the end disease we're trying to deal with.

In essence he showed that it was going to be approximately two decades before any effect was seen. I think this rather important information, something to take into consideration both in looking at whether we're going to approve or recommend approval for accelerated licensure. But two to three decades is a long time to actually see an effect from something. I suspect that the actual effect on surgical procedures for high-grade dysplasia or CIN 3 with the actual public health effect and efficacy of this is probably going to be in terms of decades as well. I think some kind of presentation of that kind of information would have been very helpful.

DR. DAUM: Dr. Felix and Dr. Kohl, this very issue we're on.

DR. FELIX: I appreciate the concerns that Fleming has. I have the identical concerns. However, if you're proposing that by producing 80 percent or 90 percent of the HPV you may not be reducing the 10 percent that will proceed to cancer. The same identical argument can be used for the more distal surrogate endpoint which would be high-grade dysplasia or CIN 2, CIN 3.

If you prevent 90 percent of CIN 2, CIN 3 it is perhaps that 10 percent that you don't progress that you don't protect for that will progress to cervical cancer. I don't think it is reasonable to expect a trial with an endpoint of cervical cancer. I don't think it will happen if that's the case.

I think Dixie was correct. I think we need to keep assurances that all of the reasonable endpoints will be looked at, that we're going to look at persistent virology, and the issue that I'm most concerned with that I hope we are going to address very soon, the issue of what interval does persistence truly become meaningful.

Then have relative assurance that we are going to see CIN 2, CIN 3 data. It is, I think, within the realm of this committee to insist that the trial for the latter be finished by the time the accelerated approval for the virological endpoints come out so that we could guarantee that the second trial or the second observation would happen.

I don't think that it is reasonable to expect anymore than the surrogates that are still going to leave doubt as to the efficacy of the vaccine.

DR. DAUM: I think in our own way we are starting to build consensus.

Dr. Kohl next. This very issue. We're still there. Then Dr. Unger, Katz, and Sheets.

DR. KOHL: I'm feeling a consensus also hopefully, what I'm thinking is the consensus as the same thing other people are thinking.

DR. DAUM: We'll find out.

DR. KOHL: Being at this end of the table, all the way at the end, this is the Dixie memorial seat down here, I'm trying to think as a virologist now. We are dealing with a virus but a virus that has an interesting effect, namely cancer.

I'm trying to think what we know about -- at least what we've been presented about the immunology or the protection against first infection. Perhaps more importantly the immunology against cancer, the prevention of cancer.

I don't think we've heard much to anything about the immunology or the prevention of cancer. Most of what's in the literature, that I'm familiar with at least, regarding the viral like particles is the elicitation of neutralizing antibody.

Yet, we know that the -- or we think we know that what causes cancer are the E6/E7 transforming elements. Then there's the whole issue of latency. What I want to get around to in a sort of sequitious way is following some of what Dr. Fleming is talking about. What if we have that heterogenetic population where a small percentage, because of immunological aspect we don't understand, is very susceptible.

And what if paradoxically neutralizing antibody doesn't have a positive effect but has a negative effect? It's wild. It's outside the box, but it's one of those things we just don't know about. I think all these uncertainties would push me towards a more rigorous endpoint as we think about surrogate endpoints.

DR. DAUM: Thank you.

Dr. Unger.

DR. UNGER: I just want to remind everybody about the difficulty in the assays in talking about infection and persistent infection.

DR. DAUM: Talk right into the microphone.

DR. UNGER: Okay.

DR. DAUM: Thanks. Sorry.

DR. UNGER: I'll start again. I just want to remind everybody about the difficulty of establishing infection, the difficulty both in the assays and the sample. I think that we need to be sure that the sample is taken appropriately and the appropriate amount of the sample is put into the assay.

You can have the most eloquent and sensitive assay in the world but if the sample is not the appropriate sample and enough is not put in, it's going to make your definition of endpoint and infection a moot.

I think that the literature is very clear that the assays and the sampling will muddy the kind of pictures that you see. We need to be clear on what kind of documentation we want to see or should be part of looking at this kind of persistent infection.

I think that the better the assays have become and the more standardized the sampling has become, the clearer the picture is as to what the situation -- not that it's clear now but there is starting to be some consensus.

I think part of the confusion is the definition of persistent infection and the timing that should be required in order to say what persistent is versus a normal endpoint of clearing of an infection that would go away on its own.

DR. DAUM: Dr. Katz at last.

DR. KATZ: I would like to go back again to what Dixie has said and what Karen has said. Viruses are all different and it's very inappropriate to make generalizations because this virus does this, that virus does that.

But the example that's been used is a reasonable one of hepatitis B. Why did they start looking for hepatitis B vaccines? Not just to prevent acute disease but because Palmer Beasley showed in Taiwan where they had a high incidence of hepatocellular carcinoma and that hepatitis B infection led to hepatocellular carcinoma.

The studies that have gone on there over the years now have shown they markedly reduced hepatocellular carcinoma to a rare disease in Taiwan because they gave vaccine to young people.

Now, it does prevent hepatitis over disease but it doesn't prevent occult infection and you may have occult abbreviated infection. This, as I mentioned in response to Karen, is shown by the fact that the vaccine only gives you antibodies to one antigen, the surface antigen.

You can show that vaccinated people, though they don't have the disease, develop antibodies to the core antigen which indicates they have not only been infected but they have been infected sufficiently to arouse an immune response.

Those people who have totally lost detectable antibody to the surface antigen nevertheless resist developing clinical disease or chemical disease. We have a model which is not a perfect paradigm, but I think we do have a model, at least, of where preventing an infection from developing beyond an abortive state does prevent the development of cancer.

I think the long-term effects of this can be in some ways analogous, if you will. Not a perfect one but I'm encouraged that if you can prevent infection with these oncogenic papilloma viruses you may well prevent cancer.

DR. DAUM: Thank you, Dr. Katz.

Dr. Sheets.

DR. SHEETS: I guess I'm simply a gynecologic oncologist. Not a virologist nor immunologist, nor a vaccinologist. When I think of human papilloma virus infection, I think of the transvaginal infection that may or may not ever be systemically manifested. Even invasive cancer you may or may not show systemic antibodies to E6/E7 my understanding is.

When we think about this vaccine and we think about this vaccine and we think about proximate surrogates or distal surrogates as to what that might eventually prevent invasive cancer, we have to think about what's happening with the mucosal barrier.

The vaccine is supposed to prevent infection by neutralizing antibodies being present in cervical mucosal discharge that keeps the virus from infecting the epithelium. That's my understanding of what the vaccine is supposed to do.

We don't know, and I think hepatitis is certainly a good surrogate systemic infection to look at for the development of a cancer. But what we don't know is the latency issues of HPV. We don't know that.

You are discussing the fact that latent virus associated with hepatitis B may cause -- does cause hepatocellular cancer. Eradicating that virus you may get infected but having the antibody potential systemically to kill that viral infection does preventing the latent state leading to hepatocellular cancer is a surrogate marker for HPV infection.

I simply don't know that. I don't know if that's true, but I think it's out of the venue of this discussion to decide whether we're going to move forward with a fast track for this vaccine or not.

I think what it underscores is the fact that we don't know how HPV induces ultimately cervical cancer in the epithelium and what the immune response plays in that role for therapeutic interventions or prophylactic interventions.

But we have to assume that the stuff the vaccine is causing is simply to block the infection itself and may have secondary effects of T cell responses, etc., etc., should there be a small amount of virus that penetrates the epithelium and causes a T cell response and we do get efficacy in that system. We don't know that yet.

Maybe the NCI knows that but I don't at this point in time. I think we have to look at the data that's here and the decision that we have to make is one step away from cervical cancer. The question is how many steps away will we allow. If we allow it to be too far away, will we ever know the real answer. My concern ultimately, and it's a step beyond what we're talking about now, is the apparent efficacy of the vaccine is so great for preventing infection will we ever be able to carry a placebo group forward.

DR. DAUM: Can you help me with one more expansion of your comments?

DR. SHEETS: Maybe.

DR. DAUM: Are you suggesting that there is a scenario -- supposing we had a crystal ball here and we knew that this vaccine was being universally used now and it meant that HPV infection was efficacy 100 percent prevented. Can you imagine a scenario with that being true where it would not have an impact on cervical cancer?

DR. SHEETS: 100 percent efficacy for both male and female?

DR. DAUM: Yeah.

DR. SHEETS: So that you're not re-exposing them chronically to the virus?

DR. DAUM: Yes. Let's go whole hog.

DR. SHEETS: How could it not? If you eradicate HPV it would impact. No doubt.

DR. DAUM: Okay. Good.

DR. REEVES: It would be next on the list behind measles.

DR. FLEMING: Let me just make sure your question is clarified. When you say 100 percent that suggest to me that you mean 100 percent across all types and 100 percent across all time. Then if that is the case, then I'm happy to say yes, too. There's a lot to that question.

DR. DAUM: Let me clarify. Let me say all time, yes, but all types, no, only the types in the vaccine. Yes, 100 percent against all the types.

I'm trying to get a sense from people who really understand the subject which does not include me, whether or not it is conceivably possible to prevent HPV infection completely and at the same time not assume that cancer is prevented also. That's what I'm trying to understand. Is there such a scenario?

Does anyone want to speak to this? Diane.

DR. GRIFFIN: No. If you're talking about HPV 16 induced cancer, if you prevent all infections, you will prevent HPV 16 induced cancer. I also want

-- I mean, I think the links are extraordinarily strong and we certainly understand a whole lot more about how HPV induces cancer than about how hepatitis induces cancer which, as far as I understand, we have relatively little understanding of that pathway.

We understand that much better for the HPV. We don't have a perfect understanding of that but it does require infection and it does require infection that is over some period of time and I don't know what that period of time is.

I think you are ignoring a lot of virologic data that has come in for a long period of time about these links and about the pathogenesis of this process.

DR. DAUM: I'm going to try and maintain some sense of order here. I wrote the rules myself and jumped in, but I have Dr. O'Connor first, then Dr. Felix, Ms. Fisher, Drs. Reeves, Katz, Kohl.

DR. O'CONNOR: I get the impression there are a lot of topics floating around here. I wanted to go back and address endpoints for just a minute and say very quickly that I agree with both Dixie and Juan as far as the endpoints go.

I think there is enough evidence to indicate that persistent papilloma virus infection is associated with CIN to the point that it can be considered a vaccine failure, surrogate or not. Certainly identification of it is accelerated enough that it might be considered surrogate.

I think there is excellent evidence to indicate that CIN 3 is associated with cervical cancer, although the information regarding CIN 2 is not as clear because criteria for diagnosis are not that reproducible. I think there's enough there to say that CIN 2 should be lumped in with CIN 3. CIN 1 doesn't work just because it's a polyglot and the diagnosis is extremely irreproducable.

The last thing I want to say is that we're talking really about histologic diagnoses and not cytologic diagnoses. You need to be clear on that. Even though the specificity of cytology gets better as the abnormality gets worse, you still have a significant number of HSILs, and I'm talking about cytology, that will have no dysplasia or low-grade dysplasia on biopsy.

I think it's best to leave that as a screen test. When you're talking about endpoints talk about a directed biopsy or excision procedure that will give you a histologic diagnosis.

DR. DAUM: Thank you, Dr. O'Connor.

Dr. Felix.

DR. FELIX: I am going to have to politely reverse Dr. Sheets' disagreement with me and disagree with her. I don't think that necessarily the function of the vaccine is to prevent infection. I think that you can have an extremely efficacious HPV vaccine if you abort infection early.

In other words, induce regression at an accelerated rate. We know that regression results in prevention of cervical cancer. I don't think that you necessarily have to prevent infection in order to make an effective vaccine. Obviously the examples have been brought forth for hepatitis B.

I think that it's a very reasonable analogy to make at this point. I think if you have cellular immunity that will act in aborting a lesion early on, you can, in fact, enhance prevention of cervical cancer.

DR. DAUM: Thank you very much.

I have Ms. Fisher, then Drs. Reeves, Katz, Kohl, Palese, and Kim.

MS. FISHER: In terms of the idea of eradicating HPV infection by vaccinating all women and men, how do you know you're not going to put pressure on an organism to change into a vaccine resistant form when you're only using certain types such as HPV 16 and 18?

DR. DAUM: That's a provocative question. I don't think we do know.

DR. GRIFFIN: You won't change those into new types but you may have the opportunity for other types to now fill those niches and we're not going to know that until we do the studies. That's the reason one of the things that needs to be incorporated into the studies is looking at these other types.

DR. DAUM: People would have to be mind of those things, I would think.

Dr. Reeves.

DR. REEVES: I had a couple of points and they kind of go back a bit. I disagree completely that if we eradicated HPV from the face of the earth, all types of infected genital mucosal, that we would necessarily prevent cervical cancer.

If, for example, we eradicated hepatitis B with a vaccine program and we eradicated hepatitis C, we would not, in fact, eradicate hepatocellular carcinoma or cirrhosis.

DR. DAUM: Due to those agents?

DR. REEVES: I'm talking about the disease because the disease is a complex multi-factorial disease of which HPV is currently the most important associated risk factor.

Unfortunately, I remember in the old days, and I think same probably remembers this, too, when herpes II caused this disease. It is not necessarily a simple disease. We have an ideologic agent highly associated with the disease and vaccine will probably have a major effect on it.

I go back to hepatitis B. The timing of the vaccine, as I recall, was very important in the prevention of hepatocellular carcinoma. It was infections of young children I think more associated around the time of delivery or transplacental transmission that was important. The timing in which this vaccine is given is very important. We talk about naive women, women who have not been infected with the agent before, that's probably not the group that's going to be vaccinated. I don't think we always know what naive means in terms of this agent.

Finally, there is at least one agent that I am aware of, unless it has changed, respiratory censishal virus, an apparently very good vaccine made worse. That possibility --

DR. KATZ: It wasn't a good vaccine. That's not fair.

DR. DAUM: Can you clarify one thing that you said? If you prevented, let's say, two serotypes of HPV, would you prevent an infection by those two viruses? Would you prevent cancer caused by those two viruses?

DR. REEVES: I think what we want to do is prevent the affects of the infection, so preventing the affects or ameliorating the affects of the infection. Preventing the infection would obviously do that but one would not have to prevent the infection to ameliorate the affects of that infection if that involves integration, over expression of E6/E7, etc. I think obviously preventing the infection would prevent the disease that resulted from that infection, yes.

DR. DAUM: Thank you. Thanks very helpful.

Dr. Kohl, Dr. Palese, Dr. Kim.

DR. KOHL: I want to genteelly object to one of my chairman's constructs.

DR. DAUM: For the first time.

DR. KOHL: Absolutely. He proposed the possibility of 100 percent prevention of an infection and then the resultant 100 percent prevention of a disease associated with that infection, namely, prevention, let's say, HPV 16 and then prevention of HPV 16 associated cancer. I would agree that is probable.

But I think as one of my favorite people, Ross Perot, said, "The devil is in the details." He's not really one of my favorite people. I can't think of any vaccine -- any vaccine, let alone a mucosal vaccine, that is capable of preventing 100 percent of infection. I can't think of that as a possible scenario.

Therefore, I'm left with some finite percentage of people in whom the prevention won't be complete, who will still get infected. It's that small percentage, because they have some unknown immunological situation that Barbara Fisher alluded to yesterday, that causes some people to progress who I'm most concerned about.

Do they have latent infection of some kind? Does antibody make that worse? I don't know. It's that little group of people, 5 percent, 10 percent, 15 percent, that I'm concerned about and that's a big unknown with this vaccine.

DR. DAUM: Steve, my comments were by way of requesting information from experts to try and get at the solidness of the link between infection and the consequences. Of course, it can't be 100 percent effective but in hemophilus there are some people who are clearly still at risk of disease because we still have a few cases occurring despite full immunization.

DR. KOHL: In some of them we know why.

DR. DAUM: The 100 percent was a hypothetical discussion.

DR. KOHL: But it muddies the water, I think, because it leaves out that 5 or 10 percent who will still be infected for sure and whom we know very little about why that's the case and what a vaccine will particularly do in that setting in those people.

DR. DAUM: Thank you.

Dr. Palese.

DR. PALESE: I just want to raise the question about the safety of the preparations which are being discussed right now. These are, if I understand it right, inactivated so they are viral like particles.

I want to ask whether there is any evidence that they have any unacceptable side effects, or that there were any exacerbations of any kind of disease associated with giving this experimental preparation. Is anything known and what is the longest time period that we can consider here so we can have the earliest preparations being administered? Is there anything known? Have we heard anything?

DR. DAUM: I'm going to call on Dr. Goldenthal because my sense is that although safety is crucial to any plan to go forward with the deployment of this vaccine, it hasn't been among the things that we've been asked to consider today, at least head on. How would you like us to take up this question of safety, Dr. Goldenthal?

DR. GOLDENTHAL: Well, I think that it's a reasonable topic for discussion, especially when we get to question No. 2 because one of the issues there would be potentially the amount of safety data that we would have prior to licensure. I think it's a legitimate thing.

In answer to your question, I think I can say in general there's been maybe three years or so of follow-up on individuals who have received VLPs in various trials. The numbers are fairly limited at this point. Maybe a few thousand at most.

It would be hard based to make, you know, based on -- while there's nothing that's been troubling that I'm aware of, it also would be hard to make a lot of conclusions at this point.

