FOOD AND DRUG ADMINISTRATION

P-R-O-C-E-E-D-I-N-G-S

DR. KARRON: I'd like to call this meeting to order if everyone would please take their seats. And I'd like to ask Ms. Christine Walsh to make some announcement.

MS. WALSH: Good morning. I'm Christine Walsh, the Executive Secretary for today's meeting of the Vaccines and Related Biological Products Advisory Committee. I would like to welcome all of you to this meeting of the advisory committee. Today and tomorrow's sessions will consist of presentations that are open to the public.

I would like to request that everyone please check your cell phones and pagers to make sure they are off or in the silent mode.

I would also like to request that any media inquiries be directed to either Heidi Rubello (phonetic) or Karen Reilly (phonetic) from FDA Office of Public Affairs, Karen and Heidi.

I would like to read into public record the conflict of interest statement for today's meeting. The Food and Drug Administration, FDA, is convening today's meeting of the Vaccines and Related Biological Products Advisory Committee under the authority of the Federal Advisory Committee Act, FACA, of 1972. With the exception of the industry representative, all participants of the committee are special government employees, SGEs, or regular federal employees from other agencies and are subject to the federal conflict of interest laws and regulations.

The following information on the status of this advisory committee's compliance with federal ethics and conflict of interest laws, including but not limited to, 18 U.S.C. 208 and 21 U.S.C. 355(n)(4) is being provided to participants in today's meeting and to the public. FDA has determined that all members of this advisory committee are in compliance with federal ethics and conflict of interest laws including but not limited to 18 U.S.C. 208 and 21 U.S.C. 355(n)(4). Under 18 U.S.C. 208, applicable to all government agencies, and 21 U.S.C. 355(n)(4), applicable to certain FDA committees, congress has authorized FDA to grant waivers to special government employees who have financial conflicts when it is determined that the agency's need for a particular individuals services outweighs his or her potential financial conflict of interest, Section 208, and where participation is necessary to afford essential expertise, Section 355.

Members and participants of the committee who are special government employees at today's meeting including special government employees appointed as temporary voting members have been screened for potential financial conflicts of interest of their own as well as those imputed to them including those of their employer, spouse or minor child related to Topic 1, Discussion and Recommendation on the Safety and Immunogenicity of an H5N1 Inactivated Influenza Vaccine sponsored by Sanofi Pasteur; Topic 2, Discussion of Pandemic Influenza Vaccine Strategies and Clinical Development of Pandemic Influenza Vaccines; for Topic 3, Discussion and Recommendations on the Selection of Strains to be Included in the Influenza Virus Vaccine for the 2007-2008 Season; for Topic 4, Discussion of Influenza B Strain Including the History of B Strain Circulating Lineages.

Financial interests may include investments, consulting, expert witness testimony, contracts, grants, CRADAs, teaching, speaking, writing, patents and royalties and primary employment. Today's agenda involves a discussion and recommendation of the safety and immunogenicity of an H5N1 inactivated influenza vaccine.

In accordance with 18 U.S.C. Section 208(b)(3), waivers were granted to Dr. Robert Couch, Dr. Lisa Jackson, Dr. Ruth Karron, Dr. John Modlin, and Dr. Robert Webster. Dr. Bruce Gellin and Dr. Wayne Hachey have been fully screened for conflicts of interest under usual procedures and have been advised that there are no financial conflicts of interest that would preclude participation or voting in this meeting or that might require a waiver under relevant statutes and regulations.

I note, however, that Dr. Gellin is involved in the process of pandemic vaccine procurement for the Office of Secretary of the Department of Health and Human Services in his capacity of Director of the National Vaccine Program Office. To avoid any perceptions of inappropriate influence in the actions of this committee, Dr. Gellin will not be voting on Topic 1. Dr. Hachey, who is director of Deployment, Medicine and Surveillance for the Department of Defense and whose office has responsibilities for procurement, likewise, will not be voting on Topic 1.

For the discussion of Topic 2 on Pandemic Influenza Vaccine Strategies and Clinical Development of Pandemic Influence Vaccines, Dr. John Treanor received a waiver under 18 U.S.C. Section 208(b)(3). Dr. Treanor will not participate in the discussion of Topic 1. For Topic 1, Dr. Treanor will serve as a guest speaker making a presentation. Dr. Treanor is Professor of Medicine, Infectious Diseases Unit, at the University of Rochester Medical Center. He will present data related to Topic 1 on behalf of NIH.

With regard to FDA's other guest speaker for Topic 3, the agency has determined that the information provided is essential. The following information is being made public to allow the audience to objectively evaluate any presentation and/or comments. Mr. Albert Thomas is employed as Director, Viral Manufacturing, Sanofi Pasteur in Swiftwater, Pennsylvania.

Dr. Seth Hetherington is serving as the industry representative acting on behalf of all related industry and is employed by Icagen, Inc. Industry representatives are not special government employees and do not vote. In addition, there may be regulated industry and other outside organization speakers making presentations. These speakers may have financial interests associated with their employer and with other regulated firms. The FDA asks, in the interest of fairness, that they address any current or previous financial involvement with any firm whose product they may wish to comment upon. These individuals were not screened by the FDA for conflict of interest. This conflict of interest statement will be available for review at the registration table.

We would like to remind members and participants that if the discussions involve any other products or firms not already not on the agenda for which an FDA participant has a personal or imputed financial interest, the participants need to exclude themselves from such involvement and their exclusion will be noted for the record. FDA encourages all other participants to advise the committee of any financial relationships that you may have with a sponsor, it's product and, if known, its direct competitors.

Thank you. Dr. Karron, that ends the conflict of interest statement, and I turn the meeting over to you.

DR. KARRON: Thank you, Christine. I'd like to welcome everybody to this VRBPAC meeting for what promises to be a very interesting two-day discussion on seasonal and pandemic influenza vaccines. I'd like to begin by going around the room and having all of the participants introduce themselves. I'll start with Dr. Modlin.

DR. MODLIN: This is John Modlin from Dartmouth Medical School.

DR. COUCH: Robert Couch, Baylor College of Medicine.

DR. FARLEY: Monica Farley, Emory University School of Medicine.

DR. SELF: Steve Self, Hutchinson Cancer Center.

DR. EICKHOFF: Ted Eickhoff, University of Colorado.

DR. WHARTON: Melinda Wharton, Centers for Disease Control and Prevention.

MS. KRIVACIC: Susan Krivacic, Patient Representative, Austin, Texas.

DR. HETHERINGTON: Seth Hetherington, Icagen, Inc., Research Triangle Park, North Carolina.

DR. WORD: Bonnie Word, Baylor College of Medicine.

DR. JACKSON: Lisa Jackson, Group Health Center for Health Studies.

DR. GELLIN: Bruce Gellin, Department of Health and Human Services.

MS. PROVINCE: Cindy Province, Acting Consumer Representative, Center for Bioethics and Culture.

DR. STAPLETON: Jack Stapleton, University of Iowa.

DR. HACHEY: Wayne Hachey, Department of Defense.

DR. WEBSTER: Robert Webster, St. Jude Children=s Research Hospital.

DR. McINNES: Pamela McInnes, National Institute of Dental and Craniofacial Research, National Institutes of Health.

DR. JAMES: Andrea James, FDA.

DR. BAYLOR: Norman Baylor, FDA.

DR. GOODMAN: Jesse Goodwin, FDA.

DR. KARRON: And I'm Ruth Karron from Johns Hopkins University. Our first speaker will be Dr. Norman Baylor from the FDA.

DR. BAYLOR: Good morning. I'm Norman Baylor. I'm the Director of the Office of Vaccines Research and Review at the FDA's Center for Biologics Evaluation and Research. Today I'm going to provide a brief overview, set the stage for today's meeting, in particular this session on our discussion of the BLA for Sanofi Pasteur's H5N1 vaccine.

Today we'll be presenting data in support of the first Biologics Licensed Application for a vaccine against H5N1 influenza viruses. This vaccine was manufactured by Sanofi Pasteur using the same manufacturing process as used for their licensed seasonal vaccine. The safety and immunogenicity data for the H5N1 strain were derived from a clinical trial completed by three National Institutes of Health Vaccine Treatment and Evaluation Centers.

As most of you know, there are currently no human vaccine licensed in the United States for avian influenza viruses such as H5N1. We at the FDA are working with our partners in the Government such as the National Institutes of Health, the Centers for Disease Control and the Department of Health and Human Services as well as the vaccine industry to facilitate the licensure of safe and effective vaccines for the use against potential pandemic influenza strains.

We're also trying to facilitate the evaluation of vaccines for potential use in the period prior to a pandemic including the potential uses for priming and cross-protection against evolving strains. And you will hear more about this in the discussion following this session.

We know that the risk of a pandemic is serious. H5N1 is present in large parts of Asia as well as now in the continent of Africa, Nigeria, Egypt. There is increased risk that more human cases will occur. The continuing presence and spread of the virus to new areas in poultry and wild birds increases the opportunities for human cases to occur. And we know that each additional human case provides this virus with an opportunity to improve its transmissability in humans and thus develop into a pandemic strain.

The timing and severity of the next pandemic we cannot predict. However, the probability that a pandemic will occur has increased and vaccines will be an important intervention against pandemic influenza and there are modeling studies that suggest that even a single dose of a vaccine of limited effectiveness may have significant effects early in a pandemic and reducing illness and spread of the virus.

I show this slide -- this is a cumulative number and I don't know if you can see this from the back, but the important thing is these two numbers. It's a cumulative number of confirmed human cases of avian influenza from H5N1 reported by the WHO last week. And as I said, the important thing here is as of February 19th, there were 274 cases. I believe there's 278 now. And of this 274, there have been 167 deaths which you can't see, but there are a variety of countries, as I mentioned before, Asia and the continent of Africa.

So as a background to the product we're looking at today, as I said before, this product uses the same manufacturing process as the licensed seasonal influenza manufactured by Sanofi. For U.S. licensed seasonal vaccines, no clinical data are required to substitute new strains into the vaccine such as we call a strain change.

The clinical data for Sanofi's H5N1 vaccine is designed primarily as a dose ranging study. And as a result, you'll note today that these data are limited. The immunogenicity was evaluated in the clinical studies. The proposed indication from the firm is for individuals 18 to 64 years of age for use during a pandemic or for those at high risk of exposure to H5N1. This vaccine will not be marketed commercially but is intended for the U.S. stockpile.

So in summary, we are bringing this vaccine to you today because we know the threat of an influenza pandemic is real and likely to continue. This vaccine that we're discussing today is intended to be an initial step to support preparedness and to facilitate a rapid early vaccine response. If licensed, this vaccine will become the first licensed vaccine available in the United States against an H5N1 strength.

The vaccine industry, in partnership with the Departments of Health and Human Services, is pursuing other approaches intended to elicit enhanced immune responses and require less antigen. And these are vaccines, for example, formulated with novel adjuvants which will not be the topic of our discussion today. We'll save that for another day.

If and when vaccines such as those formulated with novel antigens are found to be safe and effective, they're likely to supplant the use of the vaccine in discussion today. But we have to keep in mind that the benefit of having a licensed vaccine available now against a potential pandemic influenza strain must be weighed against the potential risk of having no vaccine at the time of a pandemic.

So that's my brief introduction and I'll be followed by Mr. Ken Guito from Sanofi Pasteur unless there are quick questions for clarification for me.

DR. KARRON: Thank you. Mr. Guito?

MR. GUITO: Thank you, Dr. Baylor. Dr. Karron, distinguished members of the advisory committee, ladies and gentlemen, good morning. I am Ken Guito and I represent the Strategic Project Office at Sanofi Pasteur. Sanofi Pasteur is pleased to the opportunity, along with our U.S. Government to present the first pandemic influenza vaccine for licensure, the H5N1 Influenza Vaccine, A/Vietnam 2004 (clade 1) 90 microgram formulation. Sanofi Pasteur views this formulation as an important first step which is based on time tested manufacturing technology, and we believe this represents unprecedented successful public-private partnership to prepare our nation for the threat of influenza pandemic.

As a recognized leader in influenza vaccine development and manufacturing, the U.S. Government collaborated with us to manufacture first generation H5N1 pandemic vaccines for clinical studies and stockpiling. Sanofi Pasteur has the only licensed U.S. manufacturing facility for inactivated influenza virus vaccines. We're also the largest manufacturer globally producing roughly half of the world's of influenza vaccine.

Our H5N1 vaccine development efforts have relied upon time-proven technology that have been licensed for inter-pandemic vaccine production for many years in the U.S. Sanofi Pasteur has extensive candidate vaccine efforts under development utilizing both traditional technology as well as novel cell-based production and adjuvant technologies. We are collaborating extensively with government agencies domestically and abroad.

Sanofi Pasteur's presence here today with the first pandemic vaccine applicant demonstrates our sense of urgency and our commitment to prepare for a possible pandemic event. We and other manufacturers continue on our efforts to develop additional strains of vaccine each and improvement on the last.

The H5N1 unit dose material used in a DMID clinical trial 04-063 was produced in 2004 under Health and human Services RFP award with Sanofi Pasteur functioning as a contract manufacturer. You'll hear more this morning on the DMID 04-063 trial from Dr. Treanor from the University of Rochester and from Dr. James from the FDA and more on the influenza pandemic RFP process from Dr. Robin Robinson from Health and Human Services in subsequent presentations.

As noted by Dr. Baylor, Sanofi Pasteur submitted a biologics license application for the H5N1 influenza virus vaccine in October 2006. In 2004 through 2005, Sanofi Pasteur produced U.S. Government stockpile doses of the same H5N1 clade 1 vaccine under subsequent HHS RFP awards. To date, in total, route 6 million 90 microgram-equivalent doses have been produced in the stockpile. It's important to note the majority of this vaccine is being held as a bulk formulation to allow longer shelf life and flexibility in subsequent formulation and in final packaging.

As Dr. Baylor and I have noted, the influenza virus vaccine, A/Vietnam 2001 (clade 1) 90 microgram formulation represents an important first in the response to influenza pandemic preparedness efforts. Sanofi Pasteur has a special responsibility and commitment to assist public health authorities in preparing for the possibility of a pandemic and to protect human health. We and other manufacturers, along with our Government collaborators, continue development efforts aimed at bringing forward subsequent more advanced candidate vaccines that will allow us to respond in the event of a pandemic emergency.

It is now my pleasure to introduce Dr. Robin Robinson, Acting Associate Director, Public Influenza, Health and Human Services, unless there are any clarifying questions. Okay. Thank you.

DR. ROBINSON: Good morning, distinguished panelists and guests. We are here today to discuss the H5N1 vaccines that the HHS has brought together for stockpiling. What I'd like to discuss briefly with you this morning is the department's and the nation's strategic plans and goals, our program portfolio matrix to carry out those and achieve those goals, the stockpile requirements for the H5N1 vaccines, the H5N1 vaccine production where we are today, and finally have a few summary remarks on the H5N1 vaccine being discussed this morning.

Why are we here today? Dr. Baylor has already alluded to that. In 1997, in Hong Kong, the city was hit with a poultry epidemic with high pathogenic avian influenza that wiped out many of the birds in the bird market and also crossed over into humans that were in contact with these infected birds. After cleansing and closure of these live bird markets, the epidemic was halted. However, in the winter of 2003 and 2004, H5N1 highly pathogenic avian influenza viruses re-emerged in water fowl and domestic poultry to cause an epidemic in Eastern Asia and also causing human deaths in Thailand and Vietnam.

In response to these events, the National Strategy for Pandemic Influenza was prepared and issued November 1, 2005. The President requested appropriations of $7.1 billion dollars of which $5.3 billion dollars has been appropriated thus far. In this strategy, the department and the nation communicated the needs for vaccine, antiviral and diagnostic research and development, stockpiling of antiviral and vaccines and the communication of other infrastructure building for the vaccine industry to address pandemic preparedness and response needs.

From that strategy, an implementation plan was prepared and issued in May of 2006. In this implementation plan, there are over 300 action items that the departments within the U.S. Government and the individual agencies within each department are responsible for implementing this pandemic preparedness and response actions. It provides guidance for each of these items and it defines the specific roles, responsibilities, metrics and timelines for accomplishing these action items. Further, it communicates to other non-federal entities including state and local governments, industry and even personal actions that can be taken for pandemic preparedness and response.

Also, within the pandemic strategy and implementation plan is, where possible, the use of licensed antiviral drugs and vaccines. From the strategy and implementation plan, there are a number of goals that have been enumerated. I draw your attention to two of these goals for vaccines. One is to establish and maintain a dynamic pre-pandemic influenza vaccine stockpile available for 20 billion persons in the critical workforce including first responders. Secondly, and built onto that, is to provide pandemic vaccines for all U.S. citizens within six months of the onset of a pandemic and, therefore, if a pandemic vaccine is two doses per person, that would mean that we need 600 million doses.

How did HHS try to accomplish and account for these goals? Well, we've developed an approach that was considered a program portfolio matrix, and I draw your attention to this approach for vaccines, antivirals and diagnostics and the areas of advanced development, stockpiling acquisitions and infrastructure building. Specifically, for this particular discussion, H5N1 vaccine stockpiles were established and developed in association with our sister agencies, the NIH, CDC, FDA and our industry partners that are U.S. licensed influenza vaccine manufacturers.

In 2004, we set out to establish these stockpiles giving industry the experience necessary to produce these vaccines at commercial scale, and we had a number of requirements to establish and maintain this stockpile. First, that it should be fore 20 million persons in the critical workforce including the first responders. Second, it would be for the usage at the onset of an H5N1 virus pandemic prior to the release of a well-matched pandemic vaccine. Thirdly, that this vaccine stockpiling manufacturing should be done without disrupting seasonal influenza vaccine manufacturing campaigns. Fourth, usage of apathogenic reassortants of high-risk virus strains as virus reference seeds were mandated for this manufacturing and that the manufacturing should be done at influenza vaccine sites, because these sites are the professionals at making influenza vaccine, and they use the commercial scale manufacturing process for the licensed inactivated split monovalent influenza vaccines. Therefore, as Dr. Baylor said, it would be a strain change for an antigen alone vaccine.

This vaccine, as already pointed out, is stored both as bulk and final container vaccine, and stability testing has been ongoing since September of 2004 when the first contracts were let. Further, by most of the vaccine being involved form, we're able to formulate the final container vaccine as antigen alone or with adjuvant as safety and immunogenistic cross protective data become available and warrant its usage.

The industry was given liability relief in the form of the PREP Act earlier this month. And finally, the goal of securing U.S. licensed vaccine product prior to usage was a mandate where possible.

So where are we today with this H5N1 vaccine stockpiling production? We see that we have two clades, clade 1 and clade 2, the clade 1 being the Vietnam strain 1203; the clade 2 being the Indonesian 0505 strain. I draw your attention that a dose for these calculations was based on 90 micrograms per dose and that a vaccine course is two doses per person. In a 2004 campaign, .47 million vaccine courses were produced by Sanofi Pasteur. In subsequent years, in 2005, multiple manufacturers were producing stockpiles. So in 2005, we had 8 million vaccine courses produced of clade 1 vaccine. In 2006, last year, we had 1 million clade 1 vaccine courses produced and an estimated amount of 4.8 million vaccine courses of clade 2 vaccine.

At this present, we have contracts for at least 1.6 million vaccine courses for this year. And there may be more produced later on in this fall.

So currently, for clade 1 vaccines, we have enough vaccine for 9.5 million persons. And clade 2, we have enough for probably 6.5 million depending on what the actual potency assay data has come out to be. That's an antigen preparation.