DR. DAUM: I think when we focus more on this accelerated approval question and I think we need to return to this issue, at least only to state what we believe would need to be done before we would be comfortable.

Dr. Kim.

DR. KIM: Well, we heard a lot about some aspects of HPV infection and how infection would either regress or persist. Again, we also talk a lot on the issues of a persistent infection which has been very arbitrarily defined and interpreted amongst all of us.

I have not got the concept yet. What is the biological relevance of persistent infection, particularly as it relates to CIN 2 and 3, not cervical cancer at this juncture?

DR. DAUM: Thank you.

Dr. Goldberg.

DR. GOLDBERG: My question -- it's a comment really. We saw a lot of data on different intervals for defining persistence, the time between the two successive observations.

It seems to me that a study such as the one we heard from the NCI yesterday should allow us to be able to look at the distribution of lengths of persistence in a large population and then relate back to later events.

I would like to see that kind of thinking incorporated into the trials that are designed regardless of what endpoints we choose because I think this will be relevant as we go forward.

DR. DAUM: Dr. Sheets is scheduled to speak next and maybe I would ask before you make what comment you wish, could you address Dr. Goldberg's question in that if persistence is going to be used as an endpoint, vis-a-vis question 1b, then what definition does a real expert in this recommend that we use? Clarify your question first.

DR. GOLDBERG: Okay. I'm not convinced that I saw anything that would give me great comfort in any of the definitions. What I'm suggesting is that as we design trials going forward that we incorporate the ability to look at the distributions of the lengths of time between the successive positive tests for HPV. I think particularly the information from the control groups will inform out thinking with regard to the influence of this on the later endpoints such as CIN 2 and 3. What I'm thinking is that if you cut that interval too short, you're taking away all the cases.

You're using cases that would have resolved by themselves that will have no impact. If the interval is too long, you may be practically there. I think you can get some information.

I think the NCI trial that we heard yesterday, if I understand the data correctly, and if the remainder of the cohort other than the ones that were positive at entry are examined over time and you may get some important information.

It's sort of like developing a receiver operating characteristic curve on different cut points for the definition of what the interval between successive positive HPV tests are that would be meaningful later.

DR. DAUM: Now, Dr. Sheets. Thank you.

DR. SHEETS: I don't think I can speak specifically to the NCI close session data talked about yesterday.

DR. DAUM: Nor do we really want you to.

DR. SHEETS: But I do think it's relevant to say at this point in time in 2001 that we don't know the answer to your question in regards to what represents a persistent infection that's clinically relevant, or if that is even important depending on your time point or endpoint or what you want to prevent.

Ultimately we want to prevent invasive cervical cancer, both adeno and squamose. That's the goal here. We don't know whether we can translate HPV presence, high-risk oncogenic presence, specific to the viral type being vaccinated against as being a surrogate for that or not. That's the discussion I think is on the table.

I'm not an expert in that in terms of virology persistence, but I would say that I don't know yet from the data that I've seen in the world's literature, nor heard in closed session that I can make that statement. It should be incorporated into whatever trial we decide is endpoints.

I guess within the bounds of what can be presented here in open session compared to closed, I would like to hear from Doug Lowy his point of view in terms of what this vaccine or vaccines in general that are prophylactic would probably be the best way to phrase this. A prophylactic vaccine using VLPs theoretically should be doing for us in terms of how it interrupts the HPV cycle or if we know that at all in this point in time because I think it's an important consideration in answering Dr. Felix's disagreement with me, disagreeing with me over what we're agreeing as to whether you actually do get an infection, yet dissipate that infection so it doesn't become clinically relevant and the vaccine does do that for us.

My concern is that we're using a systemic system both across the table and over here, an hepatocellular infection that is not necessarily the same as what we're talking about here today in the mucosal immune system. I'm just interested in hearing what Dr. Lowy could point out in that. Is that possible?

DR. DAUM: Is Dr. Lowy here?

PARTICIPANT: Yeah. Right there.

DR. DAUM: Do you care to comment on this? You're not obliged to.

DR. LOWY: Ellen, thank you very much. I think that the issues that are being raised are very pertinent and relevant to the discussion. My colleague, John Schiller, may want to amplify on some of my comments.

My sense of the VLP vaccine is that it is going to be doing -- it is basically going to reduce the inoculum. We haven't talked much about viral inoculum but with most infectious diseases the size of the inoculum has a very important impact on disease downstream.

By reducing the inoculum there should be one of two outcomes. One is that you would completely prevent infection, and the other would be that you would reduce the number of infectious hits.

There also is a possibility it's ambiguous whether the target, which is the transition zone of the cervix, is the immediate site of infection or whether there might be a remote site distant from that.

You could imagine that antibodies might have a further impact on reduction of going to the site of the target, if you will, analogous to the hepatitis situation where you get infection but it doesn't get in sufficient numbers to the target. I think it's ambiguous which way that would work. In the best case scenario it would be a complete prevention of infection but I certainly wouldn't expect it to do that in all individuals.

In terms of the long-term persistence of antibodies, I suppose it's hypothetically possible that might have an adverse impact, but there is no theoretical reason to believe that you are going to be -- that it would have such an impact.

We haven't seen in the limited trials that we have done which involves maybe 100 individuals, we haven't seen a group of people who are particularly resistant to responding in terms of immunity or particularly susceptible when we look at the bell-shaped curve.

The concern of Dr. Kohl that maybe you're picking out a particular group of people, I think while it's hypothetically possible, I don't think we have a coherent notion that the latent infection would be more likely to be more serious because of antibodies being present, although I think hypothetically that might be a possibility.

With regard to persistent infection, I think that Dr. Fleming is, of course, raising a very important issue about the duration of infection.

It's one of the reasons why when one picks persistent you would like to have a relatively long period of time, thereby increasing the probability that by reducing those persistent infections that you really would be having an impact on the clinically important downstream events.

The precise number whether it should be six months, 12 months, 24 months is going to be somewhat arbitrary which is what Mark was trying to point out yesterday. There are some data now, and there will be better data.

Even when you get the better data it's going to be a balancing act. I guess with Dr. Wilkinson, I think that he raises a very important issue of the question of referring people for colposcopy.

My question for Dr. Wilkinson would be if you get referred to for colposcopy, what's the likelihood that you would be biopsied? Because if you were going to be biopsied, then you presumably would be out of a clinical trial.

DR. DAUM: Having said all that and being practically oriented, given all your expertise and given the arbitrary nature of the decision that I'm about to ask you for, if you were to pick, and emphasis on the word "if," persistent infection as an endpoint, what definition would you use for that?

DR. LOWY: I really am relying on Mark because he is our expert. He is our medical epidemiologist. He feels that an appropriate balance would be a year, that you will be clearing out most of the, if you will, clinically irrelevant infections and it will have a high predictive value of preventing significant proportion of the downstream events.

DR. DAUM: One year. Thank you very much. Okay. The next people to speak are Drs. Snider and Myers. Are these clarifications of this very issue?

DR. SNIDER: Yes.

DR. DAUM: Okay. Then Drs. Griffin and Snider. Dr. Snider, you're next anyway. Why don't you go first and then Dr. Griffin on this very issue. Then we'll go on to Dr. Myers next.

DR. SNIDER: I would like to just pick up on the comments that were just made. The trade off, as I understand it, is even more profound in the sense that it's not just specificity of the study endpoints, but there are some clinical implications, some ethical implications in terms of the intervals you choose.

If you choose a shorter interval, of course, then you have the possibility that's already been mentioned or the certainty that's already been mentioned, that you'll be calling a lot of endpoints, significant endpoints which are not significant in the sense that they will regress.

There also is a clinical corla in the sense that it sounds as if whatever interval is chosen, there will be some interventions that again will have some not only economic cost but some morbidity and at least psychological morbidity and physical morbidity associated with them.

The longer you go the more specificity you get, but then if I understood correctly, for two reasons you may wind up with more cancers. One is because there are a few women who would rapidly progress. If you went to two years, for example, there are a few women that I think he said you would lose. I think he said you would lose but I think what he meant was they would progress quite rapidly to cancer.

The other, of course, is this whole issue of compliance in clinical trials. The longer you wait, the more you signal that this is not all that important and women start dropping out and they don't come in for that two-year visit.

Again, you run the risk of having these very serious outcomes that have more morbidity and perhaps even mortality associated with it. It is a delegate balancing act.

I just wanted to at least indicate some of my understanding about some of the value judgements and some of the ethical and other implications of making that choice. There is no exactly right answer right now.

The other thing that has to do with persistent infection in terms of how you define it is not just the interval but it has to do with how many specimens you want. Also, as has been pointed out, how you obtain those specimens and what assay you use. These are all critical issues in terms of defining persistence to have to be looked at very carefully. I'm not sure this committee can get into all of those details but there are some general principles, I think, we could probably articulate about what we would like to see with regard to the intensity in which one investigates and the characteristics of the test.

DR. DAUM: Thank you, Dixie.

I have Drs. Griffin, Myers, Freeman, Pagliusi, and Kohl.

DR. GRIFFIN: I just wanted to comment on and get Doug to expand on the reason that persistent infection is such an important part of the pathogenic process that we are trying to prevent and also examine in these women.

It's my understanding that the longer the virus continues to replicate in its site, the increased likelihood that you'll get integration, which is a random event, and other oncogenic changes in those cells that will then eventually result in carcinoma.

That is sort of the biologic principles under which one becomes interested in persistent infection and the length of persistent infection. But I would like Doug's comment on that.

DR. LOWY: This is a series of genetic changes presumably and the more opportunity you have in terms of chronologically, the more likely it is to happen.

DR. DAUM: Thank you.

Dr. Myers, please.

DR. MYERS: I'm not a papilloma virologist but I think we need to be careful about some of the terms we're using like inoculum and persistence and latency and replication because we really don't know how to measure those in the circumstances that we're talking about.

I guess the question is, to go back to the comment that you made, the reducing inoculum. Is that really the intent or is what we're trying to do is alter the natural history of persistent outcome? I think that's important if you go back to Dr. Reeves' comment earlier.

This vaccine is not going to just be given to naive individuals. I think we need to explore -- and we haven't really talked much about this but we really need to explore the intent to immunize the outcome from the intent to immunize.

When we're talking about persistence and when we're talking about high-grade disease, we need to address that in individuals that are both HPV 16 and 18 positive as well as naive because we really don't understand these virologic events in the natural history of the clinical setting. I think we've been skirting that issue.

DR. DAUM: Thank you, Dr. Myers.

Dr. Freeman.

DR. FREEMAN: I just wanted to make a brief comment that the choice of these endpoints and, in particular, the precision with which these endpoints can be determined, I think, are incorporated into a trial that would lead to an accelerated approval or further I think are very important.

Not just from the point of view of demonstrating that the vaccine works but in convincing the subjects who will eventually receive -- the males and females who will eventually receive this vaccine if this thing actually works.

I'm reminded of the comments of one of the advocacy groups from yesterday that if the vaccine is approved, it really has to be meaningful in order to get compliance and usage to do what it's supposed to do.

The other thing is the physicians who are going to administer the vaccine and monitor these patients safely have to have a good idea about the -- have to be convinced that the trial really demonstrated the efficacy of the vaccine in terms of the way they understand the disease process.

I'm not sure from all that I've heard, although I am convinced of the association that has been mentioned, the association between the HPV viruses and this disease. I'm not sure how easy it's going to be to rely on the HPV endpoints as indications for usage of the vaccine practically.

DR. DAUM: Thank you.

Dr. Pagliusi.

DR. PAGLIUSI: Thank you. I would like to come back to the persistence of infection to the balancing act. I would like to address a question to the experts. Maybe Doug Lowy could help me here.

If we would think of measuring persistent infection twice, three times, four times, what that increase the confidence and the effectiveness of the vaccine or the efficacy of the vaccine?

DR. DAUM: I'm not sure I understood the question. May I ask you to clarify?

DR. PAGLIUSI: My question is addressing persistence of infection. Dr. Lowy proposed that one year may give us more confidence on the results. My question is now if within this one year we would see three positives or four positives.

DR. DAUM: Or two.

DR. PAGLIUSI: Or two, what are the balancing here?

DR. LOWY: Sonia, I think that certainly it would be preferable under ideal circumstances to sample more frequently and to have recurrent positive results. My impression is that if you had two positives separated by whatever interval it was, you could then go back and be quite sure that this was with the same variant or not with the same variant. Then you could be quite sure that the individual was continually infected with the same virus or a different one. I think that would be adequate.

DR. DAUM: Thank you.

Dr. Kohl.

DR. KOHL: I was a little confused. Oh, Dixie's not here. I think Dixie was implying that different time periods, i.e., a 12-month persistence or 24-month persistence would some how affect and I think he used the words "lose some women."

I was not under the impression that the amount of time that was chosen for persistence would per se affect how woman are followed for cervical cancer screening and that women would still be followed according to standard of care no matter what the time period were for what was decided as persistence. Is that correct? So no matter what you pick as a definition you're not going to "lose people."

DR. DAUM: I think we're getting to a point where people are locking in their ideas about what would be the endpoint they most favor. But before we really start systematically debriefing everybody of those few points, I would like to ask people to consider this.

Is it possible, Karen, and I hope this is within the spirit of your first question. Is it possible to consider multiple endpoints? For example, is there a possibility of designing research, a clinical investigation, a vaccine trial if you will, that looked at different endpoints in sequence with each other and had sort of separate analyses for these different endpoints and sequence? Is that a feasible way to think about this, or do we need to focus on just one?

Karen, I would like you to respond to that first and then maybe others.

DR. GOLDENTHAL: I believe you could design a trial that way. That isn't the way we have ordinarily proceeded for preventive vaccines, but I think it's theoretically possible obviously with rigorous prospective statistical analyses plans and designation of endpoints.

DR. DAUM: Does anyone want to comment on that thought or that idea?

Dr. Kim.

DR. KIM: I was given the information that somehow linkage has been not clearly delineated and some question. I support the concept that perhaps two endpoints can be incorporated. For example, the first one would be persistent infection but second would be truly translated into CIN 2 and 3 as secondary endpoint.

DR. DAUM: Okay. Thank you.

Dr. Decker. We haven't heard much from you today.

DR. DECKER: Or yesterday.

DR. DAUM: Or yesterday. Here's your chance.

DR. DECKER: I'm glad you brought up a point you just did because I've been thinking about that. It seems to me that if it ends up being decided that the primary endpoint for a trial would be nonvirological.

Then I think it would be imperative that their be co-primary or strong secondary endpoints that were virological, if for no other reason than so that future trials would be guided by the understanding of the links between the virological and the clinical outcomes.

To me it almost goes without saying that it would be essential that there be virological surveillance and virological endpoints measured in any trial whose primary endpoint was clinical.

DR. DAUM: Thank you.

Dr. Snider, Dr. Fleming next.

DR. SNIDER: I's just like briefly to address that point, too, I think in view of the evolving knowledge base, the rapidly evolving knowledge base around this issue, having that kind of a trial may be not only advantageous to the FDA, this committee, but to the manufacturer as well because it allows you in one trial to make adjustments as new information becomes available rather than going out and having to redesign the trial. There's a lower risk of having to redesign trials.

DR. DAUM: Dr. Fleming is next.

DR. FLEMING: Bob, I strongly endorse your thought. I think in this setting where there is such uncertainty, and I think Steve had mentioned it earlier, it certainly leads me to be more cautious. The benefits of looking at a multi-dimensional or multi-variate type of outcome certainly does give us chance of capturing a broader spectrum of the nature of what the effects are.

After we discuss this broad issue, in fact, I had two or three other specific issues that I was hoping to discuss that really relate to this, to two of the domains of what I would think of as what might be the dimensions of this surrogate.

DR. DAUM: Go ahead.

DR. FLEMING: Okay. Well, one of them is we've -- my understanding is we're going to be focusing predominately on vaccines that would target the HPV 16 and 18 types. Certainly as we look at outcome marker surrogates, the ones that will be most sensitive to the effects of these vaccines will also be type specific.

At least as I'm thinking through my own formulation of what might be a surrogate or an accelerated approval measure versus what might be a full approval measure, it would be the distinction in accelerated approval of allowing focus more on those type specific outcomes but full approval on more validation of a global benefit.

My sense of how important that distinction is I have uncertainties. My understanding from the data that was presented yesterday is something on the order of 60 to 70 percent of CIN2/3 as associated with HPV 16/18.

Before you comment on that, you can confirm or refute that, but my more important question is what is the nature of the -- how much of cervical cancer is attributable to 16/18? Will there be an opportunistic influence here? If you essentially reduce to eliminate 16 to 18, what influence does that have? Could we expect that will have on the global rate of cervical cancer?

So there's two elements to this. The one element is in the current milieu of the mixture of these types, what fraction of cervical cancer is attributable to 16/18 and if you eliminate that causal influence, are there opportunistic influences that would alter what the ultimate reduction in the rate of cervical cancer would be?

DR. GOLDENTHAL: Well, across studies I would say that about 60 percent of cervical cancers overall globally are due to 16 and 18. That's a rough approximation.

I would suspect that in the U.S., as I mentioned yesterday, the adenocarcinoma components is becoming of increasing importance so that the 18 component, in my mind, has a lot of importance here.