Finally, again, why are we here today? Well, one of the things is that today represents the cooperative leveraging of resources throughout HHS, the NIH, CDC, FDA and ASPR with industry to develop, manufacture and test an H5N1 vaccine candidate most similar to the U.S. licensed seasonal influenza vaccines. Also, today is a discussion of the first H5N1 vaccine candidate that could be licensed for immediate usage if an H5N1 pandemic emerges this year.

Thank you. Any questions? Otherwise, Dr. Linda Lambert from the NIH will share with you the important work that they've done on development of this vaccine.

DR. LAMBERT: Thank you so very much. I've been asked to give you a brief introduction to NIAID's pandemic vaccine research development efforts and then really to set the stage for Dr. John Treanor who will present results from the New England Journal of Medicine article and comment on safety data from some of our follow on studies.

The overall goal of the National Institutes of Health and the National Institute of Allergy and Infectious Diseases in particular is to serve the public health by conducting and supporting research on infectious and allergic diseases. And as you've heard Dr. Robinson previously indicate, we are all part of a broader Department of Health and Human Services pandemic influenza research plan.

For NIAID, that means research on controlling, preventing and treating seasonal and pandemic influenza. And at NIAID, we do that through a variety of different levels of research from assessing the basic biology of the virus to understanding the immunology and host response to characterize newly emerging influenza strains and understanding the molecular basis of virulence and transmission and to develop and clinically evaluate new diagnostics, drugs and vaccines and to coordinate and collaborate these efforts with other parts of the U.S. Government, most notably DHHS, NVPO, FDA, CDC and other public health service efforts, and finally, to generate information that will further inform ongoing global pandemic preparedness efforts.

So let me take you back in time. This map looks a little different from some of those that you are familiar seeing with. This is actually the map that is from late January 2004, and you heard Dr. Robinson allude to the outbreaks that were going on in Hong Kong in 1997. But in this map in January of 2004, we were dealing with yet another level of unprecedented outbreak. And so as of just a little over three years ago, there were outbreaks in humans in two countries and poultry outbreaks in a number of countries. And you know subsequently to this slide and over the last several years, that has expanded greatly. But in early 2004, this is what the map looked like.

So NIAID's response to that, that unprecedented level of outbreaks, both in human and poultry, was to obtain H5N1 vaccines for manufacturers with licensed products as quickly as possible. And in May of 2004, NIAID awarded a contract on behalf of DHHS to Aventis Pasteur, so Sanofi Pasteur, to produce a pilot scale lot of H5N1 using a scaled down manufacturing process that was as similar as possible to their licensed vaccines. And we asked for two formulations, 30 micrograms and 90 micrograms per mil.

So the goal of this -- there were many goals associated with obtaining this vaccine, certainly to gain experience overcoming both technical and logistical issues, and that was for the U.S. Government as well as the manufacturer, so to demonstrate the use of reverse genetics to generate an H5 vaccine reference virus and obtain select agent exemption from the U.S. Department of Agriculture; to produce reagents -- and this was done largely between Sanofi Pasteur and the FDA to generate the types of reagents that were needed to assess the potency of the vaccine; to develop assay capacity to be able to measure antibody responses to individuals who received the vaccine. And then, really, of all this set the stage for developing a framework and groundwork by which the companies could move to, if needed, commercial-scale manufacturing.

Other objectives -- clearly, to rapidly implement well-controlled Phase I and Phase II clinical trials; to obtain data on the safety and immunogenicity of the vaccine. And the goal for this was to provide initial data comparing dose ranging immune responses to form the basis of additional clinical trials and to assess multiple populations, so just not in health adults but also in the elderly and pediatric populations, and then support the development and use of an H5N1 hemagglutination HI assay and microneutralization assay and be able to have an infrastructure that supported rapid data analysis data collection.

So specifically now focused on Sanofi Pasteur, in June of 2004, NIAID provided a clade 1 H5N1 reference virus to Sanofi Pasteur, and that virus was an A/Vietnam 1203 2004 strain with a neuraminidase and genetically modified hemagglutinin gene and the remaining six genes from the PR8 virus. In March of 2005, Sanofi Pasteur delivered that vaccine to the NAIAD. In April of 2005, NIAID initiated the first H5N1 vaccine in healthy adults. And as you've heard Dr. Baylor say, that was done at three of our VTEU sites, and the study started in early April but was fully enrolled as of May 20th. And then NAIAD transferred preliminary and safety data sets for that study, 04-063, to Sanofi Pasteur for their BLA submission.

So at that point, I'd like to turn it over to Dr. John Treanor who will give you an update or a summary of the results of the adult study. That's NAIAD 04-063 that was published in the New England Journal and a brief overview of our follow on studies.

DR. TREANOR: Thanks, Linda. What I'm going to talk about then is the evaluation of the Sanofi subvirion vaccine made from the reverse genetically engineered virus, created it at St. Jude and put on the PR8 background that was done in health adults at three of NAIAD's VTEUs, our site at the University of Rochester, the University of Maryland led by a co-investigator, Jim Campbell, and the UCLA led by Ken Zangwill in collaboration with SRI which performed the immunologic assays and EMMES Corporation which did data management and statistical analysis.

Now this slide is an overview of the study design. You can see here where the vaccine was administered, the red triangles; where safety assessments were done; and where antibody sera were obtained. The study was done in a two stage design. Because this was the first human experience with the vaccine, approximately one-quarter of the subjects were enrolled in Stage 1 and were randomized to receive either placebo or vaccine at 90, 45, 15 or 7.5 micrograms. And in addition to assessing safety by memory aids and medical histories and follow-up visits, these subjects also had laboratory safety done before vaccination and on day seven including clinical chemistries, liver function and renal function tests and blood counts.

Now after assessment of the safety data, including a laboratory values, at day seven, the data were reviewed by a DSMB and based on that analysis, the remaining subjects were enrolled and randomized to receive vaccine or placebo. Similarly, the safety data seven days after the second dose were reviewed by the DSMB prior to Stage 2 subjects receiving the second dose.

After the day 56 or 28 days passed the second dose, the immunogenicity data were available from the Stage 1 subjects and based on all of the available safety date, the decision in terms of designing the protocols for follow on studies in the elderly as well as in pediatric populations were done. And in the elderly, we chose to look at 90 microgram and 45 microgram doses, and in pediatrics, at the 45 microgram dose. Subsequently, these subjects also received a booster dose at day 180 of the same vaccine that they had received initially.

Now in today's presentation, we're going to focus on what was published which is the safety and immunogenicity data that was available at day 56, that is 28 days after the second dose. Now this is what was published in the New England Journal. It includes all the safety data and immunogenicity data that had been collected up to 28 days after the second dose of vaccine. Just to remind you, the study was done in healthy adults aged 18 to 64 inclusive. It was a prospective, multicenter center, randomized and double blind clinical trial, and the interventions were two intramuscular doses separated by 28 days of either vaccine at 7.5, 15, 45 or 90 micrograms or placebo, and there were 50 placebo recipients and approximately 100 vaccine recipients in each group. The end points that were assessed for safety included both solicited and unsolicited AEs on memory aids and medical history that were done at follow-up visits, and as I mentioned in Stage 1, clinical laboratory tests, and two co-primary immunogenicity endpoints, the development of neutralizing antibody assessed in NDCK cells using a microneutralization technique and the development of hemagglutination inhibition antibody assessed using horse red blood cells, and both of these assays used the vaccine virus that are reversed genetically engineered virus on the PR8 background as to test antigen.

Now as a handy way of comparing the responses between doses which was the primary goal of this study, we dichotomized the results based on the proportion of subjects who achieved a titer of 1 to 40 or greater in these assays. And that 1 to 40 titer was chosen based on the experience with the neutralization assay in doing sero-epidemiologic studies in the 1997 Hong Kong outbreak as well as our expectations of what background levels of antibody might be in a population in the U.S. and historical experience with HAI data in assessing protection due to conventional influenza in the inter-pandemic period.

And so this was sort of a composite, but it's important to understand that this choice of a 1 to 40 endpoint is not validated in any way as an actual assessment of protection against H5 in humans. And in fact, it might be just as valid to choose a 1 to 20 or a 1 to 80 or a 1 to 10 endpoint. But it's really more as a convenient way in order to discriminate responses between groups.

Now this is the demographics of the enrolled subjects just to point out that there were approximately 100 subjects enrolled in each of the active dose groups groups and half as many subjects enrolled in the placebo group. The study population is predominantly Caucasian. About half the subjects are female. About 40 percent of the subjects had reported receiving conventional trivalent inactivated vaccine in the year prior to the study, and the age range was between 18 and 64 with a median just slightly less than 40 years of age.

Now as far as safety is concerned, the vaccine was well-tolerated at all doses that were tested. There was very clearly an increased rate of local pain and tenderness with the higher doses which was different from placebo. Those complaints of pain and tenderness were almost exclusively mild. There were no severe complaints of pain and tenderness. And this gives the results at the 90 microgram dose -- zero complaints of severe, 7 percent complaints of moderate pain or tenderness, and 53 percent of the subjects complaining of mild pain or tenderness at the injection site. I haven't shown the data, but the responses to dose two were almost identical.

There were no differences between any dose group and placebo in the rates of systemic side effects such as myalgias or fatigue or headache and there were no individuals who developed fever after either dose of vaccine. There was one serious adverse event which was a death which was not judged by the investigators or by the DSMB to be related to the vaccine which occurred within 56 days of dose one.

Now this is a representation of the neutralizing antibody on day 56, that is 28 days after the second dose of vaccine. It shows the reverse cumulative distribution of neutralizing antibody in each dose group. You can see here that the way this chart works is it chose the percentage of subjects in each dose group who achieved the indicated titer or greater so that you can see that as you increase the dose, there is clearly a more vigorous neutralizing antibody response. Using the 1 to 40 criteria that we had chosen, you can see that individuals who received the 90 microgram dose, which is shown in red, achieved a titer of 1 to 40 or greater 54 percent of the time with 95 percent confidence limits of 43 percent to 64 percent. You can see that the relative superiority of the 90 microgram dose holds true no matter what cut point of titer you chose to analyze. It's also true the 90 microgram recipients achieved a titer of 1 to 20 more frequently and achieved a titer of 1 to 80 more frequently compared to the other dose groups.

Very similar results are seen when the sera are assessed using the HAI assay with horse red blood cells. Again, you can see that 58 percent of the subjects achieved a title of 1 to 40 with 95 percent confidence limits of 47 to 67 percent. Again, there is a very clear dose response relationship in the immune response with subjects who received a 90 microgram dose of showing more vigorous and higher titered antibody responses than those who received lower doses.

Now as you know, after the study was published in March 2006, there were further discussions with FDA and based on a guidance document which was published in March and further discussions with the agency in April and later in 2006, there was a recommendation for a change in the analysis of the data. And the two changes are that the hemagglutination inhibition test became the primary focus of the immunogenicity analysis based on increased confidence of the accuracy of the HAI test using horse red blood cells, which was a relatively new development and a recommendation that we redefine the value assigned for the first dilution that was tested from 1 to 20 to 1 to 10. An HAI sera response was then redefined with consultation as requiring both a fourfold increase over baseline and achieving a titer of 1 to 40 or greater, again, redefining the titers as calling the first dilution tested 1 to 10 rather than 1 to 20.

It's important to note that this re-analysis involves recalculations using the 1 to 10 definition of the starting dilution but does not involve any retesting of the sera. It's simply a recalculation. And to show you what this does, this is the data as published. It's a reverse cumulative distribution curve of the HAI data 28 days after dose two. And you can see that the first dilution tested is defined as 1 to 20 so that subjects that showed no HAI activity at the first dilution are assigned a value of 1 to 10 or less. That is why 100 percent of the subjects have a value of at least 1 to 10 or less.

If we redefine the starting dilution as 1 to 10, you can see that this does not change the shape of the curve but does change the values assigned to the x axis. And if we use a criteria of achieving a titer of 1 to 40 or greater, this changes that estimate to 44 percent with 95 percent confidence limits between 34 and 55 percent. So just to show you, this does not change any of the data but simply changes the way the x axis is defined and the calculation of whether we're looking at this point or this point for dichotomizing the results.

Now as you know, there have been further studies of the Sanofi vaccine. This is just an overview of what other experience exists specifically with the 90 microgram dose. The following number of subjects have received the 90 microgram dose in randomized trials which have included doses of 1, a second dose a third dose. These are the numbers -- the subjects who have received 1, 2 or 3 doses of 90 micrograms.

In addition, there have been open label studies, one of which was a study looking at revaccination of people who had been in a prior H5 study back in `1998. That involves 37 subjects that we're going to talk about this afternoon. In addition, the vaccine has been given as a 2 times 90 microgram dose to a number of workers involved in making the vaccine at Sanofi as well as laboratory workers at St. Jude's, and you can see the total numbers of subjects who have received vaccines in those open label studies. There have been 363 individuals who received at least 1 dose of 90 micrograms, 304 who have received 2 doses and 166 individuals who have received a third dose of vaccine.

Now in the open label studies, which include the use of the vaccine in manufacturing workers as well as laboratory workers, there have been no serious adverse events related to the vaccine to date, and the rates of local and systemic solicited adverse events are very similar to what had been seen in the control trial at 90 micrograms in health adults in Protocol 04-063.

The controlled evaluation in the elderly is not finished yet, and so the database has not been locked. There have been 259 elderly subjects enrolled in that study and randomized to receive either 90 or 45 micrograms or a placebo at a 2 to 2 to 1 or a 2 to 2 to 1 ration.

In addition, I'll mention that a subsequent study has also been done in children 2 to 9 years of age. This study only evaluates the 45 microgram dose. Neither database is locked and so only aggregate analysis is available, but no vaccine related serious adverse events have been reported. The local and system reactogenicity has mostly been reported as mild or moderate and appears to be very consistent with the observations in the study in adults.

So with that, I'll end. I'd be happy to answer any questions, or we could do questions at the end. Okay.

DR. JAMES: Good morning. My name is Andrea James, and I'm a Medical Officer in the Division of Vaccine and Related Product Applications. This morning I'll be presenting the results of the FDA analyses of the immunogenicity and safety data as submitted in the Sanofi Pasteur's H5N1 vaccine BLA.

This slide outlines my discussion points. First, I will give a summary of the product. Following that, I will describe the clinical study supporting this BLA, FUG01, and then discuss the immunogenicity and safety results of the study. I will end my presentation by summarizing the BLA, discussing the limitations of the data and posting the FDA questions to the committee.

The BLA was submitted on October 17, 2006. The product under review is H5N1 influenza virus vaccine A/Vietnam/1203/2004/ Clade 1. The proposed dosage is 90 micrograms, and the proposed administration is 2 one-milliliter IM injections administered 28 days apart.

Sanofi proposes the following indication: H5N1 influenza virus vaccine A/Vietnam/1203/2004/Clade 1, 90 micrographs per milliliter is an influenza viral vaccine indicated for active immunization against influenza disease caused by H5N1, A/Vietnam/1203/2004/Clade 1 influenza virus and primary vaccination of healthy adults 18 through 64 years of age.

FUG01 was the single study submitted in support of this BLA. FUG01 is a Phase I/II randomized, double-blind, two-stage, placebo-controlled, dose ranging study. Subjects were eligible for the study if they were healthy and between the ages of 18 and 64 years with extremes included. Subjects were stratified by age and prior seasonal influenza vaccine receipt and then randomized in a 1:2:2:2:2 fashion to 1 of 5 doses, either saline placebo or 7.5 micrograms, 15 micrograms, 45 micrograms, or 90 micrograms of vaccine. Subjects then received their randomly assigned dose as two intramuscular injections administered 28 days apart.

The study objectives were as follows: One, to determine the dose-related safety of subvirion inactivated H5N1 vaccine in health adults; two, to determine the dose-related immunogenicity of subvirion inactivated H5N1 vaccine in health adults approximately 1 month following receipt of 2 doses of vaccine; and three, to provide information for the selection of the best dose levels for further studies.

In FUG01, the investigators looked at three co-primary immunogenicity endpoints. Two of the endpoints dealt with neutralizing antibody and these data were not submitted to the BLA as per a prior FDA applicant agreement. The BLA submission included data for the following endpoint analyses: Fourfold rise in HAI antibody titer and HAI antibody greater than or equal to 1 to 40, both measured at 28 days after each dose of vaccine and 6 months after the receipt of the first dose of vaccine.

Of note, the first and last time points are of interest. However, traditionally, HAI titers 28 days post the last dose in a vaccine series is the data usually requested and analyzed in the FDA licensing process.

FUG01 was designed as an exploratory study, so all of the results I'm about to present and to be received with the following information in the forefront of your mind. This study was not statistically powered to provide estimates of immunogenicity at any specific dose. And the study was also not powered to detect rare safety events. Therefore, the results only provide trends.

In terms of subject demographics and baseline characteristics, a total of 452 subjects were enrolled in the study. The majority of subjects were Caucasian female with a mean age of 40.5 years with a range of 18.1 to 64.9 years. The majority of subjects, 58.4 percent had not received the 2004-2005 seasonal influenza vaccine. And interestingly, 3.3 percent of all subjects had detectable H5 antibody at baseline.

Now to go on to the immunogenicity results. On the slide we're looking is a tabular presentation of the first endpoint of percent of subjects who achieved a fourfold rise in HAI titer. In a moment, I will show this data in graph form. However, you can see in this table that the 90 microgram group with 91 subjects in the per protocol population had approximately a 23 percent response rate 28 days after the first vaccination, and a 95 percent confidence interval ranges from 14.9 to 33.1, and a 45 percent response rate 28 days after the second vaccination with a 95 percent confidence interval ranging 34.6 to 55.8 with waning of this response by six months post vaccination two.

This is a graphical presentation of what you just saw on the table. I'll take a moment to orient you to the slide. On the x axis, we have time in days, and on the y axis, we have percent of responders. The blue diamonds represent the placebo arm while the red squares represent the to be licensed 90 microgram dose group with their respective 95 percent confidence interval bars in their respective colors.

There are four distinct time points plotted for each of the study arms: baseline; 28 days after receipt of the first vaccination; 28 days after receipt of the second vaccination; and 6 months after the receipt of the second vaccination. Please note that the dose groups are slightly separated in time on the graph but that the separation is for graph clarity only.

All subjects were evaluated at the same study time points. This purple hatch mark represents the 40 percent response rate threshold that the FDA currently recommends in the draft guidance document on clinical data needed to support the licensure of pandemic influenza vaccines. It is important to note that neither Sanofi, the BLA applicant, nor NIH, the IND sponsors, were privy to the recommendations held within this guidance, because this guidance was not available until March of 2006, which was nearly a year after FUG01 was conducted.

You can see in this graph that the 90 microgram group is trending, at least the lower bound of the 95 percent confidence interval, is trending towards meeting the criteria, the 40 percent response rate criteria 28 days after the second vaccination. However, it falls shy of the lower bound threshold by about 5 percent which may be at least partially due to a small study sample size.

This is a graphical presentation of what you just -- actually, this is a dose response graph which I'm putting up to show two things: one, that at all of the vaccine doses tested, there is a dose response as you can see here. And then the second thing that I want to show is as you increase the vaccine dose, you see a dose-dependent increase in fourfold titer rise. So there does appear to be a dose-dependent response.