In terms of CIN 2/3 I think somewhere in the range of 50 to 70 percent of CIN 2/3 may be attributed to type 16 and 18.

DR. FLEMING: So you are confirming the approximate CIN 2/3 numbers that I gave, around 60 to 70 percent. You're suggesting that under the current milieu that there is a corresponding comparable percentage of cervical cancer that can then be attributable to 16/18.

The third aspect of it was is there any sense if you eliminate that component, can we conclude that we'll be left with then 40 percent, or could there be an opportunistic aspect here such that the actual reduction in the rate of cervical cancer may be less than that?

DR. GOLDENTHAL: I think what you're asking about in part is replacement. In other words, if you eliminated some types would you have replacement with others. I've actually looked in the literature for that very -- to address that very question.

I didn't see evidence from the literature that removing, let's say, type 16 would be more likely to cause persistence of other types. Again, none of this is in the context of a vaccine trial so that has to be kept in mind also.

DR. SNIDER: Thank you. Karen, on that particular point --

DR. DAUM: Dr. Snider, is this on this very point?

DR. SNIDER: Yes.

DR. DAUM: All right. Let's finish this point. Go ahead. People are waiting in line so please go ahead and finish this point.

DR. SNIDER: I just wanted to point out that in the mathematical model yesterday this was taken into account and they assumed the reason it was taken into account is because women get infected with multiple genotypes of HPV so that just because you're infected with 16 or 18 and develop CIN 2 or 3 cervical cancer as a result of that doesn't mean you are immune to.

It means, in fact, you're not immune to some of the other oncogenic genotypes. There would be not so much a replacement but there would be cervical cancer in people who receive this vaccine as a result of their being infected with other oncogen types if I understood the model correctly. It's a small proportion but, I mean, it's there.

DR. GOLDENTHAL: I don't think I want to comment on that model.

DR. DAUM: Let's go on then. Finally, Dr. Reeves.

DR. SNIDER: I apologize.

DR. REEVES: I have two comments, one just a follow-up on this. I would agree with everything that Dr. Goldenthal said. I mean, I think there's not going to be a rush of other types to replace it. If it works, if there is an effect with 16 and 18 associated cancers, then that gives us more evidence to make better vaccines.

I would agree with two endpoints. Actually, I was unfortunately looking at the slide and I'm wondering if we shouldn't discuss three. To me, CIN means histologically confirmed disease and high-grade squamous intraepithelial lesions means PAP smear diagnosed disease.

I'm wondering since PAP smears are what, in fact, women screen in on, are a relatively easy procedure to do rather than following women all the time with colposcopy and biopsy to get a CIN diagnosis whether, in fact, studies should not have a large PAP smear component and include obviously with colposcopic follow-up but include reduction in high-grade squamous intraephithelial lesions as well as reduction in CIN 2/3 as well as potentially a decrease in infection or persistence of shedding.

DR. DAUM: Thank you very much, Dr. Reeves. My sense is that we've really had a fairly thorough sort of go around in terms of issues that bear on question 1, the issue of endpoint choice.

So what I would like to do is take a short break. Then upon return to begin systematically polling not as a vote but systematically hearing from each member of the panel in terms of the endpoint question. It's 10:25 here in the eastern time zone, or 10:20 according to this green clock. We'll take a 15-minute break and reassemble at 10:35.

(Whereupon, at 10:22 a.m. off the record until 10:39 a.m.)

DR. DAUM: Would everybody take their seats and get ready? Thank you very much. We're missing a few people and that's out of the table I must say. Do you know where everybody is?

PARTICIPANT: No, but I'll go find them.

DR. DAUM: Okay. We are now going to sample opinions, so to speak, on question 1. Before we do that, Dr. Fleming has a couple of succinct unspoken points to raise.

Dr. Fleming.

DR. FLEMING: Thanks, Bob. I'll just keep this to one theme, and that theme is we heard some brief discussion at the beginning of this morning about as we're struggling with defining which of these markers are really the appropriate ones to use as surrogate or replacement endpoints in accelerated approval or, for that matter, even full approval we've noted that some of these markers, certainly CIN 2/3, maybe even persistent infection, influence how interventions or care is delivered.

There has been at least some uncertainty about how that then impacts our view of the appropriateness of those markers as surrogates.

I guess the point I want to make is it's not uncommon in clinical practice in many disease settings for markers to be used and their use can be in several different ways.

Markers can be used as prognostic factors to guide patients and caregivers on risk of outcomes. They can be used as triggers for when and how to intervene. They can be used as surrogate endpoints which by definition we should mean as endpoints that serve as replacements for other ultimate more important clinical endpoints.

The point that I want to make here is that those are three distinct purposes and it may be that some markers are appropriate for some purposes and not others.

As a quick example, in the HIV world where we're looking at interventions to prevent transmission of HIV, it's clearly known that STDs are a prognostic factor indicating higher risk for transmission of HIV. But that doesn't mean that even though they are clearly prognostic markers that they are appropriate surrogate markers.

A couple simple examples of this, there were a couple of major trials done of STD inventions in developing countries to look at whether we could prevent transmission of HIV by preventing STDs. In the RICAH trial with a mass intervention, we were successful in reducing STDs but we had no impact on HIV.

Conversely in the MELANZA trial with syndromic interventions we had no impact on a number of STDs but we reduced HIV. You can readily have a prognostic factor. Because it's a prognostic factor, that doesn't mean that it's specifically a replacement endpoint or surrogate endpoint.

It can also trigger an intervention. Classic example, in cardiovascular diseases we know that arrhythmias are risk factor for sudden death. It's clearly a prognostic factor. For that reason, it triggered many people to then use anti-arrhythmic interventions, echinide and flecunide, for example, to reduce arrhythmias which they do do with the intention of reducing sudden death.

Two hundred to 500,000 Americans a year were using them on this premise. Ultimately a placebo controlled trial was done that showed that they actually did reduce arrhythmias but they tripled the death rate.

A marker that is clearly prognostic that may trigger a physicians use to intervene doesn't necessarily mean it's a reliable replacement measure to ultimately judge the effect of the intervention on the clinical endpoint.

The final example that I might give is early HIV infection. We can treat early HIV infection using HIV levels as a guide for how to tailor the intervention and that may well be an appropriate strategy.

If you want to mix the types of anti-virals we're using to achieve undetectable levels for an early infected HIV person, but that doesn't at all mean that reducing viral lows to undetectable levels in a certain manner is a clear surrogate endpoint for achieving prevention of long-term transmission of HIV, long-term occurrence of systematic disease and death.

Ultimately what is important is that when we consider markers in a case like this, which would be persistent infection, for example, or CIN 2/3 to distinguish the fact that they are clearly prognostic. We know that they are prognostic. They may be used to trigger intervention. Certainly CIN 2/3 is. But whether that makes them -- it doesn't at all address whether the question that we're really interested in, which is whether they are appropriate replacement endpoints.

Although I will say -- certainly I will acknowledge that if CIN 2/3 is a trigger for an invasive surgical intervention, then a vaccine that would prevent the need for that invasive surgical intervention, that is a direct intrinsic value, but that doesn't also lead to the additional conclusion that we're doing anything specific relative to preventing cervical cancer.

DR. DAUM: Thank you very much. I think that is a very clarifying and helpful perspective.

Dr. Snider, you wanted to speak to this very issue?

DR. SNIDER: Actually, a very quick point that is a little different, and that is that it was mentioned to me at the break and sort of shamed me as an epidemiologist that I hadn't brought this up earlier. An FDA staff member by the name of Dr. Ellen Birch pointed out to me that in our discussion of concerns about eradicating HPV 16 and 18 infections in individuals.

We didn't think about the secondary effects of reducing the prevalence of those infections in the populations and, therefore, even if there were certain individuals who were not protected by the vaccine and got cervical cancer, if we were able to reduce HPV prevalence in the population by 80 percent or so, that other people who were susceptible to cervical cancer from this infection may not even acquire it because the prevalence in the population had been reduced. I just felt that it was an important point that needed to be brought out in this consideration.

DR. DAUM: In other words, an effect in the transmission perhaps.

DR. SNIDER: Yes.

DR. DAUM: You're absolutely right. It hasn't been said and it sort of goes in the thinking of how vaccines work when deplored over a whole population.

DR. SNIDER: And Sam points out the magnitude of that would be greater if you gave it both to males and females. Still I think even if you gave it to females it would have some effect.

DR. DAUM: No, I think that it's good that it's been said. I'm sure it's been on many of our thoughts as we go through.

I would like to sort thicken the soup a little bit with raising one more issue. That is, this issue of accelerated approval. What I think we can do is have Karen Goldenthal remind us exactly what the agency means by that which she has agreed to do during the break.

Then I think we can try going around and getting everyone to speak to these issues to incorporate this idea into your comments. I had thought initially we would go around twice but I don't think we need to. If that view needs to be reassessed, then I'm happy to reassess it.

I think that given your choice of endpoints, that you can also say how you would phase it in, how you would advise the agency to phase it in to their strategy for approving these vaccines.

In order to prepare us for this discussion, I'm going to call on Karen first to remind us in very precise succinct language, which she has agreed to do, what exactly is meant by accelerated approval and how it might phase into your choice of endpoints or multiple endpoint scenario.

DR. GOLDENTHAL: Thank you. I have a couple of points to make here. Accelerated approval is basically the use of a surrogate marker that's reasonably likely to predict clinical benefit as the basis for an approval, but that's not the end of it. You have to have a confirmatory trial that would need to be well controlled and well underway at the time of approval. I would even think at the time of a BOA or license application submission for the accelerated approval endpoint.

Just a few things to keep in mind in that regard. This means that FDA would be asked to do an approval based on interim data with the accelerated approval application. That particular interim data would need to be presented to an advisory committee. You need to be thinking of how you would feel being an advisory committee member and having that particular accelerated approval endpoint to base your decision on.

Another critical thing pertains again to the timing of the confirmatory trial, particularly with regard to its completion. I think you need to think very carefully about whether randomized trial could realistically continue following an accelerated approval.

My suspicion is that at least in the U.S. that would be problematic. When I thought about applying accelerated approval, as I mentioned yesterday, I thought about the fact that it cuts about -- it would potentially make the vaccine available maybe a year earlier than it would be otherwise thinking of the FDA review and approval process.

DR. DAUM: Thank you, Karen.

Dr. Kohl, you have been placed at jeopardy by Dr. Stephens' departure. I'm sorry about that.

What I would like to do is to ask each person now, and we'll go around, to comment succinctly on the two questions, that's 1 and 2. I would like you to see if you could incorporate into your comments an issue that the agency has raised and asked for your comment, and that is the indication.

In other words -- I guess in other words, and Karen Goldenthal, correct me if I don't understand it, if there were an accelerated approval scenario where something were approved based on an interim indication, what would the approval indication say? I need you to sort of put that into your comments as well.

Dr. Kohl, let's start with you and we'll just get a feel for how this goes. We have a little over an hour to do this and I think we can get it done.

Not quite yet. Clarifying comment.

DR. GOLDENTHAL: Also the indication. Our question about what should the indication be also would apply to traditional approval.

DR. DAUM: Thank you, Karen. One more bit of food to swallow.

DR. PALESE: And this is not a vote, correct?

DR. DAUM: This is not a vote. This is your comment. They will be noted, recorded, and thought about, I can assure you, line by line.

Dr. Kohl.

DR. KOHL: We are being asked to consider endpoints for a vaccine that hopefully will provide long-term, possibly lifelong protection against cancer. The things that give me pause in terms of an early surrogate, and I'm not sure what is distal and what is proximal but in terms of a virological surrogate is we have no idea what the duration or protection is yet for any of these type vaccines. There's a significant question about population heterogeneity and detection in different types of populations which I think needs to be addressed, or looked at, at least.

We don't have a clear definition of what persistence of viral infection is yet from the experts, although that may evolve in the next year or two possibly. The considerations for size obviously have to include what sample size and how long a duration would be necessary for safety as well as some kind of efficacy in terms of markers.

The last point that Karen brought up, early licensure, I think, will seriously preclude subsequent studies of this vaccine, the hypothetical vaccine, and, in fact, future vaccines for HPV prevention.

Bearing those issues in mind, what I would call for in terms of primary efficacy is a CIN 2/3 model. I would urge that this study be powered such that CIN 2/3 could be clearly defined in terms of efficacy, but it would include sequential virology yearly or every six months.

I'm not sure what is appropriate and what the best technique will be at the time that this study is undertaken. Right now it looks like it's PCR. I would include definition of all oncogenic HPVs, not just the ones that are in the vaccine so we can look at replacement.

And also would include immunological parameters that would allow us to determine what the correlates of protection are against both infection, persistent infection, and CIN 2/3.

Given that as question No. 1, then I come to is there something acceptable for me for accelerated licensure.

If this study were to proceed as I envision it, then for provisional licensure or accelerated licensure I like something that Dr. Fleming suggested yesterday; namely, that my primary endpoint is going to be CIN 2/3 efficacy but accelerated could be a proof that there is a significant difference in CIN 2/3 between a placebo and a control.

That is, as soon an interim analysis shows a significant difference, accelerated approval might be asked for, but in that study will obviously come efficacy.

DR. DAUM: Thank you very much, Dr. Kohl. We're off and running. What you did that was really wonderful is you actually managed to address all the things I asked for and the agency asked for. If everybody could sort of make a little checklist in their minds as we go around to try and touch each of those points, I think we'll have a wonderful discussion.

DR. KOHL: The only thing I didn't address, I guess, would be what the package insert would say about what this prevented. I think it would say prevention against CIN 2/3 with probable effect on cervical cancer but not proven.

DR. DAUM: Thank you.

Dr. Griffin.

DR. GRIFFIN: Okay. With respect to question 1, I guess my choices are A, B, and E. I think that I'm of the opinion that if you prevent A, incident infection, you will, therefore, by definition prevent persistent infection.

Now, whether you need then to -- and the main objection to saying preventing incident infection is what a criterion is for the efficacy of the vaccine and that may be much too stringent. As many people have brought up, you may get infection that is rapidly cleared and, therefore, preventing persistent infection would be a more realistic surrogate.

I think that data will just have to evolve so if you required prevention of persistent infection, you would accomplish that if you were also preventing incident infection. Therefore, I guess B would be the main virologic endpoint.

I guess I am most convinced by the one-year endpoint for persistence, definition of persistence, but, at the same time, realizing that this is a bell-shaped curve, about when the actual oncogenic activities for a virus infection would actually kick in for any individual person you can't predict.

In some people that's going to happen early and in some people that's going to happen late. In some people that's not going to happen at all. There isn't going to be any way to predict for an individual.

Then, lastly, I think that the outcome that's most closely related to development of cervical cancer is the CIN 2/3 pathologic endpoint. What I would like to see in a trial is imbedded both outcomes, that you have two parameters.

Don't ask me how you would design this but I'm convinced that it's happening with other kinds of interventions with respect to HIV, etc., where you have early endpoints that then allow early accelerated approval, etc. But, at the same time, the same trial has enough individuals in it that you follow them for a longer period of time.

That's ongoing at the time that you're getting your early outcome data and you avoid the problem of then having to have a new trial with a vaccine that you've now got approval for and one would say you ought to be using. I would think that I would much favor a larger trial to start out with that you look at both of these, basically virologic and pathologic endpoint.

Embedded in that is then the fact that I can see accelerated approval using a virologic endpoint, i.e., persistent infection with the HPV types that are in the vaccine. Then it's a little more problematic to say what you are preventing.

You're not going to be able to say in the package insert at that point that you are preventing cervical carcinoma or even that you are preventing CIN 2/3 if you have that data yet. You may or may not be able to say you are preventing infection depending on what the data show.

If you actually have prevented infection, you can say that with this oncogenic types. Otherwise, I guess you're stuck with saying you are preventing persistent infection if that's your outcome.

DR. DAUM: Diane, thank you very much.

Dixie, you're up.

DR. SNIDER: Thank you. First of all, I would just like to congratulate everybody who's been involved in all this work. I mean, it's really exciting to be sitting around the table talking about a vaccine that may prevent a cancer that globally and even in the United States is of great significance. I would express my appreciation to everybody in the academic community, NIH, FDA, pharmaceutical companies and so forth for getting us to this point.

As I expressed in my frustration yesterday, it is a moving target and it does create a difficult situation in terms of making definitive recommendations but with the clarification that Karen gave us. I think it is possible for us to address these questions, at least as we view them today.

My recommendations are, I think, very similar to those who preceded me in that, at least at this point, I would be interested in designing one trial that would look at two endpoints, persistent infection and CIN 2/3. I'm assuming, of course, if you're looking at persistent infection, you're going to be looking at incident infection but that wouldn't be a primary endpoint.

I do have some concerns about persistent infection that others have already talked about as have I. How is it going to be defined not only in terms of the issues brought up there as it relates to the appropriate number of tests in the interval between tests, but the sampling methods and the assay methods and all of that needs to be carefully worked out to be sure that it's very highly sensitive in detecting the presence of infection.

Then there's the whole issue of whatever you want to call it, latent infection or an apparent infection using techniques that aren't highly rigorous that concern me. Those are issues that have to be dealt with.

With regard to the labeling, I guess I would lean toward what Dr. Kohl, I think, has already mentioned, that the label would say if this endpoint was reached that the vaccine prevents CIN 2/3 which is associated in a high proportion with the development of cervical carcinoma.

I personally would not be inclined to support an accelerated approval approach right now. However, right now that is two important words because there is an evolving scientific database that might change my opinion and obviously the opinion of many other people if we got information.

We were able to get some information that, for example, identified subgroups of women who clearly had an extraordinarily high risk of cervical cancer or progression to CIN 2 and 3 with persistent infection. It's conceivable that somewhere along the way during this trial that an alternative could be revisited. But, at this point in time, I think there are enough uncertainties about the significance of persistent infection and how you define it that I would be a little reticent to advise FDA and the manufacturer to proceed along those lines and bring those data in. At least with the database we have right now, I think it might not lead to a happy outcome.

But if we were able to move to the point where we became convinced that persistent infections or, at least, persistent infections in certain identified populations, whether that's personal characteristics, viral loads, who knows what, really progressed to cancer, then the labeling would be of this sort that the vaccine prevents persistent infection with these particular types which is associated in some individuals, or maybe at that point in time it could be a high proportion of individuals, with progression to CIN 2 and 3 and cervical cancer.

DR. DAUM: Dixie, thank you very much.

Dr. Kim.

DR. KIM: I also support the concept that the trial can be designed large enough to address perhaps a minimum of two endpoints.

I guess this is in part that as we heard that there are many issues that are not only heterogenous but also answers are not in our hand at this time so that I think it is important to be able to monitor all the issues which have been addressed during this meeting as part of perhaps a trial so that, again, going back to the specific questions would be my preference would be looking to a persistent HPV infection as a primary endpoint since the HPV infection per se can be difficult to predict whether you regress or you persist.

I would at least like to see that a vaccine has been shown to be beneficial in preventing persistent HPV infection due to vaccine types.

Then I guess the question which we do not have based on the discussion is whether that can be translated into the bottom line which is reduction in cervical cancer. I think it's because of that I would certainly like to see some data related to those issues as the study is coming along.

Particularly I would like to see the information on CIN 2 and 3. Again, I think cervical cancer would be very, very difficult to achieve as an endpoint so CIN 2 and 3 as a secondary endpoint as part of a trial.

So what that means, at least to me, is that when this vaccine can go through and then would be presented as an accelerated format, then I would like to see that vaccine has shown to be beneficial in significant reduction of persistent HPV infection due to serotypes.

Also, I would like to see that time of discussion, some information on a significant reduction on CIN 2 and 3 as a sort of assurance that, indeed, prevention of persistent infection has, indeed, a sort of right kind of target that we all want to see as part of this vaccine.

DR. DAUM: Thank you, Dr. Kim.

Dr. Katz.

DR. KATZ: I don't think I have any disagreement with what my preceding colleagues have stated. To me the most important issues or endpoints would be persistent infection and the CIN 2/3.

But I have several questions which perhaps are tangential but I would like to see some nested studies within the large trial and some nested studies that in a smaller cohort might be able to answer some of the questions we've tossed about to which we don't have answer about, the role of mucosal or secretory antibody, the role of salmeated infection and what could be done in looking at that along with virus cultures.

The other issue that concerns me is the way we've conducted conventional vaccine studies, and I would call this one somewhat unconventional, is once you've reached a point where you're comfortable that you've achieved your goals, the controls then receive the vaccine that the original recipients have the benefit of.

I don't know when you would feel you've reached that point. If we accept the endpoints of persistent infection and CIN 2/3, then maybe that's the time that you would give the controls the vaccine. But that wouldn't answer what Dr. Kim wants which is the next step which is cervical cancer. I think I would have to consider that in my overall format as I have put together the longitudinal protocol.

DR. DAUM: Thank you very much, Dr. Katz.

Dr. Faggett.

DR. FAGGETT: I disconnected my phone.

DR. DAUM: I'm very grateful.

DR. FAGGETT: I really learned a lot these past couple days. Just sitting next to Dr. Katz is always an hallucinating experience.

Really, I think more of us primary care providers need to hear this kind of very high-level discussion of the science of the vaccine approval process. I think it would make us better able to discuss it with our patients and encourage them to get the immunizations available.

I think question No. 1, I agree with previous speakers. Again, it would appear that persistent infection in CIN 2/3 consensus endpoints, the accelerated approval I think we should do it. I think we should have as long as possible.

It would appear that the longer we go, the more evidence we'll have in terms of the efficacy in preventing cancer. I would just say I concur with previous speakers with those comments.

DR. DAUM: Thank you. I need to press you a little bit, though. The accelerated approval, where do you sit on that?

DR. FAGGETT: Yes. I think we should have accelerated approval.

DR. DAUM: The endpoint you pick would be? I just want to make sure I'm very clear on what you're saying.

DR. FAGGETT: Again, as previous speakers, that you would prevent the disease. I think the longer you go the more evidence you will have that you can prevent cancer so I would say it would be probable prevention of cancer would be an endpoint.

DR. DAUM: Which endpoint would you pick? Did I miss it? Did you say it? If you did, I apologize. Persistent infection?

DR. FAGGETT: Right. Persistent infection.

DR. DAUM: Okay.

DR. FAGGETT: I said persistent and CIN 2/3.

DR. DAUM: So you picked two different ones.

DR. FAGGETT: Yeah, those two.

DR. DAUM: I apologize. You want two different endpoints?

DR. FAGGETT: Right.

DR. DAUM: I think we know what you want now.

Ms. Fisher?

MS. FISHER: I thought of this in two separately so I'm making two separate statements.

DR. DAUM: Stop for one second. You said something that upset Dr. Mitthune.

DR. MITTHUNE: Just to clarify, Dr. Faggett, would you want both the virology and the CIN 2/3 for accelerated approval or would you want only the virology on regular approval? You want both? Thank you.

DR. DAUM: Thank you, Dr. Mitthune. Thank you, Dr. Faggett. Thank you, Dr. Katz.

Now, Ms. Fisher.

MS. FISHER: Well, first I'm going to speak about the endpoints. From the information that was presented to us in both closed and open sessions of this meeting, it appears that if an HPV vaccine demonstrated prevention of persistent HPV infection with certain types such as HPV 16 and 18, it would suggest that it would be effective in preventing cervical cancer associated with those types.

However, much appears to be unknown about potential cofactors involved and why some women clear HPV infection and some do not and go on to develop cervical cancer. I think there needs to be more known about these potential cofactors because they may be important independent of HPV infection.

In prelicensure clinical trials demonstrating efficacy, the standard used should include follow-up of all participants to prove not only prevention of persistent HPV infection, but also prevention of CIN 2/3 as well as prevention of cervical cancer because CIN 2/3 is a more certain predictor that cancer will most likely occur and demonstration of prevention of cervical cancer is the only way the vaccine user could be reasonably confident that it is, indeed, a vaccine that could prevent cervical cancer.

The other statement is on the accelerated approval process. I think, needless to say, certainly cervical cancer is a terrible disease for women, especially in developing countries and we need safe and effective ways to prevent it.

Because the majority of women clear HPV infection and a very small number go on to develop persistent infection, and an even smaller number go on to develop cervical cancer, I'm concerned about an accelerated approval process for licensure.

If the request for accelerated approval was for an HPV vaccine that would only be used by women known to be at very high risk for developing cervical cancer, then I might feel differently.

However, this discussion has been about an HPV vaccine that would target all healthy adolescent girls and adult women, perhaps even female and male children. That's entirely another matter. We need to have a better understanding of the biological mechanisms of long-term immunity of HPV infection.

We need more information about safety including the potential ability of this protein vaccine to induce autoimmunity in a subset of genetically susceptible individuals, as well as the potential negative impact on women with preexisting HPV infection.

Clearly it should not be an a priori assumption that this vaccine has no long-term negative health consequences whatsoever. Long-term studies need to be done to measure for all morbidity and mortality outcomes.

I'm not talking about paying attention to car crashes and ski accidents that occur during the study but taking serious development of post-vaccination deterioration of health such as multiple sclerosis-like symptoms, arthralgia, arthritis, thyroid disease, etc., as well as exacerbation of preexisting autoimmune conditions during long-term follow-up.

If we don't ask for these kinds of studies prelicensure, an unknown number of young women who may indeed avoid infection with HPV and cervical cancer by using an HPV vaccine could be left with other vaccine induced chronic health problems because the vaccine was licensed too quickly without enough data. I do not think the accelerated approval process is appropriate for this vaccine.

DR. DAUM: Thank you, Dr. Fisher.

Dr. Palese.

DR. PALESE: This is obviously a very complex issue here. Human papilloma virus we don't have a good system, no good animal model, and certainly no antivirals and no vaccines. On the other hand, cervical cancer appears to be almost 100 percent associated with infection by HPV.

Now, if we have a vaccine which basically prevents infection and we can't demonstrate virus, I'm sort of persuaded by persistent HPV as an endpoint and I would go along with Dr. Lowy's recommendation of a year. I guess he meant two assays. He didn't give a specific amount because of an interval but I think six months may be okay.

Clearly as a virologist I feel if there is no virus, then there is no disease so this is really for me very, very compelling that one would be able to prevent infection and replication of the virus that this must have some consequences and that's why I feel very comfortable with an endpoint which measures persistent HPV infections.

And having that rationale, I sort of also feel that an accelerated approach would be -- I would support that. Accelerated approval I would support, particularly if there is a provision for a long-term analysis in there and that the time would be large enough in terms of measuring other parameters. Again, I would be happy enough if it turns out that there is no virus replication that we would vote for an accelerated approach.

In terms of the labeling I would also say if the vaccine prevents infection, then it is also most likely prevents cervical cancer so I would be quite happy with that kind of labeling.

DR. DAUM: Dr. Goldenthal, if I understood you, the accelerated approval scenario would be one that would be granted by the agency only if there were a confirmatory trial in progress or enrollment was completed. Is that correct?

DR. GOLDENTHAL: That's the way accelerated approval ordinarily works, yes.

DR. DAUM: So, Dr. Palese, let me come back to you for just one moment. You mentioned that you would have accelerated approval based on viral persistence, if I understood you.

DR. PALESE: Yes.

DR. DAUM: And if that's the case, then you would accept that caveat that the confirmatory trial be in progress but I didn't hear you say that.

DR. PALESE: Yes, with the same endpoint. I mean, I'm not -- maybe I didn't understand your question.

DR. DAUM: Okay. Agency people listen and if I'm not saying it right, please jump in. It seems to me that you might ask for traditional approval with persistent viral infection as your endpoint.

DR. PALESE: No, that's not what I am --

DR. DAUM: Right. And then accelerated approval though would have to have an interim endpoint that approval would be granted for but a confirmatory trial in progress or underway as well for the

agency --

DR. PALESE: What kind of endpoints? I mean, that's the question. For this confirmatory trial that's --

DR. DAUM: That's what we're asking you to comment on.

DR. PALESE: Okay. I will be happy with persistent -- if there's no virus, there's no disease so I will be happy with the confirmatory trial with the same endpoint of persistent HPV.

DR. DAUM: Does that fit with agency guidelines or are we okay with that?

DR. GOLDENTHAL: It almost sounded like he's more advocating traditional approval.

DR. DAUM: I think so, yeah. With persistent viral infection as the endpoint. I think that's what he's saying.

DR. PALESE: So what am I saying?

DR. DAUM: I'll be damned if I know.

DR. PALESE: I will keep going for an accelerated approval. If that requires a confirmatory trial going on, I would support that but with the assumption that the endpoint again would be persistent infection.

DR. DAUM: Okay. I understand what you're saying and it doesn't completely gel for me but that's okay. And the indication would be what?

DR. PALESE: That a vaccine, if it turns out that it really prevents infection, most likely prevents infection to a certain percentage and, therefore, is most likely to prevent also cervical cancer.

DR. DAUM: Very good. Thank you.

Dr. Myers.

DR. MYERS: The endpoint of interest is cervical cancer and I think the data on CIN 2 and 3 as a part of the natural history is sufficiently robust that it predicts a clinical benefit directly I think is clear and it probably serves as a surrogate for cervical cancer.

While I agree with a lot of the preceding comments, it's intuitive that prevention of infection, and specifically prevention of persistent infection even without dislogic changes, it's intuitive that those could be endpoints. I don't think the data at this time are sufficiently robust. Like somebody said previously, at this time that is a qualifier.

I think some of the data we heard in closed session yesterday may imply that a year from now or so we may be able to say that persistent infection, in fact, in the absence of histology could be a marker but it's not at this point. I would suggest that infection are secondary endpoints and not the primary endpoint.

I would want data on the other high-risk HPVs as well looking for emergence of those. But also because I think for the next generation of vaccines that will be very important.

Going back to the secondary endpoint, the infection endpoints, I think, are also critical to collect that now as part of the study so that the next generation of vaccines we will, in fact, know whether we can utilize these as surrogate markers for the histology.

I mentioned this before a couple times. I would just like to say it again. I think it is important to understand that it will be important to examine the outcomes on an intent to humanize perspective.

I think to look just at HPV 16 and 18, naive individuals, would be a mistake and that we need to understand what immunization of previously infected young women is as to whether that reduces the risk of persistent infection or has an adverse outcome because this vaccine will not be directed just at naive individuals. It will be targeted to specifically young women who are at high risk and, therefore, may already be infected.

As to accelerated approval, I'm unable to support that conceptually in that I think it would be very difficult to complete a study even if enrollment is completed. Once the vaccine is approved and is being marketed, I think it would be very difficult for the placebo arm to be maintained. Therefore, as I think the definitive endpoint is CIN 2/3, then I think it would be very difficult to support an accelerated approval.

With that said, I think if, in fact, early on in the process there were a significant difference between the groups for CIN 2 and 3 before the full duration of the study is completed, then I would consider accelerated approval at that point.

From the labeling perspective I thought Steve Kohl and Dixie said it quite well and I would agree with that.

DR. DAUM: Thank you very much, Marty.

Dr. McInnes.

DR. McINNES: My certainty and uncertainty about papilloma viral infections and their relationship to cancer and the role of this vaccine waxed and waned. I think I'm left here with a fair amount of certainty that we are reasonably uncertain about lots of things here.

I'm moving forward on the assumptions that human papilloma virus infection is necessary for and does precede cervical cancer, although it's not sufficiently causal. I do understand that HPV infection with the oncogenic type is much more common than the resulting cancers.

Nevertheless, I'm also reasonably comfortable with the assumption that persistent infection is linked to risk of CIN 2, CIN 3, and invasive cancer.

With the reality of having to accept a surrogate, I am comfortable with persistent infection as an endpoint, surrogate endpoint. The timing of that, I am somewhat concerned about the short interval that has been proposed.

Given the data that incident infections may clear within eight months, I am bothered by time frames that are less than that. I think I envision protracted trials rather than condensed trials.

I am somewhat persuaded that cytologic abnormalities are an endpoint for consideration in the trials because they certainly would give us a sense of the bad player HPV infections with more rapid progression to the CIN 2 and CIN 3.

I do not dismiss the role of cytological evaluation. Certainly it's a question of where it would be within the framework of endpoints, primary, secondary, tertiary endpoints.

The case definitions, I think I'm not totally resolved on which of the spectrum of clinical disease has a place and which doesn't. At this point I'm not persuaded that any of the spectrum doesn't have a potential place in articulation of an endpoint. I would leave open the possibility of a spectrum of clinical disease being incorporated with persistent viral infection into the endpoint.

Regarding the accelerated approval, I at this time am having a great deal of pragmatic difficulty understanding how sufficient safety data, how a considerable safety database will be brought to bear for consideration for the accelerated approval.

Pragmatically when I lay out a time frame I don't at this point see much to be gained. I'm obviously open to being persuaded of something other than that. At this point I am advocating a traditional approval and I'm having difficulty understanding the role of accelerated approval for this vaccine.

DR. DAUM: Thank you very much. Quite clear.

Dr. Reeves.

DR. REEVES: Okay. To begin, I would be in favor of accelerated licensure because of the nature of the disease and the long time period and actually seeing the disease of interest. The disease of interest is prevention of cervical cancer.

For that reason I believe studies should be done in high-risk populations, the woman that actually get cervical cancer in the United States, and to the extent possible so that one can begin early on. They should involve populations that have cancer registry so that a long-term effect can be seen.

I think the only appropriate surrogate endpoint, and perhaps an endpoint in and of itself is CIN 2/3. In the United States in terms of body count that is actually the primary cost, the primary morbidity. If treatments for that could be cut down significantly, it would be a significant public health advance so I think that is a very appropriate endpoint -- surrogate endpoint or endpoint.

I believe that the virology and immunology are also terribly important to studies and must be included as either co-surrogate endpoints or data that must be measured. I really don't have an opinion on what persistent infection is.

I think that viral studies must be very complete. The patients, or the subjects, should be followed by cytology as well. Every time that a cytologic sample is taken, a virologic sample taken also. Presumably high-grade SIL will go down but the patients with low-grade SIL, their virology is the important comparison along with the placebos.

Cervical immunology is terribly important to this and I believe should be included in any of the studies. I think the question of incident HPV infection is probably going to be an impossible one to address.

I think the term is used wrong. The first culture positive is not an incident disease. It's the first incident infection. It's the first infection in a person that has not been infected with the agent previously.

I think it's going to be impossible to cover in the studies but it's been brought up a couple times. One of the primary epidemiologic risk factors is age at first intercourse. That group of women in the high-risk group is going to be part of this but only a small part of it.