On this slide we're looking now at a tabular presentation of the second endpoint of proportion of subjects who achieved an HAI titer greater than or equal to 1 to 40. In a moment, I'll show this data in graph form. The numbers are very similar to the numbers that you saw for the fourfold rise. You can see in this table that the 90 microgram group had approximately a 24 percent response rate at 28 days after the first vaccination and a 46 percent response rate 28 days after the second vaccination. And again, we see waning at 6 months post vaccination 1.

This graph is very similar to the one I just showed you for fourfold rise. Again, in orienting you to the graph, we have time and days on the x axis and percent responders on the y axis. Once again, placebo is represented by the blue diamonds and the 90 microgram group is represented by the red squares with the respective 95 percent confidence interval bars in their respective colors. Once again, the four time points are graphed here, and we have baseline; we 28 days post vaccination 1; we have 28 days post vaccination 2; and we have 6 months post vaccination 2.

Once again, the points are separated in time slightly on the graph just for graph clarity. Up here you'll see this purple hatch mark, once again at the 70 percent mark. And this, again, is the FDA recommended or requested threshold for HAI titer greater than or equal to 1 to 40. And once again, this is recommended as of March 2006 in the draft guidance.

Once again, you can see that at the to be licensed dose 90 microgram, this group is trending upward. However, it falls well short of the 70 percent threshold that FDA is now currently recommending.

In addition, to the pre-specified endpoint analyses, I performed additional analyses of the following subgroups: gender, race and ethnicity, and the pre-specified strata of age and prior influenza vaccine. Of course, the ends are small but if you look at this per protocol gender subgroup analysis of the 90 microgram dose, you will see that 56 percent of females had a fourfold in HAI titer compared to just 46 percent of males in the study.

Moving on to race and ethnicity. Here the ends for most of the groups are even smaller. You can see that in the race groups, the percent of responders in terms of fourfold increase in HAI titer were fairly equal across the different races. However, if you look at ethnicity, Hispanics appear to respond at a higher rate.

In this slide, I'm presenting the pre-specified strata of age and prior seasonal influenza vaccine, and if you -- the thing, I guess, that jumps out very quickly at you is that the younger group, less than 40-year-old subjects who had not previously had the 2004-2005 influenza vaccine appear to have a higher response rate in terms of fourfold rise in HIA titer. And this is as compared to their counterpart who did receive prior vaccination. So you're looking at a 75 percent response rate versus a 37.5 response rate, again, noting that the ends are very small.

However, if you look at the group who is 40 or greater in age, you see pretty much the exact opposite where this group did much better, if they received the prior seasonal influenza vaccine versus not having received the prior influenza vaccine; and again, I must stress that the ends are small here and that this is a subgroup analysis.

So in summary, the immunogenicity results suggest that this H5N1 vaccine appears to have a dose-related immune response. And of all the doses studied, the highest dose, 90 micrograms, appears to have a higher response rate with approximately 45 percent of subjects responding after two doses of vaccine. However, immunogenicity observed in the study is less than what is usually seen in seasonal influenza vaccine studies, and the impact of gender, ethnicity and prior seasonal vaccination on H5 immunogenicity is unclear and may warrant further exploration.

On to the safety results. Safety was assessed by frequency and incidence of immediate reactions occurring 15 to 30 minutes post vaccination, solicited local injection site and systemic reactions measured a day 0 through day 7 vaccination and unsolicited AEs and SAEs measured at day 0 through day 56 of the study. Solicited injection site AEs included pain, tenderness, redness and swelling, and solicited systemic AEs included feverishness, malaise, body aches exclusive of the injection site, nausea and headache.

There were four SAEs in the study, none of which were considered vaccine-related. There was one death in the 45 microgram arm, and this subject was a 52-year-old male with a history of chronic alcoholism, and his death was considered secondary to sequelae of his chronic alcoholism. There were three other SAEs, a breast cancer in the placebo arm, menorrhagia in the 15 microgram arm, and a cerebrovascular accident in the 90 microgram arm; again, none of these considered vaccine related.

If we look at local reactogenicity events, there appears to be a dose-dependent increase in the frequency of injection site reactions with the 90 microgram group having the most with approximately 85 percent of subjects experiencing at least 1 injection site reaction. The majority of these injection site reactions in this group were pain and tenderness and approximately 14 percent of subjects had injection site reactions that were considered of moderate intensity.

When we look at systemic events, we see that overall they were a lot less common than injection site reactions and that system events did not appear to be dose related. In looking at the specific AEs, you see that the most common in the 90 microgram group was headache at 38 percent and malaise at about 30 percent. However, the rates for a systemic injection -- or systemic events were similar across all dose arms.

So in summary, the safety results suggest that there is a dose-dependent increase in frequency of local reactogenicity events with the majority of events being pain and tenderness occurring in the 90 microgram group. And these data reveal no other apparent safety signals.

So to summarize, Sanofi has submitted an application seeking licensure for their biologic product, H5N1 Influenza Virus Vaccine A/Vietnam/1203/2004 (clade 1) at a recommended dose of 90 micrograms to be administered as two 1- milliliter intramuscular injections 28 days apart. Based on the data submitted with the BLA, it appears as though the two 90 microgram doses provide a higher immune response. However, the immunogenicity observed in study FUG01 is less than what is usually seen in seasonal vaccine studies with approximately 45 percent of subjects responding after two doses of vaccine.

Again, there are no apparent safety issues. Unfortunately, there are many limitations of these data contained in the BLA. Therefore, our ability to make firm conclusions about the data are limited. First, the clinical database is small, and as such, is not statistically powered to detect rare adverse events and is not statistically powered to produce statistically significant results. And in fact, these results can only provide trends.

Additionally, the clinical efficacy of this vaccine is unknown. A correlative protection against H5 is unknown. And the impact of gender, ethnicity and prior seasonal influenza vaccination on the immune response to this H5N1 vaccine is unknown.

With that, I will move on to give you a brief look at the questions to the committee reminding you that Sanofi's proposed indication is that their vaccine will be indicated for active immunization against influenza disease caused by H5N1 A/Vietnam/1203/2004 (clade 1) influenza virus and that primary vaccination of healthy adults 18 through 64 years of age --

The questions we will be discussing later on today and presenting to the committee are: Are the data sufficient to support the effectiveness of this product for use during a pandemic or in situations of potential high risk exposure; are the data sufficient to support the safety of this product for use during a pandemic or in situations of potential high risk exposure; and lastly, please comment on studies to collect additional information about the effectiveness and safety following this vaccine's use. The questions will be presented again later on, prior to our discussion.

Before I end, I'd just like to acknowledge all of the people who helped me in developing this presentation. I specifically would like to give great thanks to Dr. Tammy Massey, Dr. Zhiping Ye, Dr. Melissa Baylor, Dr. Antonia Gerber and Dr. Joe Toerner whose time and resources and knowledge and expertise made this presentation possible. Thank you.

DR. KARRON: Thank you, Dr. James. At this point, we'll take questions for Dr. James or for any of the previous presenters. Dr. Couch?

DR. COUCH: Most of my questions are procedural. I guess I'm directing them to Dr. Baylor maybe. But I need a little -- maybe some of the other committee members -- a little better understanding of the role of the FDA and maybe of this committee in licensing a vaccine like this. You know, we've said and many of us have earlier understood that this would represent a strain change. You see? And yet we're considering a licensing application because we wouldn't license the strains we're about to select for next year. But on the other hand, if we license H5, do we also have to license H7, H2 if those come down the line? And where do we stand with regard to considering individual vaccines that are using, as you pointed out, a pre-existing approved procedure for preparation?

DR. BAYLOR: I can start out answering that, Bob. I mean, you know, this is sort of new ground here. And the -- the procedure is basically -- I mean for this vaccine it's -- we're saying it's manufactured by the same process as the currently licensed vaccine. And so in some sense it's a strain change, but you have to keep in mind that we at least need some dose ranging studies. So we need to figure out what the dose is for this vaccine and, therefore, we have a clinical study which has, you know, gone down that road to try to do that. And so we're -- also, this vaccine will be labeled with a different name to differentiate it from the current seasonal vaccine, so we're calling this an application.

Now if we -- let's start with something like a new clade. Well, how would we handle a new clade? So it's a H5. That's more like a strain change supplement, like, for instance, tomorrow when you decide on what the strains will be for next season's vaccine. However, for the H5, since we have very little experience with that, we may require clinical data for the next clade. And in fact, we know that some studies are being done with the H5N1, Indonesia. So that would be -- that would come in with additional more supportive clinical data as far as looking at the dose, because we -- we just wouldn't be able to predict that.

Now if you move into NH7 or that H7, if it was manufactured by a licensed procedure, it would follow the same process. But we still would need some kind of -- and again, same process in the sense that we would need some kind of supportive clinical data to, at a minimum, determine the clinical -- the dose required. And so I think that sort of addresses your question how we would do that.

DR. COUCH: Yes. I think it does, but I think you would agree then in the process of doing this, then we're not literally looking at a brand new vaccine proposal. For example, with regard to something like this, you see, this is an established procedure and the H5 has made itself into a green monster disease wise, but I don't think the virus knows that. And we've changed the hemagglutinin up on the top. You see? So there's an absolutely safety data for this procedure and for other vaccines that ought to be, I would say, considered from that point of view.

Now from the point of view of looking at the dose and things like that you see, you can see individual considerations for each one that comes forward. But it would -- we would not want it to be considered a brand new virus starting from scratch to look at everything. I guess that was part of my question.

DR. BAYLOR: And you're correct and we agree. I mean you -- we're not saying we're going to bring -- every time we do this, we're going to bring one to you. But I mean this is the first and so we believe it's really important to have this discussion, have you look at the data, although the data are limited. But this is -- and I don't think we should get sort of wrapped up in what we call this thing as far as the submission. I mean it's not a brand new product as for example we came in and changed the manufacturing process completely or we had an adjuvanted vaccine. That would be brand new product. But -- so what we're -- what I'm saying is this is -- don't get confused by what we're calling this. You know, this is a first of its kind and we're bringing it to you with the limited data for the reasons that we explained earlier.