As far as the package insert, I believe it's premature to be discussing that.

DR. DAUM: Thank you very much, Dr. Reeves.

Dr. Goldberg.

DR. GOLDBERG: Thank you. I think that the accelerated approval --

DR. DAUM: Sorry. We may have a procedural problem.

DR. MITTHUNE: I would just like to ask for a clarification, Dr. Reeves. You said that you thought that CIN 2/3 would be your basis for accelerated approval?

DR. REEVES: That's correct.

DR. MITTHUNE: What would be your confirmatory study endpoint?

DR. REEVES: My confirmatory study? I think CIN 2/3 in and of itself would be sufficient. I think that's an important enough public health problem that if that could be dramatically reduced, I would be quite happy. I think cervical cancer is going to take decades which is, in fact, the final end product.

DR. MITTHUNE: Right. So are you actually advocating a traditional approval based on CIN 2/3 as your endpoint?

DR. REEVES: That's correct, with obviously evaluation of the virologic and immunologic data that is collected along with it.

DR. MITTHUNE: Thank you. One further clarification. Dr. McInnes, you said that you did not support and you advocated a traditional approval. It wasn't clear to me what endpoint would support that traditional approval.

DR. McINNES: I would use vaccine type DNA persistence so viral persistence. I talked about some spectrum of clinical presentation ranging from vaccine type DNA positive, cytological abnormalities to some clinical endpoint. I'm not opposed to the CIN 2/3. I'm just considering that it may need to be broader.

DR. MITTHUNE: Thank you.

DR. DAUM: Thank you.

Before we go to Dr. Goldberg, Dr. Goldenthal, I think there's a little bit of confusion still about accelerated approval. I would like you to just say the sentences you said before so that the remaining committee members can have it straight. What does accelerated approval mean?

DR. GOLDENTHAL: Okay. Accelerated approval means that you would have, I guess, a drug development plan in place where a product would be initially approved, receive the accelerated approval based on a surrogate and, at the same time, there would be a confirmatory efficacy trial that was also well controlled and well under way at the time of license application submission.

This would mean, again, a committee and FDA would have to review the interim data. That interim data could be used for the accelerated approval. Then we would be very interested in the timing, of course, of the confirmatory trial and when it would be completed in comparison to when the license application was submitted. You can have -- you know, we've heard various scenarios.

One thing that was mentioned was, I guess, sort of an early look at CIN 2/3 and then those people might be followed for another year. You might get more follow-up data on other participants in the trial. That was one example of the sort of accelerated approval. In that case, it was the same endpoint.

Usually you think of a different -- when I've seen it used in other context, it's been usually two endpoints, one for accelerated approval and one for the confirmatory trial endpoint which is something different. Perhaps we can work in the CIN 2/3 for both.

DR. GEBER: I just wanted to add that maybe a way of thinking of it that the accelerated approval is in a way a preliminary approval and if that were granted, the sponsor would have to then meet their endpoint in the confirmatory trial to keep their approval. If one decided that a persistent infection was an endpoint for a confirmatory trial, or if one were not satisfied for an endpoint for a traditional approval, then that would be a preliminary approval.

DR. DAUM: Question about this?

DR. FELIX: Procedure.

DR. DAUM: Please go ahead.

DR. FELIX: If in the confirmatory trial does it have to be a failure to achieve significance or would it -- I'm sorry. Would it have to achieve significance or failure to achieve significance? Would that belie the preliminary approval gain at accelerated?

DR. GOLDENTHAL: Certainly we can withdraw approval if the confirmatory trial is a failure, so to speak. In other words, if they don't find a significant result, than we can withdraw the approval. We do have that authority.

DR. DAUM: Dixie and Steve, we need to hear from the other half of the table.

DR. SNIDER: I think they need to understand the question.

DR. DAUM: Right to this point. Go ahead but we're running behind. Go ahead, Dixie, and then Steve.

DR. SNIDER: My question is that, if I understand correctly then, under accelerated approval the product would be licensed and available and, therefore, the individuals who participated in the confirmatory trial would have to be informed about the availability of the vaccine and presumably the IRBs would require that is included in the consent form and the IRBs would have to approve such a trial.

DR. GOLDENTHAL: Right. That actually speaks to my major concern which is continuing the trial following approval. I don't believe that there is a major issue in continuing a trial during the FDA review of the BLA.

As I have said, I think that accelerated approval might buy you a year but it's still got to be -- that's a very good question. Would an IRB concur with, you know, an ongoing placebo-controlled trial following approval. In the U.S. that's pretty unlikely.

DR. HILDESHEIM: If I could provide some factual information.

DR. DAUM: Tell us who you are and your affiliation.

DR. HILDESHEIM: Allan Hildesheim with the National Cancer Institute. If we did any interim analysis to submit to the FDA, we would have to present it to our IRB data and safety monitoring board.

We've discussed this and it's clear that any trial that had early CIN 2/3 as an accelerated outcome with confirmatory long-term CIN 2/3 would not happen because we would vaccinate our placebo group at the instant that we saw any evidence of protection against CIN 2/3.

DR. DAUM: Other the other hand --

DR. HILDESHEIM: Possibly even persistent infection.

DR. DAUM: On the other hand, if the first basis for interim approval were some viral marker like viral persistence and there was an indication for viral persistence and the trial were ongoing to look at CIN 2/3 or cervical cancer, that trial wouldn't necessarily have to be aborted because the vaccine would be available for prevention of viral infection. IRB would certainly have to address that and it's hard to know how it would come out. It's not as clear as the example you gave where it's very clear an IRB shouldn't go along with it if they were willing to even.

DR. HILDESHEIM: You are correct, it's more murky. However, my sense from the discussions I've had with IRB and DSMB members for our trial is that if we showed something was protected against persistent infection for a reasonable amount of time, that we may not have to abort the trial but we would be required ethnically to inform all of the women.

If they wanted, they could withdraw from the trial. In effect, any follow-up data after that would be highly biased by who stayed and decided not to stay in the trial.

DR. DAUM: Perhaps. I think that's very helpful. I think we understand what accelerated approval means and I'm really anxious to hear from this side of the table so let's go on.

DR. GOLDBERG: I'm not sure I know anymore but I believe there can be a well-designed trial for vaccine efficacy based on the CIN 2/3 or worse including any cases of cervical cancer that would be included in that endpoint.

With that said, I do believe you have to monitor for persistent HPV and the length of the interval has to be studied. I guess from everything we've heard so far, probably a year is the interval.

What I would suggest as an interim analysis for CIN 2/3 efficacy requiring that persistent -- that the endpoint is supported at that interim analysis if there was a recommendation to stop the trial early with the persistent HPV also showing efficacy.

I believe that we should be studying this in high-risk populations as well. I think there should be some stratification and HPV positive at entry should be retained as a stratum file so that you will have some information on possibly higher risk women.

The length of the study is an issue. It could be a much longer study than was anticipated, but I don't believe that even if we did a vaccine efficacy trial with persistence as the endpoint that we would ever be able to complete a confirmatory trial.

I think we have to make the effort to expand both enrollment initially and lengthening the follow-up but having a carefully planned interim analysis or interim analyses based on the best available planning mechanism that you can put in place.

I also think that there needs to be a mechanism in place for the long-term follow-up for the occurrence of untoward events as well as cancer. You also during the trial should be monitoring for other types of CIN 2/3 and/or cervical cancer associated with other types of HPV than just 16 and 18 to be able to access the impact of this on that.

DR. GOLDENTHAL: Could you just clarify are you advocating CIN 2/3 for traditional approval?

DR. GOLDBERG: Traditional approval.

DR. GOLDENTHAL: Okay.

DR. GOLDBERG: I don't believe accelerated approval is really possible here. But I do believe that if the trial were well designed and the results compelling somewhere during that trial, there could be an early stopping but there still needs to be follow-up for safety.

DR. DAUM: Thank you, Dr. Goldberg.

Dr. Fleming.

DR. FLEMING: Thank you. Clearly the prevention of cervical cancer is a critically important public health problem and that leads to an obvious need for effective, responsible, and timely evaluation of various vaccines here for targeting the risk or for targeting the reduction and the risk of cervical cancer.

However, the use of surrogates has always raised complex issues. There is clearly a tradeoff when we're using surrogates between the timeliness and the reliability of conclusions. In this specific setting what we have in hand is strong evidence that HPV infection is a necessary factor.

But there is significant uncertainty from the information we have at this point regarding whether reduction in various virologic, cytologic, and histologic markers and what duration of effects on those markers translates into being reasonably likely to predict benefit which is the condition put forward before us by the FDA.

To me this leads us to the wisdom that Steve Kohl had pointed out yesterday, "When in doubt, be cautious." I think there is a lot of doubt in this setting. I think there is additional reasons to be cautious. Dr. Fisher pointed out how broadly this vaccine is going to be used. To my way of thinking, that does mean this is a setting where we have to be cautious.

I'm also concerned that the trials that lead to the initial approvals have a particular burden to be well designed. It's going to be extremely difficult in the future. You won't be looking at future vaccines addressed through placebo-controlled trials. Inferior trials will be even that much more problematic.

All of this leads me to being very cautious. My sense is from what we've heard the marker that has the strongest evidence for reliability, even though it, too, is not fully reliable, is CIN 2/3 and, in particular, CIN 3.

My sense here, as I think through the two stages, Bob, in this accelerated approval leading to full approval, ultimately the full approval from my perspective needs to have considerable evidence that we're influencing CIN 2/3 in two ways here. One is relative to the targeted types of HPV 16 and 18.

I think we need to be ruling out 33 or 50 percent reductions. We need to have sufficient evidence that we have something on the order of an 80 percent reduction that we can rule out 33 to 50 percent reductions. Relative to untargeted types, I would argue that we need to have evidence that over all we're seeing a reduction in CIN 2/3.

Ultimately the data that we would have in hand in the final approval, I think, has to address all of the dimensions of strength of evidence, breath of effect, and durability of effect.

Now, working backwards could we do an accelerated approval? I think this is very controversial. I think there is a potential here for doing an accelerated approval.

Just to give you a sense of what I'm thinking, the type of trial that I think can address what I would think we would need for an accelerated and for full approval is one that might only involve

-- I say only in contrast to what would be in some settings even bigger trials, 10,000 to 15,000 participants in a two-arm trial where they would take a year of accrual and about three years of follow-up to get what I'm getting at here for the accelerated approval target and an additional two to three years of follow-up for the full approval, essentially what could be the accelerated approval.

If we had significant evidence of a reduction in targeted HPV 16 to 18 type CIN 2/3, that would take on the order of 20 to 25 specific cases. The problem with that is what we may be seeing with that is simply the ability of the vaccine to reduce this risk of progression to CIN 2/3 in the rapid progressors which may not represent a more global effect.

For that reason I would strongly support that there would be a duel endpoint in the accelerated approval that would be based on persistent infection, specifically persistent HPV 16/18 infection.

I'm not comfortable at this moment, though, defining over what interval. I think there is a lot that is unknown, although the good news is in the course of finalizing the design and implementing these trials, there will be some additional time to tap into natural history data.

What I would specifically focus on here is getting additional data from prospective cohorts that allows us to follow up incident cases that are going to allow us to understand more clearly what will be the time of persistent infection as well as potentially viral loads. This could be multi-variate. What is the duration of persistent infection and level of viral burden that translates into fairly reliable evidence that when these markers are achieved, there is going to be a high level of risk of progression of CIN 2/3.

The reason I want this as part of the accelerated approval evidence is that if an accelerated approval is going to be based on 25 cases relating to CIN 2/3 which is all it takes to rule out quality when you have an 80 percent reduction, I want to have additional evidence to give me a sense that when I'm going to get more durable evidence later on about more global effects on CIN 2/3, that this is going to be achieved and that is where I think the added data on persistent infection as a co-marker for accelerated approval is complimentary in its insight.

What I haven't mentioned is there's a third dimension I would ask for to be explored in the accelerated approval. In addition to persistent infection where you are going to define much better than we can today exactly what that is after you follow these incident cohorts, I would like to see the CIN 2/3, HPV 16/18.

I would also like to see consideration of need for reduction in invasive therapies because that is, in fact, part of the tangible benefit that is going to be achieved here.

Where I would define invasive therapy, certainly what I'm thinking are of, for example, what typically would be clinical care when you have a CIN 2/3 infection, diagnostic decisional procedures, electro-loop excisions, for example.

Interventions which by their very nature have such clinical importance that it exceeds the cost inconvenience and toxicities and side effects of the very intervention, i.e., the vaccine you're going to be delivering to prevent these. They have to be significant events.

Now, with this as an accelerated approval, what we don't have, just to repeat what I said earlier, in my view is adequate insight about strength of evidence, breathe of effect, and duration of effect.

Ultimately this trial would need to continue and I'm guessing to approximately a six-year median follow-up point to be able to have sufficient evidence to rule out a 33 to 50 percent reduction, that we have even a better effect than that.

If we have an 80 percent true reduction, it's going to take 40 to 60 events to do this. If we have at least an 80 percent reduction on targeted HPV 16 to 18 based on Karen's projections, that's going to translate into about a 50 percent global reduction. At the same time this trial is going to be powered to rule out no reduction. My standard for untargeted CIN 2/3 is to at least conclude that you are achieving this roughly 50 percent reduction ruling out no reduction.

But for targeted 16/18 I want to see an 80 percent reduction ruling out 33 to 50 percent combined with the additional evidence that invasive interventions are also being reduced.

Final comment. Is this doable? My sense is this is a strategy of about a 10,000 to 15,000 person trial. We're going to be about four years into this trial when we would have this information that I outlined for the accelerated approval. We're still about three years away from having the final data.

There will be certainly a lag time of approximately a year from the time the data are essentially realized and when they would be analyzed, presented, and reviewed for regulatory approval which essentially would mean if at that point the vaccine was now available for potential access to the control arm, there could be some cross-ins.

This gets to, if I don't call it a flaw, a risk of accelerated approval and it's been acknowledged here. If you have an accelerated approval, does this truly compromise your ability to answer the question of interest. If it does, then I don't think the accelerated approval is acceptable because I think we need to get the answers here for the full approval.

On the other hand if the judgment is this is late enough in the process that any crossing in or lack of adherence to the control intervention would only in a minor way dilute this assessment, then I would consider it to be acceptable.

I'll just note here I acknowledge my NCI colleagues as they point out the ethical dilemmas. I've had this ethical dilemma for a long time as we've implemented accelerated approval in HIV settings and oncology settings.

Is it ethical to say I have enough evidence to bring forth to regulatory authorities approval of a new intervention and, yet, I'm still going to enter or follow people in a controlled trial to be able to get at what we recognize to be the ultimate answer that we know we have to get. That's a dilemma that I think all of us have to face.

But if in our judgement it is ethical, and we can adequately maintain adherence, then I think it is appropriate to consider this accelerated approval and this strategy so long as we're insured that this six to seven-year answer is going to be achievable.

DR. DAUM: Thank you, Dr. Fleming.

Dr. Sheets.

DR. SHEETS: Can I say I agree with everything you said?

DR. FLEMING: You sure can. The committee would be grateful and so would everyone else.

DR. SHEETS: I find the constraint in the argument of Dr. Fleming to be the ethical consideration of early termination or accelerated approval. Although the trial that he outlines and others have outlined ahead of him is probably doable, I think you would get to the point where clinically speaking it would be very difficult to continue on with the placebo arm.

I believe that the endpoint is CIN 2/3, although I would caution that we should continue to collect data on cytology, as was pointed out previously, because we are assuming here that colposcopy as the gold standard has no downsides and that's not true by any means.

As much as we would like to say that we are experts at this and we are perfect, we're not. The sensitivity and specificity of colposcopy leaves something to be desired. I think we need to look at the endpoint of a high-risk cytology and histology together.

People who have colposcopicly negative yet persistent high-grade abnormalities on PAP smears will in excisional data have high-grade histology. We just missed the lesion so we have to keep that in consideration as we go forward with the trial that you might outline.

I would advocate also to approach, as has been said before, high-risk women and try to endeavor to try to keep these trials demographically representative of what the future may be very soon in America and certainly include high-risk groups within that not only in terms of ethnic groups but socioeconomic groups that are at greater risk for the development of high-grade precancer and invasive disease with or without screen being present in those communities.

I do think that safety data, as has been pointed out before, is very important. Although we are using recompetent material here that represents what is probably exposed to a woman in general transvaginally, we are giving it now systemically in certain a larger dose than it has ever been inoculated into someone before or vaccinated to someone before. I do think it's important to continue on with rigorous safety controls in regards to these trials.

I'm not a clinical trialist and I won't tell you how to do that, but I think it's important to bring that data forward because in America although cervical cancer continues to be persistent, we are not going to impact on that cervical invasive rate, as has been pointed out, for 10 to 20, maybe longer years. In the interim the safety data is very important for these women who have been given this vaccination.

That's all I have to say. I would vote for only accelerated approval if it does not impact on the placebo group for the CIN 2/3 outcome. In regards to labeling, I think it might be premature but it would be for the prevention of persistence of high-risk infection and also CIN 2/3.

DR. DAUM: What was your interim endpoint?

DR. SHEETS: I would use CIN 2/3 as the interim endpoint with persistence of viral high-risk oncogenic type of analysis like Dr. Fleming has pointed out with the final endpoint being CIN 2/3.

DR. DAUM: Thank you very much.

Dr. Unger.

DR. UNGER: I think that this is a nontraditional vaccine and probably for that reason I think it's very important that we are cautious. For that reason I feel that this is our chance to really understand what's happening with this virus in a natural history setting so the studies have to be able to help us understand what we're preventing.

I, therefore, feel that we need a CIN 2/3 histology as an endpoint. I agree that the study has to be designed to help us understand what that CIN 2/3 endpoint means in terms of viral persistence of all the types and immune response.

The trial does need to be conducted in appropriate populations. I really don't see a public health imperative to do an accelerated approval. The package as far as what the recommendation would be, I think that we only -- we have to stick with what we know and what we've shown.

If we end up approving it based on prevention of persistent infection, that's how it's labeled and we can say what we think that means. Until we show something else, I think it's ill-advised to label it as showing something we haven't demonstrated.

DR. DAUM: Thank you very much.

Dr. Wilkinson.

DR. WILKINSON: First I would like to congratulate all involved in this commendable meeting on prevention of cervical neoplasia, a very important issue in women's health. I encourage accelerated approval considering the evidence at hand and the importance of the issue that we're dealing with.

I would favor an accelerated approval study format with the CIN 1, CIN 2, or CIN 3 as the accelerated interim and confirmatory points considering that CIN lesion is the usual source of the HPV infection and that CIN 3 lesions rarely regress. I would also make emphasis that cytology and HPV testing need to be applied in this process and safety net issues need to be addressed. Thank you.

DR. DAUM: We thank you, sir.

Dr. Felix.

DR. FELIX: I will make first a comment. I think one of the things that we saw presented were power calculations. Eventually it won't make a difference because if their studies are inappropriately powered they will not reach statistical significance.

But I will say that the numbers that I've heard in the power calculations look to me to be in error because of the studies examining or the studies that were used to determine that were lacking in taking into consideration incident disease.

I think the preponderance of data on persistent HPV is robust enough to use it as a primary endpoint. I think that a negative predictive value of nonpersistent HPV is very robust for a prediction of CIN 2/3.

I would favor using that as an endpoint or a co-endpoint. I would favor only accelerated approval once. If for primary endpoint for accelerated approval, I would chose persistent HPV at a minimum of one-year interval.

But I would favor only granting accelerated approval if FDA is aware that there is a complete maturation of the data for the confirmatory process already delivered, and that was very nicely stated would reduce the interval by one year.

The secondary endpoint for the confirmatory trial I would consider CIN 2/3 with a co-endpoint of cytology because, again, we don't know what a vaccine could potentially do to the reduction or change of predicted ability of cytology once a patient has a reduced inoculum. I hope that that data would come out.

As far as the labeling, for the preliminary approval I would label it, if effective, a reduction of HPV, a persistent infection by HPV that has been associated with development of precursor cervical cancer lesions and, if confirmed by the secondary endpoint, a prevention of cervical cancer precursor lesions.

DR. DAUM: Thank you very kindly.

Dr. Freeman.

DR. FREEMAN: I'd like to start with complimenting the NIH, NCI, and the industry and many others not here who have contributed so much to this important problem of HPV.

After giving this careful consideration, I feel that the traditional approach with CIN 2/3 as an endpoint is the safest in this particular situation having been involved and actually chaired the data monitoring committee.

I understand the complexity of how data monitoring committees can evaluate data and decide to make decisions for early termination. It's not a trivial matter and I would emphasize the data monitoring committee needs to be an independent committee totally and the principal investigator should not be involved in these decisions.

The reason that I favor this as an endpoint is this is an important decision here that will affect many millions of lives. We don't know the outcome. There have been randomized trials, and I can think of one in particular recently in lung cancer where a product was given to patients with lung cancer to try to prevent secondary lung cancer where, in fact, it turned out that the treatment was worse and more patients were dying in the treatment.

It was a subset of patients who continued to smoke as it turned out. It was very, very good preliminary lab data and clinical data to support this clinical trial. But that is one of the reasons I think that we need to be especially cautious with a study that can impact so much on patients' safety and their outcome.

However, I would say that the virologic, immunologic studies are very important and other studies. For example, we talk about mucosal immunity and factors in the vagina that could impact on infection.

For example, women in certain countries are prone to use vaginal medications more frequently than in others. Particularly if you're doing a study in different locations, these factors could possibly impact on the infection. We need to study these as well as part of the design of the clinical trials.

I would use the CIN 2/3 as an endpoint that both men and women will understand who will be receiving the vaccine. Eventually men possibly in other trials. Also that the physicians that have to take care of these patients will understand the significance of the trial and the endpoints.

Also I am concerned that if you give an accelerated approval, the information that's on the label may very well influence what happens to the definitive trial.

Obviously informed consents would have to be changed based on that your control arms may be affected, particularly if patients have not been fully entered and followed for enough time. The definitive answers here, or the most proximal answers to the real question which is whether the vaccine actually prevents cancer may never come to us. I think that is particularly important.

DR. DAUM: Thank you very kindly.

DR. GREENE: Thank you. A few comments. One is the idea that even a perfectly effective and perfectly administered vaccine could possibly eliminate all cervical cancer is somewhat naive. I would say it's comparable to expecting the elimination of cigarette smoking to eliminate all lung cancer. Nonetheless, it would be desirable if no one ever smoked cigarettes. I think there is a certain analogy there.

Next is a question that came up earlier in the discussion. What is the number we need to treat with a vaccine in order to eliminate one case of cancer? The number needed to treat, of course, is one over the absolute risk reduction.

In this case we don't know what the absolute risk reduction is so that we can't a priori calculate what the number needed to treat is. It may be that this kind of a trial might help to give us some notion as to what the number needed to treat is.

But it should be fair to people interested in vaccines that the number needed to treat could be very large to avoid one very serious outcome. Certainly that's true for hepatitis B.

It would be true for the number needed to treat with varicella vaccines to avoid one death from varicella pneumonia so that a very high number needed to treat I think would still be an acceptable indication for vaccine.

Next is that to me the most important point that we don't understand yet is the durability of a vaccine affect and this has very important implications in terms of who and when you would suggest to receive the vaccine. Who should the recipients of the vaccine be?

In theory, if the immunity was life long, then the appropriate recipients would be children at birth because that way you could be 100 percent confident that all persons would be protected at the time of first intercourse.

However, if the vaccine effect wans after 10 years, you would wind up with no net reduction in the incidence of infection so that we definitely need to know what the duration of the effect is in order to know who to treat with the vaccine.

The question of identifying "high-risk population" is extremely difficult. Women who are celibate life long basically do not get cervical cancer. Everyone else is "at higher risk."

Among those people you can identify people who are even at further greater risk on the basis of age at first intercourse and total life time number of partners. Women, for example, who are commercial sex workers have ultimately the highest risk next maybe to persons who are immunosuppressed, HIV, transplant recipients, etc.

It would be difficult for me to imagine asking parents to bring in their girls for immunization if they expected their daughters to have a very young age at first intercourse or to have a very large number of sexual partners life time. I would think that would pose some difficulties. As the father of a 16-year-old daughter, I'm particularly sensitive about this issue.

I'm not sure who we should label, how the vaccine would be labeled in terms of who the vaccinees should be and who should the targeted population be.

Finally, with respect to the definition of persistent infection, it's quite clear from the data that is already available in the literature, Woodman's paper and Lancet, Hoe and Burk's paper in the New England Journal of Medicine that two cultures separated by less than 12 months really describe only incident infection and not persistent infection.

Not cultures but PCR assays so that you would need -- it would seem to me that you would need to have two positive assays at least a year apart to define a persistent infection.

Finally, I am persuaded by the discussions that I've heard from the past two days as well as Dr. Goldenthal's assessment that the difference between accelerated approval and final approval might be 12 months.

I am persuaded from that, as well as the difficulties that have been acknowledged with attempting to complete a definitive trial in the wake of an accelerated approval, that the accelerated approval mechanism is not appropriate for this vaccine. My recommendation would be that the standard approval mechanism be used and that the endpoint be CIN 2/3.

DR. DAUM: And I thank you for your very cogent comments.

Dr. Pagliusi, please.

DR. PAGLIUSI: I'd like to thank the FDA for the opportunity to participate in this meeting first.

Secondly, I would like to express our respect to all the scientific community who worked on the development of these vaccines and brought this field so far.

Now, at this stage we believe that cervical cancer is not a visible endpoint and that an intermediate endpoint should be considered. We would favor CIN 2 and 3 as the most appropriate endpoint. However, if trials should take longer than expected, infection at endpoint would be a surrogate endpoint to consider only if a sustained protection is proven because from the public health point of view, a vaccine that should be boosted every year or every second year would be a challenge for coverage and compliance and may not be useful at all.

In this sense the laboratory wishes to accelerate the vaccine development and consistent with this line we would welcome the accelerated approval but provided that robust data is created to support long-term duration of protection and that the persistent infection is well correlated with CIN 2 and 3. We would favor the one-year interval between PCR positive points.

DR. DAUM: Thank you very much.

As a penultament speaker, Dr. Decker.

DR. DECKER: This vaccine is an exciting prospect. This has been a fascinating discussion. Prevention of ICC cervical carcinoma is a major public health goal so the stakes are high, as are the penalties for missed steps.

Based on our presentations and discussions, I conclude that first there is not presently the necessary proof regarding the effect on cervical carcinoma of vaccine induced changes in virologic or cytologic endpoints. As much as we think they're correlated, there is clearly some uncertainty as to what the effect on ICC would be are changes in those prior measures.

Secondly, a study endpoint of cervical carcinoma itself is infeasible for multiple reasons.

Third, as CIN 1 is not a clearly defined homogenous group, it's use as a study endpoint would be problematic.

Fourth, it appears that the study using CIN 2/3 as the primary endpoint could be conducted within time frames and expenses that are commensurate with other modern vaccine efficacy trials.

Finally, my own observation, I believe that licensure of a vaccine is likely to lead to widespread use irrespective of license indication. This has implications regarding the need for high confidence regarding the outcome of confirmatory studies to reduce the risk that a widely used vaccine would later be shown to be poorly effective.

At the same time, concern regarding the ability to conduct these confirmatory studies, and moreover the need for complete safety data in both males and females before any licensure because it seems to me that any rational use of such vaccine would target both males and females such as we do with rubella vaccine.

Accordingly, I would recommend that the primary endpoint should be efficacy against CIN 2/3. The study should incorporate secondary or observational objectives regarding virological and cytological outcomes so that we can improve our understanding of the relationships between these outcomes and CIN 2/3, and perhaps allow for simpler studies subsequently.

Similarly with Dr. Katz, I would encourage the inclusion of nested substudies to explore related epidemiologic and clinical questions. The study design should not be predicated on a plan for accelerated approval but a design such as I have just described would permit reconsideration of accelerated approval should findings during the course of the study warrant that reconsideration.

Finally, the license indication should be based on the outcomes proven in the study which should also explain the basis for a belief that these outcomes are relevant to the prevention of cervical carcinoma.

DR. DAUM: Thank you. To bring this discussion to a close before I forget to say it, I would like to really commend all members of the committee and members of the sponsors who presented data to us and, of course, our FDA colleagues.

I think this has been a very wonderful discussion. I think the committee has transcended their usual degree of excellence by having a very lively debate and consideration of all points of view.

Having said that, I will very briefly give you mine. I think that one of the things we haven't said a lot about is that if we have a potential preventive strategy to stop a disease like cancer that we ought to do everything in our power to ensure -- I think everyone in this room would agree with this -- to ensure that it be developed to get the definitive answer of does it or doesn't it prevent this horrible disease as quickly as we possibly can.

Having said that, I favor an accelerated approval strategy but only if things can be put into place during that. First of all, what persuades me to be in favor of it is Dr. Goldenthal's notion that we might be able to get a confidently effective vaccine to the public a year earlier. I am sort of shooting for that.

It would have to be done predicated on something in place to definitively answer the question about an important endpoint. I agree with everybody else that CIN 2/3 is a reasonable surrogate for the definitive endpoint.

I think viral persistence were it to be shown could be the interim endpoint. I don't know what the exact definition is. We've heard many different attempts at it. For lack of a better one, I think I would accept the one-year cut off as a definition of viral persistence.

I would not favor interim approval if it compromised gathering of appropriate safety data, or if in the views of the people responsible for the study design compromise the integrity of the study to get to the definitive endpoint.

In that circumstance I would rather let the year pass because I would not want uncertainty or erosion of public confidence once it was decided this vaccine were good enough for general public use.

I, like others, have commented, particularly on this side of the table, am very excited about what I've heard here in terms of a prospect of getting a preventive measure like this out. I would like to encourage everybody who is working on this problem to move things forward as fast as they possibly can.

I think that brings this discussion to a conclusion. Before anybody moves, there are two items of importance to deal with. One is we have a minor and very quick presentation to make. It will take about 10 or 15 seconds to walk the presentation over. Another two or three to gather it.

This is your moment in the sun here, Ms. Cherry. Nancy, on behalf of the agency, the committee, and I think really everybody else in this room if I could just extrapolate for a moment, there is only one Nancy Cherry, folks, in this universe and she cannot be replaced and will be missed sorely. Thank you so much.

MS. CHERRY: Thanks to all of you. I was going to wait until the end of the closed session today to say goodbye to my committee because it's been such an honor and a privilege to work with them. Those are hackneyed words, I know, but I really, really mean them.

I've taken this job probably more seriously than I should have and more personally than I should have and I've never gotten over my feeling that a kid would have of being totally awed by the group I'm working with. Not just your reputations, not just the importance of what you're doing, but also how good the people are. How good all of you are.

It took a lot of thinking to decide when to announce that I was ready to retire and you all are certainly making it difficult today.

DR. DAUM: Maybe you'll reconsider.

MS. CHERRY: Thank you all.

DR. DAUM: Now, before everyone starts milling around, I would like to briefly take a line of demarcation here from Dr. Goldberg over and just ask the troops whether you would like to break for lunch or whether you would like to continue working through. Break for lunch? Work through? Okay. We will then take a 10-minute pause, let the room clear, potty break, etc., and we'll reassemble with the review of the lab.

(Whereupon, at 12:24 p.m. off the record until 12:43 p.m.)

DR. DAUM: Okay. Welcome back everyone, committee members, FDA folks. This is an open session on the briefing on activities in the Laboratory of Bacterial Toxins. We are going to try and complete this in a succinct but thorough style and begin by calling on Dr. Walker.

Did I see him? There he is. Thank you, Dr. Walker. Welcome. He will talk to us about the organizational structure and overview of research and regulatory responsibilities in the Division of Bacterial Parasitic and Allergenic Products.

Dr. Walker.

DR. WALKER: Thank you. Good afternoon. Hopefully everybody can hear me now.

In a few minutes you're going to hear presentations of the research of Dr. Vann and Dr. Schmitt. For this reason I've been asked to give a little introduction to their presentations by giving you an overview of the Division of Bacterial, Parasitic, and Allergenic Products.

I'll do that in two ways. First, I'll talk about the functions of the division and then I'll talk a little bit about the organization of the division to meet these functions.

Very briefly, our mission of functions is to assure safe and effective products for control of bacterial, parasitic, and allergenic agents affecting human health. This involves a number of activities by the people in the division. One of those activities is research. The other is review.

Something I might mention about putting these two together, it's sometimes hard for these people to manage their schedules because things coming in for review are not always something that somebody can plan for so this creates a scheduling problem. As you'll see when you hear the presentations this afternoon, these people do get the work done anyway.

In addition to the review, there's post-licensure surveillance with the things that it involves like inspections, log-release testing, and review of label and promotional activities. Then we continue to consult with outside organizations like WHO and others that are dealing with problems that are pertinent to the FDA.

I just want to show this to illustrate the involvement of the FDA research and reviewers in the lifetime of a product. As you can see here, the important take-home message from this slide is not all the individual components under each section of the development of a product, but the fact that there's activities that go on under each section of the development of the product.

Like here in very early stages meeting with sponsors, providing some guidance, review of original submission and subsequent amendments, technical advice for product and assay development, review of product manufacturing data, determination of product specs, and, of course, continued discussion with sponsors.

I don't want to belabor this since we're moving along with regards to time but, as you can see, after license activities there is present the product to the advisory committee, continue to have dialogue with the sponsors, and continue to evaluate the products and review the procedures that are being used in manufacturing and so forth.

The point, like I said, that comes out of this is once a product is licensed, the story is not over. The job continues. Post-licensure we still have to review biological deviation reports from industry, participate in inspection of licensed products, view post-approval commitments, so forth. It's a long-term ongoing process.

To make the challenge even greater in dealing with all these products, there's a tremendous variety of products that research and reviewers have to deal with. You can see by looking at this figure about new and improved products that might be possible in the next 10 years.

There are respiratory pathogens dealing all the way from life-threatening diseases to those that cause ear infections, sexually transmitted pathogens, diarrhea-causing pathogens like campylobacter and so forth, and other mucosally trafficking pathogens like salmonella, heliobacter, and so forth.

There's quite a variety of pathogens, most of these mucosal pathogens that are shown on this slide. Also we need to think about pathogens that are not necessarily mucosal like those that are like materia, lyme disease that are encountered by penetrating inoculation.