DR. COUCH: This one is just -- one more and I'll quit -- minor. And then with an licensed approval for this, does that -- what kind of freedom does Sanofi have with that? I mean, for example, most of us would say if we could hang a shingle out on the streets that we have a bird flu vaccine for sale, we'd get rich in a big hurry. Now that would be politically unwise for them, but what sort of freedom does this give them?

DR. BAYLOR: Well, I can let Sanofi respond to this, but I mean we all presented in our slides, or I did and I believe Sanofi did as well, this vaccine will not be commercialized. It will be for the stockpile, and Dr. Robinson has stated that as well.

DR. COUCH: The license will be for the stockpile, specified that way?

DR. BAYLOR: Well, that's a little -- that's -- you know, we have to make those decisions, but this vaccine -- if we license this vaccine, it will be licensed, but it will be licensed for what it is.

DR. KARRON: Dr. Webster?

DR. WEBSTER: We've heard that this is not a new vaccine, but indeed, it is a new vaccine being made by totally new strategies, by reverse genetics, and this is really a very historical event when we're faced with the use of reverse genetics virus to make a vaccine and then provide that vaccine to humans.

And it's a genetically modified organism that you're talking about putting in human. This was mentioned in passing, stress where we made issues that come from the use of a reverse genetics. I can get past, is this the reason for the poor immunogenicity in this thing? Is this why it produces that poor amount of hemagglutinin? These are all scientific messages that are out that, but my point is that this is a whole new strategy we're using to make this vaccine. And we have to have that on the table as we think about it.

I think that the use of such a vaccine is the roadmap to the future. We've been using reverse genetics within the States over many years. And now we make these viruses by reverse genetics exactly as we need them, and this procedure has shown that these vaccines are genetically tainted. The question that was raised earlier is if you use reverse genetic process on this highly pathogenic virus, will it be safe for manufacture, will the manufacturers be safe. And I think that these are issues that every worker in immunogenicity, I would have nothing to do with this vaccine. I will conclude B I don=t know whether it's necessary, but I just wanted say that.

DR. KARRON: Dr. McInnes.

DR. McINNES: Rob, I want to clarify one thing. I want to be sure that you did not state that genetically modified organism is being put into people. At one point, that was where I thought you were heading, and I want you to please clarify that.

DR. WEBSTER: (Inoperative microphone)

MS. WALSH: Excuse me. May I interrupt? I'm sorry. I was just told that your microphone is not working, so if you could use Dr. McInnes=? Thank you very much. We appreciate that.

DR. WEBSTER: The light was working. Sorry about that. The genetically modified aspects of this organism, yes, a genetically modified organism was made. It was inactivated and made into vaccine which we've heard this morning, so it was a genetically modified organism that we began with.

DR. KARRON: Dr. Modlin?

DR. MODLIN: I have a couple of unrelated questions. I guess the first is for Sanofi, and that is what are the plans for extending the age range for approval for this vaccine to children and to the elderly? Obviously, we have studies under way, but I'd be real curious as to what the thinking is with respect to the timeline for bringing forward what I assume would be a supplement.

MR. GUITO: So keeping in mind that the discussions around this license application started roughly a year ago, the data that was available at that time was the data in 18 to 64 year olds. There were subsequent trials done with the NIH in the pediatric population and in the elderly population. That data has only recently become available. Dr. Treanor and Dr. Lambert are ready to discuss that data today. I think when we reach conclusion on this issue with the 18 to 64 year old indication with the FDA, we will then initiate discussions about broadening that population.

DR. MODLIN: Maybe I could ask Bruce Gelling or some of the others that have been actively involved in these discussions what might happen in terms of use of this vaccine if it were stockpiled and we have a -- we're faced with a clade 2 epidemic?

DR. GELLIN: Well, I mean we started the process -- I think in Robin's slide -- you may want to address some of this -- in 2004, and the goal was to have vaccines in the stockpile that would be relevant to what was circulating at the time and this has begun to move forward. We don't know whether or not a vaccine like this would provide some, any, much protection and I think the idea is that since it could provide some, I think the concept is that in the setting with an imminent pandemic, you would begin to use what you had available.

There will be discussions later today in the second session about how other -- how vaccines might be used more in a different way and regarding immunologic priming. But I think that right now the idea is that you'd use the vaccine that you had and hope that it provides some protection. And this is the sort of a stopgap as you begin to make the vaccine against the pandemic.

DR. MODLIN: But that would be the case even though the label would say this is indicated for use in the event of a clade 1 epidemic?

DR. GELLIN: I guess there is the -- you know, given that labeling, I guess I'll ask others to respond to that, because, again, we don't know. We do know that with other vaccines when there is a mismatch, there is some protection. So I think that the idea would be that you could get some but not perfect protection, but maybe FDA would like to respond to that.

DR. KARRON: Dr. Couch?

DR. COUCH: Perhaps. I just wanted to add a comment sometime -- this may be appropriate -- that we don't really know, as Dr. James said, what is required to predict against an H5 pandemic strain any more than we'd know about H7. And so when we're looking at the criteria that she showed us, those, our European colleagues have perpetuated those fairly extensively, but we haven't used them much in this country. But those are frames of references that way I think of them when you're talking about H5, for what kind of immune responses you're getting, they cannot be used, I think most of us agree, as a criteria for an approval based on some idea about protection. We just simply don't know what we need, and I'm one of the views that anything is better than nothing which then relates a little bit to Bruce's question, and that should be what we have in mind when we decide to approve a vaccine, not where how close it came to the lines that Dr. James showed us.

DR. KARRON: Dr. Jackson and then Dr. Farley.

DR. JACKSON: Well, Dr. James presented some information on fourfold response by age and prior vaccine stratification, and those data, while limited, suggest potentially important interactions in vaccine response by age and possibly by prior receipt of seasonal influenza vaccine. And so it seems relevant to know more about that. While the study was conducted among persons 18 to 64, that does not necessarily mean that there was homogeneity of response or dose response across that entire age range.

So I wondered if there was additional information available on one, the distribution of age among the groups less than 40 or greater than/equal to 40, specifically interested in the proportion of individuals in the higher end of that age range; if there is information on the RDC curves to give estimates of both effective age as well as whether dose response actually varies by age; and then whether safety data has been evaluated by strategies of age and/or prior vaccine response?

DR. JAMES: In terms of your first question, what I've looked at were the stratification, so as I presented the stratification of age and prior influenza vaccine, I do not have currently have information on those particular strata. But I did look at safety data based on gender and based on age, and there -- again, the data are limited. There are no apparent signals with those.

Can you repeat -- you asked me another question on --

DR. JACKSON: Yes. Thank you. You presented the fourfold rise data. I wondered if the response to achieving a titer of 1 to 40 are greater, specifically the RDC curves, if there were any information on the relationship of age and possibly vaccine receipt on those other measures of the vaccine response and dose response?

DR. JAMES: Okay. In terms of the stratification, I did look at -- I didn't look at all of the doses, but I did look at the 45 microgram dose for the stratified groups and the results were similar to what was shown for the 90 microgram group. I didn't look at the 15 or the 7.5 microgram group. And I need to answer another question for you I think.

DR. JACKSON: No. I think that's it. Just an interpretation of the data, I mean the data are consistent although not -- they do not prove that the dose response and the evidence for some response are actually restricted to a particular subgroup which is the less than 40 with no prior vaccine receipt, and I think that's important considering the implications for the overall results.

DR. KARRON: Dr. Farley?

DR. FARLEY: Well, I guess I'm struggling a little bit with the guidance that has now been published which was after the fact, so the March 2006 guideline, they're not binding but suggestions for parameters of immunogenicity. And while I understand we're in a situation of wanting to be ready in responding, how will this impact -- I mean, those in general, that wasn't met, the guidance was not met with this vaccine in terms of immunogenicity which may be okay if it's better than nothing, you know, in an urgent situation. But will we -- will this be modified over time? Are we going to expect more with each additional or each further refinement of these vaccines as they go along?

Or, you know, it's a struggle here to say it didn't really meet -- it isn't all that immunogenic if we are -- if this reflects anything close to correlates of protection and we don't know that. But I guess I'm struggling between urgency and needing to have something available versus sort of where -- how low to set the bar for immunogenicity.

DR. KARRON: Dr. Goodman?

DR. GOODMAN: Well, I was going to comment anyhow and follow-up on what Dr. Couch said which is, I think, relevant. And he might want to comment. These guidances, just like the European criteria, are set as, in this case, as a goal, as this is something that would be desirable. As Bob Webster said, this H5 is poorly immunogenic. Also, as Dr. Couch said, and he's written extensively about it, what you see with these levels of hemagglutinating antibody is basically the higher the levels are, they correlate in a population with more protection.

However, that does not mean at levels lower than this, in many circumstance, there is not substantial protection. So there's no a perfect correlate mapped out. We know at least from seasonal influenza that levels lower than 1 to 40 can have a protective affect, and as Norman mentioned, and I'll mention later this afternoon, some of this in modeling also plays out as showing a beneficial affect.

So I think the guidance was intended to set goals. The better an antibody responds, the better. We're all hopeful that new technologies will achieve a better antibody response with this antigen. But right now, in terms of a vaccine with a safety profile that is well-established and could be acceptable in broad use this is where we're at.

DR. KARRON: Actually, just a comment that I wanted to make in response to that, and I'd ask other influenza experts around the table to comment, you did say, Jesse, that in general, higher titers of antibody correlate with increased protection. That's true, we think, for seasonal influenza. I don't think we have those data for pandemic influenza, and if anyone wants to correct me, please do.