And then something that is, of course, is very relevant these days, special pathogens, biological terrorism type agents like franciscella santhrasious, clostridium botulinum, franciscella tularendous, and arensia peskas.

In addition to these pathogens, we also have products, the allergenic antigens dealing with latex antigens, cockroach, and various plant antigens, and skin test antigens. I'm just trying to give you the picture that there is a variety of types of products that our people have to be able to deal with over the next couple of years.

To meet these challenges the Division of Bacterial, Parasitic, and Allergenic Products is divided into eight laboratories. There's the immediate Office of the Director with myself. I have an excellent deputy director Carolyn Deal.

Carolyn and I are supported by people who are regulatory and administrative staffs. We work together to help all these people in the various laboratories accomplish the jobs that they have to do.

We have eight laboratories. The Laboratory of Respiratory and Special Pathogens, Laboratory of Bacterial Toxins, which you'll be hearing from today, Laboratory of Mycobacterial Diseases and Cellular Immunology, Laboratory of Methods Development and Quality Control, Laboratory of Immunobiochemistry which is allergenic products, Laboratory of Biophysics, Laboratory of Sexually Transmitted Diseases, and Laboratory of Bacterial Polysaccharides.

If you look at these laboratory names, you'll see that they are identified by the types of pathogens and types of approaches they use. I think it is also important to realize that the talents and the resources that are present in these different laboratories we brought together on certain focus areas.

These are some of the focus areas that we're currently dealing with in our division. One of those is standardization of assay methods for bacterial, parasitic, and allergenic products. Also a large group is focusing on pertussis and other toximediated diseases. You'll hear a little bit of that in just a few minutes. Mycobacterial and other intercellular parasites.

It was mentioned this morning how important mucosal immunization is and how much we need to know about that. We've got work going studying mucosal pathogenesis and immunization, products to combat bioterrorism. We also have a very active group dealing with allogenic products.

What I'm trying to show in this slide on the screen now is that you can take those laboratories and those focus areas and sort of think of it in terms of a matrix. You can just see how we pull our resources together to accomplish things.

All of the laboratory names are shown across the top. I've abbreviated some of them just to make it easier to read. The various focus programs are shown going vertically.

If you look at assay standardization, that's something that involves everybody, all the laboratories to some degree or another. Pertussis and toxinmediated diseases involves work from the Laboratory of Methods Development and Quality Control, as well as biophysics, toxins, and the respiratory and special pathogens groups.

Microbacteria is really one laboratory that's dealing with that. Mucosal pathogenesis immunization, one laboratory dealing with that. We now have six laboratories out of the eight that are dealing with some aspect of bioterrorism agents and two of the laboratories, Biophysics and Laboratory of Immunobiochemistry, they are dealing with allergenic products.

I'm going to go very quickly through this. I have identified these laboratories and this will just give you a little bit of flavor of the type of research that is going on. The Laboratory of Biophysics they use various instrumentation such as NMR to characterize biopolymers.

Examples are given here. And macromolecular assemblies so they bring a lot of technology that really opens up new doors for some of us in the other sections. They also have computer or simulation methods for collector myogen analysis.

Laboratory of Bacterial Toxins, we're not going to say anything about that because you're going to be hearing about their program in just a few minutes. Laboratory of Respiratory and Special Pathogens conducts structure and fluction studies of various toxins and regulation of virulence factors, B. pertussis and B. anthraces.

Laboratory of Bacterial Polysaccharides is another rather large laboratory in our division. They characterize immune responses to polysaccharide and conjugate vaccines and work toward standardization of methods for relevant clinical applications and develop physical and chemical methods for improved evaluation of license and experimental vaccines. There's quite a lot of work there to do with the polysaccharide and the conjugate vaccines.

Mycobacterial Diseases and Cellular Immunology are evaluating immune responses to intercellular bacteria, mycobacteria and tuberculosis. They are assessing vaccine strategies particularly for tuberculosis.

Enterics is looking at pathogenic mechanisms such as invasion mechanisms of campobacteria and shigella, hormonal controls of conococcual pathogens, mucosal immunity that will help us understand not only these but other pathogens infecting mucosal services.

The Laboratory of Methods Development and Quality Control, as the name suggest, is one set up to develop, standardize, and evaluate quality control methods for bacterial vaccines and develop and evaluate and apply the serological methods to measure immune responses in vaccine trials.

Also an aspect of the overall lab accreditation project the FDA is doing now, people in this laboratory help coordinate the quality assurance activities within our division, provide leadership and initiative to accredit to the CBER Quality Control Testing Laboratories.

The final laboratory I want to mention is Laboratory of Immunobiochemistry where they are looking at not only the allergen structure inflections but the immune responses caused by these allergens, as well as ways to modulate these immune responses.

As you can see, we have a variety of things going on. We have a lot of talented people. In just a couple of minutes you'll hear from two of them and have a better appreciation for what they are doing. Thank you.

DR. DAUM: Thank you very much, Dr. Walker. Are there any comments or questions, concerns? Thank you again. I appreciate your time.

I would like to next introduce Dr. Willie Vann in one of two hats that he'll be wearing in the next little while. This hat is as the Director of the Laboratory of Bacterial Toxins, one of the laboratories that Dr. Walker mentioned.

When Dr. Vann has concluded his remarks as director of the laboratory, he will then transform into Dr. Vann in charge of his program and tell us a little bit about that. We'll actually have a hiatus in between for committee comment if there is.

Dr. Vann, as director of the laboratory tell us what's going on.

DR. VANN: The Laboratory of Bacterial Toxins is organized into three sections, Neurotoxin Section, Glycobiology Section, and Corynebacteria Section with three PIs. I'm currently the acting PI for the Neurotoxin Section. We have recently since our review completed the hiring of a new PI, Dr. James Keller for the Neurotoxin Section.

This section currently has two post-doctoral fellows and a biologist. They work primarily on clostridial neurotoxins botulinum and tetanus. The Clycobiology Section has currently a post-doctoral fellow and a biologist.

This post-doctoral fellow is working on an anti-bioterrorism project with anthrax. The Corynebacteria Section currently has a post-doctoral fellow and a microbiologist, newly hired microbiologist technician. This post-doctoral fellow is also working on an anti-bioterrorism anthrax project.

The laboratory has product responsibilities for bacterial toxoid vaccines against botulism, diphtheria, and tetanus, vaccines containing toxoids as components of polysaccharide conjugate vaccines, and botulinum toxins, active toxins, as therapeutics for diseases involving muscle contractions.

Our review responsibilities include review of biological license applications, biological license application supplements, and investigation of drug applications relating to these above products.

In addition, we have responsibilities for review of lot release protocols for botulinum toxin, annual and prelicense inspections of manufacturing establishments. We participate in efforts to monitor and improve vaccine safety and potency. We evaluate manufacturing deviations reported to CBER. In addition, we provide expertise to the Office of Therapeutics on glycoprotein therapeutics.

The FDA has an anti-bioterrorism initiative. The Laboratory of Bacterial Toxins has incorporated into its existing program research projects on bacillus anthraces which are on polysaccharide biosynthesis and iron metabolism.

The laboratory has organized to meet its existing and future obligations. Existing toxoid vaccines require a maintenance of expertise in C. diptheriae and neurotoxins. The therapeutic use of plastritial neurotoxins requires a clear understanding of how these active toxins and not toxoids work.

Glycoconjugate and recominant vaccines require expertise on molecularbiology and carbohydrates. CBER has an expanded obligation in glycoprotein therapeutics. Thus, this requires a leveraging of expertise across the center of which we are a part. We have integrated into our existing programs an anti-bioterrorism effort.

The general research objectives of the Laboratory of Bacterial Toxins want to determine the function and structural basis for the potency of vaccines and neurotoxins. Questions we're answering is, (1) what determines the specificity of the interaction of a neurotoxin with the nerve cells, (2) can we replace expensive and low-precision in vivo assays with in vitro assays that are based on biochemical measurements that are functions of toxins?

A large part of our effort is to define new targets for the control of viral bacteria. To this end we are asking two questions. (1) What are the systems for iron metabolism and C. diptheriae, an essential component of virulence, (2) and how do bacteria make their protective coats? Later you'll hear Dr. Smith talk about his efforts in this area.

That concludes my summary of the organizational laboratories.

Yes, sir?

DR. DAUM: Dr. Faggett, please.

DR. FAGGETT: I have just one question. I've had a lot of action with botox injections and all that. Do you folks look at efficacy of botulinum toxin? Is that part of your responsibility, too, or is it just the basic science of it?

DR. VANN: We do review license applications. We have the product lab for reviewing license applications for botulinum toxin indications if that answers your question. We also review IND submissions for various indications using botulinum toxins.

DR. FAGGETT: So you would have information on efficacy of toxin?

DR. VANN: Exactly. We are part of the committee that -- we are part of the committee that actually assesses the efficacy of the toxin for various indications.

DR. FAGGETT: Of course, in D.C. we've had a little experience with anthrax recently and I was wondering are you also involved in development of the anthrax vaccine or what is the relationship?

DR. VANN: The objective of our research on anthrax is to actually develop targets for development of vaccines against anthrax. For example, if there were a key protein that we found on the surface of anthrax or that was produced by anthrax involved in iron metabolism, that could be developed by someone as a potential vaccine.

Or if there was something we found in our research that was essentially for the survival of an organism or the virulence of the organism, that would provide information for someone to develop an antagonist against the organism. This is lying foundation research for the development of new things to combat anthrax.

DR. FAGGETT: Thank you very much.

DR. DAUM: Dr. Kohl, then Ms. Fisher.

DR. KOHL: Several questions. In the material we got there is a little table on the funding of the Laboratory of Bacterial Toxins.

DR. VANN: You're referring to the book?

DR. KOHL: I'm referring to the book from 1997 to 2001. Is that funding -- that does not include personnel, I presume?

DR. VANN: The funding that I have listed there does not include personnel.

DR. KOHL: Okay. So that is basically research funding.

DR. VANN: That is research funding, yes, which includes materials, services, and everything else that is not personnel.

DR. KOHL: Okay. If you had to estimate what the total budget is of the division, could you give me an estimate for that counting personnel?

DR. VANN: I didn't understand your question.

DR. KOHL: How much of the personnel work?

DR. VANN: I don't have an answer to that.

DR. KOHL: What percentage of your time is regulatory? It looks like a lot but I would like to get a feel for that. Who is primary involved in the regulatory activities?

DR. VANN: Okay. That depends upon the person. Dr. Schmitt and I, I would say, could spend up to 50 percent of our time doing regulatory work. Sometimes it's a little bit more. Sometimes a little bit less. If you look at the organization chart, that varies with some of the other people in there.

For example, the post-doctoral fellows, for example, IRTA fellows who are post-doctoral fellows don't spend any of their time on regulatory work. It depends on the personnel.

DR. DAUM: Thank you. Ms. Fisher.

MS. FISHER: It looks like you have a very important function, particularly now that anthrax has been added. Your organizational chart looked very small though to me. Do you have enough people in your organization to fulfill all of the duties that you are supposed to fulfill?

DR. VANN: Thank you.

MS. FISHER: I mean, could you use more help is what I'm saying?

DR. DAUM: What was that green stuff you guys were exchanging?

DR. VANN: We do our best and we are always trying to get additional resources to actually help. Some things are beyond our control but we do the best we can.

MS. FISHER: Well, I understand that. I guess as a consumer I'm very concerned that you have adequate resources and staff to fulfill the function that you have.

DR. DAUM: Dr. Deal, do you want to respond to that?

DR. DEAL: Yeah. My name is Carolyn Deal. One thing I think might be helpful to the committee is to clarify that in addition to Dr. Vann's lab other laboratories within the division also have a regulatory responsibility for the anthrax vaccine.

In fact, his lab is not the primary one for that vaccine's review work. That may be helpful in your consideration of some of the workload of the regulatory responsibility.

DR. VANN: Right. There's an entire division that deals with the clinical aspects that are related to these things. What we are responsible for primarily is product and things that are related to product but that is still a lot of work.

DR. DAUM: Dr. Kim, please.

DR. KIM: You briefly indicated that CBER has expanded to include glycoconjugate therapeutics. Can you give me just a couple of examples of what they are?

DR. VANN: Yeah. One of the ones that is very -- probably one of the ones that earns the most money for biotech companies is monoclonal antibodies. One of the first was TPA, erythropoietin just to name a few examples. There are many others that I don't know anything about.

The reason I actually am somewhat involved in that is that I started out and my major expertise is in carbohydrates.

DR. DAUM: Okay. I think we should move on to the next presentation which will be by Dr. Willie Vann where he will describe his own research activities and his own research program.

Dr. Vann.

DR. VANN: For the sake of clarity of presentation, I will discuss two of the research projects in the research program. The other projects are actually listed in the book.

We are investigating the biosynthesis of capsule of polysaccharides in our model system. We're studying the biosynthesis of polysialic acid. Work has been done largely by a post-doctoral fellow who is no longer with us and a technician.

E. coli and niceria meningitides which are encapsulated with polysialic acid are generally associated with invasive disease such as meningitis and urinary tract infections. These are the structures of the common known polysialic acid capsule of polysaccharides. We are using in our system the E. coli K-92 which is alternating 28/29 polymer.

As an added bonus to studying polysialic acid, we also are studying polysialic acid metabolism. Polysialic acid plays several roles in microbial pathogenesis. The same system general pathway that is used to synthesize polysialic acid is also important for synthesizing other virulence factors for pathogens like polysaccharide, capsule of polysaccharide, and the eukariotic cell receptors for toxins and adhesives.

The genes that encode polysialic acid synthesis are arranged in three regions. The central region, Region 2, is specific for the polysaccharide. We have concentrated our efforts on the understanding the function of the genes that are in this region.

Our approach has been to purify and characterize enzymes encoded by the gene cluster and in the process assign various genes to functions within the pathway.

Our recent efforts have concentrated on the polysialytransferase, the enzyme that actually perlimerizes the substraight CMP sialic acid into a polymer and exports it through the cell surface. This enzyme is a membrane bound enzyme and is characteristic of glycosialytransferase and other pathogen bacteria.

What we would like to know under the question of the mechanism of this enzyme is, one, how the reaction is initiated, how the chain is elongated, what components within that complex are responsible for initiation and elongation, and how is this chain fidelity of the repeat unit maintained?

We can summarize our findings thus far on the elongation reaction here. The K92 polysialytransferase itself cannot initiate synthesis, thus it requires other components. It will elongate or use all of the polysialic acids that we know as acceptors. It has a preference for 208 link acceptors. It seems also to have a preference for a hydrophobic aglycon.

We can use our current model to explain data elongation or list it here in the next two slides. The first model proposes that there are three sites, an acceptor site which binds the preferred alpha 2-8 acceptor and two catalytic sites, one that forms a

2-8 linkage and one that forms a 2-9 linkage.

Once the 2-8 and 2-9 linkages are formed, the newly formed 2-8 linkage moves to the preferred site and then the reaction starts over again. The alternative mechanism proposes that there is a 2-8 binding site and a single catalytic site. Once the 2-9 linkages form, the enzyme undergoes a comfirmational change to allow it to bind the newly formed 2-9 linkage and then a 2-8 is formed and then you start over again.

We have approached the initiation reaction in using two different types of methods. The question is how many things are involved in initiating the reaction and how does that occur.

One of these is using complementation. We know that we can separate the initiation reaction from the elongation reaction by simply cloning out the polysialytransferase gene. In the complementation experiment what we've done is genetically add back various components from the gene cluster and ask the question can we restore the ability to initiate synthesis.

Another approach that we've taken is to try to estimate the molecular weight of the complex that is actually doing the reaction. We've done this using a method that is particularly suited for crude systems, namely radiation target analysis. The size of the active complex is inversely proportional to the amount of radiation.

Our current model for the initiation reaction is listed in this -- given in this slide. First of all, what we've learned is that the initiation reaction and the elongation reaction probably uses the same size complex that consisting of a dimer of the polysialytransferase.

The complex actually transfers to some membrane bound glycolipid acceptor. The groin chain stays attached to the membrane acceptor. Now, what we believe is that the acceptor is probably some glycolipid. The next phase of our research involves understanding what this glycolipid acceptor is.

Our efforts to do that will be to chemically synthesis sialic acid analogs that get incorporated into the membrane but terminate, and then selectively using some of the newer chemistry to tag this acceptor and then extract it and characterize it structurally.

The other project that we are studying is the binding of tetanus toxoid C fragment to ganglioside. This has largely been done by post-doctoral fellow Heather Loach, along with a graduate student in the Laboratory of Biophysics.

Clostridial neurotoxins are very lethal and they are produced by C botulinum and C. tetani. They are the active agents in tetanus and botulism and they chemically inactivate toxins or serve quite well as vaccines.

We have mentioned before botulinum toxins are a potential bioterrorism agent. However, botulinum toxin is also a therapeutic for diseases involving severe muscle contractions.

These toxins are organized structurally into three domains, catalytic domain, central translocation domain, and a receptor domain. The receptor domain is the part that actually binds to the ganglioside on the surface.

This domain is organized to be two domains itself. One of those is elected jelly roll domain, and the second is what we call a beta tree foil.

The binding of tetanus or botulinum toxin to ganglioside is an essential first step in pathogenesis. It binds to the nerve cell, gets internalized, and then eventually goes and cleaves a snare protein that prevents neurotransmitter release.

Most of the protective antibody to tetanus and botulinum toxin is against the binding domain. For that reason scientists are currently developing recombinant vaccines against the binding domain of botulinum and tetanus toxin.

Our approach to determining the binding site is to use available biochemical and crystallographic data and then to make use of molecular model and to predict likely binding sites. We then test these likely binding sites by site directed mutagenesis and then look at the binding properties of the resulting mutants, and then in a reiterative process refine our model to hone in on the binding site.

We use two types of molecular modeling experiments to actually do this. The first is homology modeling. In this type of experiment what we would do is superimpose the three dimensional structure of proteins that bind carbohydrates on the C fragment of tetanus toxin and look for motifs.

Once we found and narrowed in on an area, we then use our molecular docking which is based on the crystal structure of the oligosaccharide and then dock or look for energy at various locations of that oligosaccharide on the three dimensional structure of the C fragment.

Using that type of methodology we came up with two regions for mutagenesis. Both of these regions are located on the beta tree foil section of the toxin. We mutated all the residues here and what we found is that one of these is actually essentially for binding. Thus, we've concentrated on Region 1 as being the potential binding site to ganglioside.

We later went back and did further modeling using this docking methodology and identified two other residues here, histamine 1271 and aspartic acid 1222. We mutated those, did binding studies. In these traces which illustrate the extent of binding, both of these indeed are involved in the binding of the ganglioside to the C fragment.

Thus, we've refined out model here and thus far we have defined the binding site on to tetanus C fragment as including these residues, the histamine and the aspartic acid I mentioned before, the tryptofame 1289. During the process of this research these residues were identified in literature.

In the future what we would like to do is refine our model using experiments based on the entire ganglioside since most of these experiments were done with a fragment of the ganglioside. We would like to determine the effective side chain characteristics on the kinetics and thermodynamics of binding. It determined the effect of these mutants on binding to nerve cells.

DR. DAUM: Thank you very much, Dr. Vann. Do we have committee questions or comments?

Dr. Griffin, please.

DR. GRIFFIN: This is probably totally obvious to anybody who is a bacteriologist, which is not me. The polysaccharide glycosialytransferase that you are studying, I assume that they are involved with developing the capsules of these E. coli and are important for virulence of these particular organisms and are potentially targets or something like that?

DR. VANN: That is exactly right. These glycosialytransferase are the enzymes that actually make the polymer. They are part of the machinery that make the polymer and export it and put it on the cell surface.

More specifically, if glycosialytransferase is negative mutants of bacteria, encapsulated bacteria are a capsular.

DR. GRIFFIN: And are then less virulent?

DR. VANN: And are then either not virulent or less virulent. In strains were capsule is essential, they are not viral.

DR. DAUM: Pneumococcus for example?

DR. VANN: Pneumococcus, for example, has been shown with E. coli K1.

DR. DAUM: Dr. Kim.

DR. KIM: I guess one question, I don't know, you may have that on your slide, do you cross-complement between K1 and K12 glycosialytransferase?

DR. VANN: We've never tried that.

DR. KIM: K12?

DR. VANN: No. What we can do is we can cross-complement it between K5 which is a totally different polymer and certain regions of that genome.

DR. KIM: My question is if you have a K1 mutants what happens if you complement with the K12 glycosialytransferase?

DR. VANN: With K12?

DR. KIM: K92. I'm sorry.

DR. VANN: K92. Oh, okay. That's a different story. K92 and K1 polycosialytransferase are interchangeable. In fact, that's the way we do the experiments. We do the experiments using mutants of K1 since most of the mutants have been made with K1. What we do is we take K92 glycosialytransferase gene and put it into K1 to study the system.

DR. DAUM: Dr. Palese, please.

DR. PALESE: Which laboratories are sort of competing in your field?

DR. VANN: Which laboratories out in the rest of the world?

DR. PALESE: In the rest of the world, yes.

DR. VANN: There are a number of them. One, Dr. Vemmer at the University of Illinois in Urbana. Dr. Silver who is also a collaborator but he is also somewhat competition. Dr. Troy who is --

DR. PALESE: Where is Silver located?

DR. VANN: University of Rochester. There is Dr. -- in the synthesis of polysialic acid, Dr. Stephens was part of the advisory committee. Then in Germany Dr. Frosch works on niceria. There are a number. And there are a few that occasionally you see another paper pop up with something on the system. There's Cheuw Wong who is a synthetic chemist who actually did some experiments directly on solid transfers which was directly competition. Then there's a group in Taiwan. I don't know whether they are still working on it. Does that answer your question?

DR. PALESE: Yes.

DR. VANN: Okay.

DR. DAUM: All right. If there's no further input -- there is further input. Dr. Snider.

DR. SNIDER: I wondered if you would just briefly comment since there is a bit about the anthrax in your section what you're doing and what the objective is and how that fits in to the rest of things that are going on.

DR. VANN: You have to understand that these anthrax projects are actually new projects. When we wrote this, these projects were just getting started. Briefly I can tell you what I'm doing and I can sort of hint to what Dr. Schmitt is doing but he can answer that question himself.

The fellow in my lab is looking at a group of genes that were discovered on a virulence plasma for bacillus anthraces which seemed to include how uranic acid synthesis and how uranic acid has actually been associated with other pathogenic bacteria such as Group A scrapococcus.

The question is why is it there, what's it doing. We are just in the beginning of characterizing those genes to see whether they're functional, what those gene products do, and then later ask questions like what does it have to do with hurdles.

DR. SNIDER: Thank you.

DR. DAUM: Okay. If there are no further questions, Dr. Vann, we thank you for both of your presentations. We will now hear from Dr. Michael Schmitt who is the Director of the Corynebacterium Laboratory and the overall structure of the laboratory of bacterial toxins.

Dr. Schmitt, welcome. We need you to probably adjust the microphone down. Talk right into it.

DR. SCHMITT: How is that?

DR. DAUM: That's fabulous. Thank you.

DR. SCHMITT: So what I would like to present today is a very brief overview of my research program in the Laboratory of Bacterial Toxins. The focus of my research is the characterization of iron transport systems in the corynebacterium and bacterium diphtheria which is the causative agent of diphtheria.

While the incidents of diphtheria has declined dramatically in recent decades in the United States and in other developed countries primarily due to the widespread use of the vaccine, a number of recent studies have indicated that greater than 50 percent of the adult population lacks adequate immunity to diphtheria and is potentially susceptible to disease. This is primarily due to waning immunity and failure to receive booster doses of the vaccine as adults.

Now, the vaccine is in the activated form of the diphtheria toxin known as toxoid and it is recommended for adults every 10 years. Additionally, since the vaccine is primarily directed against the toxin, it fails to eradicate the carrier state of the organism. Fully vaccinated and healthy individuals can potentially be carriers of highly virulent organisms and potentially introduce these into susceptible populations.

Another alarming factor regarding diphtheria was the recent epidemic in the newly independent states of the former Soviet Union which occurred in the mid to late 1990s.

This is the largest outbreak of diphtheria to have occurred anywhere in the world in the last 40 years and I think it illustrates the important point of how quickly a disease like diphtheria can reemerge when we fail to keep an adequate vaccination of the population and also when there is a partial breakdown in the medical infrastructure which had occurred at this time.

So the organism I study is corynebacterium diphtheriae. It is a gram positive aerobic nonsporulating bacteria. It is related to the microbacterium in streptomyces, a group of organisms. And it is the causative agent of diphtheria with the primary virulence being diphtheria toxin which has been extensively studied at the biochemical level. We actually know quite a great deal about its structure and function.

We also know a great deal about how the toxin is regulated which has been an interest of mine over the years and, in fact, has been known for over 60 years that the diphtheria toxin is regulated by the iron concentration in the growth media.

In fact, the human host is generally believed to be very limited for iron with regards to invading bacterial pathogens and, in fact, this low iron environment if the host is generally thought to be a signal to activate certain virulence factors such as diphtheria toxin.

The tox gene, which is the structural gene for diphtheria toxin, is regulated at the transcriptional level by DtxR, the diphtheria toxin repressor protein, and iron when iron functions as an essential co-repressor in this system.

When the organism is grown in a high iron environment, iron will bind the DtxR causing it to undergo a conformational shift which allows it to bind to a region that overlaps the promoter for the tox gene, thus inhibiting transcription and blocking production of diphtheria toxin.

In a low iron environment, which is the environment thought to exist where the bacteria colonizes in the upper respiratory tract of humans, iron is not available to bind to the DtxR and, therefore, DtxR cannot block transcription and transcription of toxin proceeds and production of diphtheria toxin occurs.

So my primary research objectives are to identify and characterize new virulence determinants and C. diphtheriae whose expression is predicted to be coordinately regulated with that of diphtheria toxin. That is regulated by iron and presumably DtxR.

My primary emphasis has been looking at heme-iron transport systems in C. diphytheriae. Heme-iron transport systems or heme-iron utilization systems have been well characterized in gram negative bacterial pathogens where they have been shown to be important virulence factors in many cases.

They have also been shown to be iron regulated in a manner very similar to how the tox gene is regulated in C. diphtheria. Some of my initial studies when I arrived at the FDA was to demonstrate that C. diphtheria could, indeed, use a variety of host compounds such as heme and hemoglobin and transferrin as essential iron sources.

However, the mechanism for how it used iron from heme and hemoglobin was not known and this was one of the projects that I initiated. I set out a strategy to try to characterize this system.

So the strategy I followed was to initially isolate mutants in corynebacterium that were unable to use heme and hemoglobin as iron sources. And to complement these mutants with a plasma library carrying C. diphtheria DNA, and that ultimately characterized the genes and the products on these complimenting clones. And also to look at the molecular mechanism of how some of these genes might be regulated. Are they regulated by iron like the tox gene.

So I isolated a number of mutants in corynebacterium that were unable to use heme or hemoglobin as iron sources and then proceeded to compliment these mutants with a plasma clones carrying C. diphtheria DNA and identified two distinct groups of clones, one represented by placid PCD293 which carried a gene that I term HmuO, and another group of clones represented by PCD842 which carried a small operon of three genes which I call HmuTU and V.

The product of the HmuO gene encoded heme oxygenase which have been well characterized in ucariotic systems but this was the first report of the heme oxygenase in bacteria. What heme oxygenase do is they degrade heme shown here, but the subsequent release of iron in a heme breakdown product.

What we think HmuO is doing in the heme iron utilization system is that it will act on the heme once it is traversed the cytoplasmic membrane breaking down the heme and releasing the iron into the cytosol making it available for the cell.

The other clone that I identified that could complement some of these heme transport mutants encoded three genes that appear to be organized in a single operon termed HmuTU and V. These showed a high degree of homology to heme transport systems that have been identified in gram negative bacterial pathogens. We think it has a similar role. Actually, we went on to demonstrate that it had a similar role in C. diphtheria.

What I'm showing here is a model of what we think is going on in C. diphtheria and possibly other gram positive bacteria with regards to heme transport and the utilization of heme as a iron source.

What we believe is at the HmuT protein, which we showed was a lipo protein is anchored to the side of plasmic membrane by means of a lipid moiety so it's basically tethered to the cell and the remaining portion of the protein which is exposed on the extracellular surface is available to bind to heme or hemoglobin.

We then believe it delivers heme to a permease complex located in the side of plasmic membrane which is composed of HmuU and HmuV proteins. This facilitates the transport of heme into the cytosol where then the HmuO protein, the heme oxygenase, can then act on the heme breaking down the molecule and releasing the iron.

As I said at the outset, my interests were not only to identify the components and proteins involved in the transport utilization system, but also to understand how some of them are regulated. Are they coordinately expressed with the toxin.

In the process of sequencing the HmuO gene, I identified overlapping the promoter region for the HmuO gene. Just upstream of the actual coding region was a sequence that showed a high degree of homology to the consensus DtxR binding site which could indicate that the HmuO may well be regulated by DtxR and possibly iron.

Subsequent studies, DNA footprinting and various promoter fusion studies went on to show that HmuO was indeed regulated by iron in DtxR in a manner very similar to how the tox gene was regulated.

However, the regulatory system for HmuO proved to be more complex than what was found at the tox promoter.

So in addition to regulation by DtxR and iron which was very similar to how the tox gene was regulated. We found an additional layer of regulation in that in order to see any appreciable expression of the HmuO gene, he heme source was required, either heme or hemoglobin.

Not only was there repression by DtxR but the promoter was also activated in the presence of a heme source. This we found to be very interesting and unusual since heme activated genes had not been previously identified in bacteria.

Additional studies to try to identify what were the factors involved in this heme activation went on to show that this heme activation was mediated by a two component signal transduction system in which one of the components of the system was involved in sensing heme at the cell surface and then transmitting this signal to a second protein located in the cytosol which then activated transcription of HmuO.

What I have shown here is pretty much a summary slide of the heme transport and heme regulatory network that we think goes on in C. diphtheria. The two component system I just mentioned is composed of a sensor kinase protein, which I have termed ChrS, which has at its end terminus a number of transmembrane regions and some loop regions that extend to the extracellular environment which we believe are involved in the binding or heme.

Upon binding heme we believe a signal is transmitted to the C. terminal portion of the protein which contains a hystine kinase which becomes phosphorylated on the binding of heme and then can transmit this phosphate group to the activator component ChrA which upon being phosphorylated will undergo a conformational change allowing it to bind upstream HmuO promoter and activating transcription.

Now, in high iron environments this promoter can still be repressed by DtxR. Optimal expression of HmuO would occur in the presence of the heme source, either heme or hemoglobin in a low-iron environment where DtxR is no longer acting as a repressor on this promoter.

Optimal levels of HmuO would then be predicted to be made under these conditions and then it would be able to act on any heme being transported to this heme transport system.

We now believe there is an alternate or second heme transport system in C. diphtheria since site directed mutations in the HmuT protein do not abolish the ability of C. diphtheria to transport mutalized hemes and iron source.

Regardless of which transport system is bringing in heme, HmuO would act on this heme coming into the cell breaking it down, releasing the iron, and making the iron available to the cell in order for it to grow in the low-iron environment of the respiratory track.

Some of my future aims are to identify this alternate heme transport system in C. diphtheria and to develop improved mutagenesis methods for C. diphtheria. There are very few molecular tools in this organism and the development of improved mutagenesis methods for the chromosome would greatly enhance our genetic analysis of this organism.

I also intend on pursuing structure-function cellular with the ChrS protein. This is the sensor kinase for the two component system to understand the mechanism by which it senses heme in the extracellular environment and how it transmits this signal to the activator component.

I would like to acknowledge some of the people who helped me in this work. Post-doc Sue Drazek, Craig Hammack, and Carrie Brickner who worked with me on this diphtheria project. Collaborators on this project include Angela Wilks at the University of Maryland and Shelly Payne, University of Texas, John Fulkerson who is currently a post-doc in my lab looking at iron transport systems in bacillus anthraces. Thank you.

DR. DAUM: Thank you, Dr. Schmitt. Downloading a lot of information very quickly.

Committee questions? Comments?

DR. KIM: I guess one question I have what is the current status of sequencing of genome of diphtheria?

DR. SCHMITT: That is actually an interesting question. It was in progress at the Sanger Institute but actually logging on to the website last night I discovered that they actually just completed the genome of diphtheria.

DR. DAUM: Who did that?

DR. SCHMITT: Sanger Institute.

DR. DAUM: Is that public domain kind of information?

DR. SCHMITT: Yes, it is. It's available.

DR. PALESE: Anthrax. How far is anthrax?

DR. SCHMITT: Very, very close. They are still filling gaps so it's not entirely complete yet.

DR. PALESE: Who does that?

DR. SCHMITT: That's Tiger.

DR. DAUM: Dr. Kohl.

DR. KOHL: I'll ask you a lead question that Dr. Vann already got. What would make your life more productive in your lab? What do you need that you don't have?

DR. SCHMITT: I'm in the process now of hiring a new post-doc for my lab. I think certainly once I get that person on board --

COURT REPORTER: Can you hear him?

DR. SCHMITT: I'm in the process now of hiring a new post-doc for my lab. That certainly will make life easier once that person is on board. Other than that --

DR. KOHL: Do you find within the constraints of the FDA that you can collaborate with people who you would like to collaborate with?

DR. SCHMITT: Right. Absolutely. I looked at a number of collaborators here. Certainly some of the important people in the field that I developed collaborations with that have been very productive.

DR. DAUM: Thank you very much.

Ms. Fisher, did you have a comment?

MS. FISHER: I don't know if it's specifically for you but any of the bioterrorism money -- the money to fight bioterrorism that Congress is appropriating, is any of that going to the FDA?

DR. SCHMITT: I believe so. I'm probably not the most appropriate person to comment on that.

DR. DAUM: Let's call on Dr. Goldman to answer that. I suspect we've already heard the answer.

DR. GOLDMAN: Yes, indeed, Dr. Fisher. In fact, the FDA has received $104 million to support bioterrorism. They got it only about a week ago.

DR. DAUM: Thank you, Dr. Goldman. Thank you, Ms. Fisher. I think at this point, thank you, Dr. Schmitt. That brings to a conclusion our open session. We thank very much the speakers and participants in it. I think we'll take a five-minute break to let the room clear and then we'll go into closed session and try and finish up.

(Whereupon, at 1:40 p.m. the open session was adjourned.)