DR. GOODMAN: Yes. Well, I think we should go around and ask people, but I think we -- there are not a lot of reasons to think that, you know, pandemic may be more like in children, for example, where you don't have a history of chronic exposure to other antigens. But I think all we can say is that we know from in annual influenza, that there's a correlate. And you're correct, we don't know with pandemic that there is or exactly what it is or that the curve would follow the same level.

DR. COUCH: I think we know --

DR. GOODMAN: There's reason --

DR. COUCH: -- in a general sense.

DR. GOODMAN: Well, I was going to say there's no reason to think not.

DR. COUCH: Well, actually, in 1957 says that indeed, if you've got a vaccine response to that antibody -- Ted can comment on this -- you were protected. Now, can you -- is there nice quantitative, correlated data with all of these titers like we tend to look at now? I can't remember any if there was. But it was pretty clear that a vaccine response induced protection. It was actually less clear in '68, but it was also there. So I think we can still use that generality even if we can't take a titer and put numbers and percentages on. Ted, you may want to comment on that.

DR. EICKHOFF: Yes, I agree, but the amount of H2 vaccine produced in 1957 was really very limited, and so those studies are very limited. However, certainly for seasonal flu, it's been amply confirmed time after time after time that higher HAI levels correlate with protection

If I may, may I ask another question? Two questions as a matter of fact. First one to either Dr. Treanor or Dr. James. I'm interested in the thinking that led to the recalculation of the results. What was accomplished here? You set the bar higher, obviously, made it a more stringent test. What was the thinking that led to this?

DR. COUCH: Could I comment on that because I understand. It was a very simple error as I understand. Well, maybe I shouldn't call it an error, just doing things in a different way.

DR. BAYLOR: It was -- I mean what we used was normal convention, and I think that the purpose -- you know, NIH was -- and NIH and John can speak as well -- but they were looking at microneuts. and HAI and so it was a different purpose in how they were calculating -- how they -- the convention they were using for the assays. But we used what was normally considered the standard convention. And so, I mean, there's no magic here or any -- you know, I don't want to dwell on it.

DR. EICKHOFF: I understand. Second question -- perhaps Bruce might comment on this -- but what would be the trigger for a use of this product?

DR. KARRON: Actually, before that --

DR. TREANOR: Just so people are clear about the difference between 1 to 10 and 1 to 20, the way these tests are done is that the sera is diluted to 1 to 10, that's 2.5 microliters of serum in a volume of 25 microliters of buffer or RDE. So that's a 1 to 10 solution. Then serial dilutions of that are made. An equal volume of virus is then added, and that is the reaction in which antibody and virus interact. So depending on your philosophy, you could call this a 1 to 20 dilution or you could call it a 1 to 10. There would be a valid argument for either. The laboratory that did the testing by convention called this a 1 to 20 dilution. But there are many other labs which would call it 1 to 10. I think there was an effort to try to harmonize the definition with what other people used that led to the reclassification. But this is essentially what we're talking about here.

DR. EICKHOFF: Thank you.

DR. TREANOR: Right. And the other important point is everything started with the microneutralization test, and that's where this definition came from. And then we wanted -- the HAIs would use the same definition so it wouldn't appear that one test was artificially more sensitive than the other. So for our studies, everything used this convention as calling what the starting dilution was. When you go back to using HAI, it's more conventional to use this definition. And that's sort of how things evolved as HAI became more important than neutralization.

DR. EICKHOFF: Thank you. Second question for anybody and perhaps Bruce. What would be the trigger for use of this quote pre-pandemic vaccine?

DR. GELLIN: So, again, the terminology gets tangled. This is a pre-pandemic vaccine. We're not talking about a pre-pandemic vaccination program. So those often get confused. So the idea is that this is what you'd have available with the declaration of a pandemic as you then were going back and creating the pandemic vaccine.

DR. KARRON: Actually, Bonnie, did you have a comment?

DR. WORD: I just had a question that part of it is following up what Dr. Modlin had mentioned when he asked about other groups. He asked about children and elderly. I guess my question was related to what plans did Sanofi have for looking at high-risk groups? Because when you start looking at that one slide when you talk about the difference in ages and how they responded, perhaps, as Dr. Jackson mentioned, most of your high-risk individuals fall into that greater than 50 age group? And I don't know if they're planning on looking at that group, because that would be the majority of people. That's why we chose that age -- or that age was selected.

MR. GUITO: So as I mentioned earlier, Sanofi Pasteur has extensive development efforts underway looking at not only traditional manufacturing methods but some novel approaches with cell-based production and different adjuvant approaches as do many other manufacturers. And we think that our direction is best served in this area rather than expand the studies with the 90 microgram formulation at this point.

DR. WORD: So the answer is no, you're not going to look at it in high-risk groups?

MR. GUITO: The answer is no.

DR. KARRON: Dr. Self.

DR. SELF: So I'd like to go back a little bit to the use of this. There's a question about the trigger but trigger for what? The -- I mean there's this prospect of pandemic which raises all sorts of images and where I'm being asked to make some balance between the risks and benefits of this vaccine. While anything is better than nothing in a general sense, there is a specific use in mind. And so there does seem to be some sort of minimum level of efficacy that we need to be thinking about in making this balance. So could you describe a little more what this -- how this stockpile would be used and what the modeling that was briefly alluded to suggests as a minimum level of efficacy that would have enough merit to warrant the investment and licensure?

DR. GELLIN: Only because I'm closer to the mic, but I'm reading off of Dr. Robinson's slides, and I'll ask Norman to address the modeling piece which he had in his, but his first two bullets on Robin's slide 6 were that the goal was to establish a stockpile for 20 million persons and the critical workforce including first responders for use at the onset prior to the release of a well-matched vaccine. So that's the purpose of this stockpile. It's different than other stockpiles for other purposes, and remember it was sized for just a small portion of the population at that -- as the first responders. But -- so you can ask me more about that or I can turn to Norman about the modeling piece.

DR. SELF: Maybe we can hear about the modeling.

DR. KARRON: Well, I think, Dr. Robinson, did you want to comment a bit more on that first and then the modeling?

DR. ROBINSON: Two things. One is that the department and the administration certainly has two goals here, and one is to sustain the constitutional government and to maintain social and economic order at the onset of a pandemic. This vaccine has been set here as a stopgap measure until a well-matched vaccine is available from the vaccine manufacturers after a pandemic declaration. When a pandemic is declared by WHO or independently by the President or the Secretary for Health and Human Services can vary, you know, a little bit. And so if it seemed to be imminent and it's worthwhile to move to declare that pandemic such that we can start moving forward, then that would be done.

Secondly, as far as the modeling studies, and Norman can certainly attest to this, too, is that what's been seen is that if you have a vaccine that has as little as 33 percent match in efficacy for the circulating strain, then it can certainly, we'll call it, flatten the curve and buy time in which that vaccine can achieve the first goals that I mentioned and to be used until that well-matched vaccine is available.

DR. SELF: Thirty-three percent is a level of efficacy of the vaccine or 33 percent match to get a certain level of efficacy? What we're seeing here is a weakly immunogenic vaccine, only maybe 40 or 50 percent responding at levels that might be protective at all. So that 33 percent, if it's efficacy, might be very difficult to achieve with this kind of vaccine?

DR. ROBINSON: In different modelings -- studies that have been done by Neil Ferguson, and others, the 33 percent I use is a mean of what they see, and that is on -- is how well-matched it is and also combining then also with the amount of efficacy one might see as preventing death or very severe disease.

DR. KARRON: Actually, don't go away just yet. I have a follow-up question for this which is related to the use of this vaccine again and some of Dr. Word's questions. So the real intent of this vaccine is to be used for first responders to maintain order, etcetera? There's not -- is there an intent on the part of HHS at this point to use it beyond that level to stockpile beyond that level?

DR. ROBINSON: At the present time with a 90 microgram dose, and you see where we are right now with about enough vaccine for about 16 million persons with clade 1 and clade 2, that is what we'd use it for. As discussions will go this afternoon and in future vaccines development and we have more -- and we can see antigen-sparing that can be safely accomplished, then we'll have to revise because then we would see -- we'd have vaccine stockpiles that could be for many more people. But prioritizations right now are based on the 90 microgram dose.

DR. KARRON: For this vaccine right now?

DR. ROBINSON: That's correct.

DR. KARRON: I have two very specific questions for John Treanor, just clarification questions. One is actually related to slide 17, and I just wanted to clarify there were a number of subjects receiving one, two or three doses?

DR. TREANOR: Right.

DR. KARRON: And I just wanted to clarify are those -- those are subsets, so --

DR. TREANOR: They're all subsets of -- there's no one who received two doses who did not receive one dose if that's what you mean.

DR. KARRON: Right. So they're all subsets of the 363? Yes. Okay. The -- yes. The other question that I had was I know we'll have discussions about boostablility this afternoon, but in this particular study or as a follow on to this study, we saw the antibody titers at six months post vaccination. Were any of those subjects boosted and do you have any information about that?

DR. TREANOR: I think that data hasn't been completely finalized yet. We will have data on the response of individuals in the 063 study who received a third dose, and that will, at some point in the near future, include both safety and immunogenicity data for those subjects.

DR. JACKSON: While John's up there, John, was there any relationship between response to the first dose and response to the second?

DR. TREANOR: I don't have that analysis.

DR. KARRON: Dr. Modlin?

DR. MODLIN: Sorry to be persistent, but I wanted to go back to the last question I had, and it's I don't quite understand the rationale for labeling this for clade 1 use, if indeed, as proposed, this vaccine would be used in the event that we had a clade 2 outbreak. Would this require -- if we had a clade 2 outbreak, would it indeed require use under IND if clade 2 is not included in the label? This is maybe -- it's a sticky regulatory issue, I understand, but potentially an important one.

MS. BACHMAN: I'll answer you, John. I think we're -- the indication is not final yet. I mean this is the proposed indication from the company. We have to keep in mind sort of what situation will we be in. I mean if we labeled this vaccine as clade 1, I mean that does not prohibit -- I mean it's -- it would sort of be like, and I think somebody mentioned this earlier, that we had the seasonal vaccine and there was a mismatch. We wouldn't say stop using that vaccine. We would continue to use the vaccine that we had decided upon the strains and perhaps there would be some protection, but, again, we would not stop using that vaccine.

So in this situation, if we had that -- even if the vaccine, the label was clade 1 and this is all we had, we did not have a vaccine against Indonesia, we -- it's a policy decision whether we would continue to use that. But keep in mind that we are, or at least there are data being generated with Indonesia, and so, again, I think we really have to keep in mind this is in the interim. I mean we're not just freezing in time. As we evaluate this vaccine to sort of get us through this period as other vaccines are being developed and other vaccines are being developed rapidly and we need to look at those clinical data and we're moving into other generations of vaccines of all types for pandemic, so we have to keep a frame of reference here.

DR. KARRON: Dr. Webster's going to ask the last question and then we'll move on. We'll have time for discussion after.

DR. WEBSTER: No. I'm going to make a comment on clade 1-clade 2 cross-protection. The information is just not available at this time in humans, but the animal studies already indicate that vaccination with a clade 1 virus in challenge for the clade 2 gives considerable protection. So it's -- I think it might come down to the labeling issue. There is more and more information coming on cross-reactivity between these clades.

DR. KARRON: Thank you. I think we'll move on now to hear from Dr. Davis on Post Marketing Safety Monitoring During an Influenza Pandemic. It's Dr. Shay on effectiveness first.

DR. SHAY: Thank you and good morning. I've been asked to speak briefly about CDC's plans to monitor the effectiveness of pandemic influenza vaccines. Of course, limited immunogenicity and safety data will be available prior to distribution of any pandemic vaccine and safety monitoring will be essential. Post licensure safety studies can begin in a pre-pandemic use of each product and continue throughout the whole vaccine program. And if desired, post licensure immunogenicity data could also be collected in a pre-pandemic setting.

Data concerning clinical effectiveness of pandemic vaccines will be essential, of course, and immunogenicity and protection from illness are imperfectly correlated. Different populations may receive vaccine in pre and post licensure situations as well. And we, of course, always need to consider issues of vaccine match and perhaps the need to change the strain of a pandemic vaccine during the course of a pandemic. But obviously, studies of clinical effectiveness must await the onset of a pandemic and illness in populations who are eligible to receive stockpiled or pandemic vaccine.

So as we talk about CDC's vaccine effectiveness plans, we'll define effectiveness as protection against influenza illness when vaccine is administered in the context of an immunization program and that is outside a randomized clinical trial. Effectiveness may vary by age, by medical history and immunocompetence of the vaccine recipient. And effectiveness, we can expect, will vary with the outcome studied as well such that it'll be lower for non-specific illnesses that may be caused by pathogens other than a pandemic virus and that it may vary with the severity of the outcome being studied whether that's illness, hospitalization, more severe illness such as need for mechanical ventilation or death. We'll also need to plan to assess effectiveness after one and two doses of vaccine.

Our existing plans for pandemic vaccine effectiveness assessment is really built on our existing influenza vaccine effectiveness projects. Two of these projects build on existing surveillance systems for influenza and those are the Emerging Infections Program or the EIPs and the New Vaccine Surveillance Network or NVSN. The third project is one with the Marshfield Clinic Research Foundation which was funded to provide rapid, within season estimates of vaccine effectiveness against a laboratory-confirmed outcome. And all our existing studies do use laboratory-confirmed influenza illness as the outcome, although the specific outcome does vary with the study.

So this is a map showing our population-based influenza surveillance at present. There are 12 Emerging Infections Program sites scattered throughout the country in those orange areas that mark the counties in which influenza surveillance is done. Currently, in the EIPs, children less than 18 years hospitalized with laboratory-confirmed influenza infection are the surveillance group, and adult surveillance began in January of '06 as a pilot in several of the sites.

The New Vaccine Surveillance Network is in three counties, Hamilton County, Ohio at the University of

Cincinnati, Davidson County in Tennessee with Vanderbilt University, and Monroe County in Rochester, New York in the University of Rochester. And those latter two counties overlap with EIP surveillance.

In these sites, children less than 5 years with inpatient or outpatient laboratory-confirmed influenza infection are the cases that are sought. And outpatient surveillance in children age 6 to 12 years started this season.

So to go over these studies in a bit more detail, the EIP study is a case control design. It was piloted last season and will continue this season in '06-'07. The setting is hospitals, last season, in six of the EIP sites and this year in nine of the EIP sites.

The cases are children aged 6 to 23 months old in '05-'06 and 6 to 59 months this season to reflect the ACIP recommended age groups for receipt of vaccine.

Cases are children hospitalized with laboratory-confirmed influenza as by test ordered by clinicians, and the most common test ordered are DFA, rapid antigen detection, and culture in that order. Controls are age and zip-code matched children not hospitalized with influenza. Vaccination data are sought from healthcare provider report and by parental report via telephone interview.

The sources of other data include medical chart review by the provider and, again, parental review. And some of the other data collected are age, gender, race, insurance status, high-risk medical conditions, socioeconomic status, smoking in the household, those sorts of variables.

The New Vaccine Surveillance Network studies now also include case control studies. Therefore, studies were done in the '03-'04 season and continue up into the present season.

The setting for these studies are hospitals, emergency departments, and outpatient clinics, again, in those three counties. The children are aged 6 through 59 months. The cases are children brought to medical attention with fever or acute respiratory who test positive when enrolled in a -- by study nurses for influenza, by culture, or RT-PCR.

And controls are children, in the current studies, with fever or ARI, again, brought to medical attention who test negative for influenza by culture and RT-PCR. Vaccination data are obtained again from healthcare providers.

Other sources of data are sought through medical chart review by abstractors and again by parental interview. And other data collected in this set of studies are age, gender, race, insurance status, again, high-risk medical conditions, socioeconomic status and other factors that are known to be risk factors for hospitalization with viral respiratory pathogens in children.

And finally, the Marshfield Clinic studies include cohorting case control designs. These studies were started in the '04-'05 season and continue to the present. The setting here is a clinic population in North Central Wisconsin where a very large majority of the population receives their care through the Marshfield Clinic and their affiliated clinics.

The age group that is studied is all individuals for whom ACIP currently recommends annual vaccination. Cases in this set of studies are patients seeking care for acute respiratory illness who are influenza positive by culture or RT-PCR. And again, the cohort is a set of adults and children for whom ACIP has recommended annual vaccination and a cohort analysis is done.

In addition, there are age-matched controls without ARI symptoms who are in the same healthcare system. And again, as is done in NVSN studies, also a set of test-negative ARI controls.

Vaccination data in the Marshfield studies, what makes them rather unique, is obtained from a regional electronic vaccine registry that includes all vaccinations essentially given in their service area, so even if you get your vaccine at Kroger. Other sources of data include electronic medical record and interview of the patients. And again, Marshfield has a totally electronic medical record, so data such as age, gender, race, high-risk conditions, and propensity to seek healthcare as assessed by previous healthcare visits is accessible, essentially, immediately.

To use any of our existing systems to look at vaccine effectiveness, we have to think a little bit about pandemic vaccine prioritization and how stockpiled and other vaccines will be used. Everyone will be susceptible, of course, and U.S.-based production capacity is not currently sufficient, as we all know, to provide vaccine rapidly for the entire population. It is assumed that the earliest doses of vaccine will be available approximately 20 weeks after isolation and characterization of a pandemic virus.

So many of you, I am sure, are familiar with the ACIP and NVAC priority groups for pandemic vaccine. This was the joint work of the two HHS committees, and the process entailed consideration of estimates of vaccine supply and effectiveness, the effects of pandemic by age and risk groups, and the potential effects in critical infrastructure and healthcare. And the recommendations from ACIP and NVAC were included in the 2005 HHS pandemic plan as guidance for state and local planning and to promote further discussion.

And so sort of the top two ACIP priority groups were 1-A, vaccine and antiviral manufacturers and medical workers who are involved in direct patient care contact and support services, so about 9 million people. And the second group were those the committee felt would be at highest risk for pandemic-associated outcomes, included persons age 65 or greater with one or more influenza high-risk conditions or 18 million people, approximately; persons aged 6 months through 64 year with two or more high-risk conditions, another 7 million; and those 6 months and older with a history of hospitalization for pneumonia or influenza in the past year, so another 700,000 people.

Again, there has been considerable discussion. After the ACIP recs, an interagency pandemic vaccine prioritization workgroup was formed and include participants from multiple federal agencies. They considered the ACIP and NVAC recommendations and considered the National Infrastructure Advisory Council recommendations on critical infrastructure sectors that would be most important, and there have been public engagement meetings and the stakeholders meeting.

And there's a summary of those meetings. At each of the three meetings, the most highly rated goals were the same, and that was maintaining critical societal functions, protecting those who would help others during a pandemic, including healthcare workers, and a priority placed on protecting children, especially against pediatric mortality.

Most other goals were considered modestly important and those included protecting those most likely to get sick or die during a pandemic and although the ranks and rank order did vary between these meetings.

This group has developed draft prioritization guidance, is going to hold additional meetings, solicit written comments. ACIP, for example, was updated very recently by Ben Schwartz of NVPO on this work. And this working group will also consider pre-pandemic vaccine prioritization and will modify guidance -- how to modify guidance at the time of a pandemic. Final guidance is expected by May and, of course, all these considerations influence how we are thinking about needing to be prepared to monitor effectiveness of stockpiled and other pandemic vaccines.

So with those considerations in mind, here are present sort of plans. We will study laboratory-confirmed outcomes. Hospitalizations, for example, are well-captured in several of our systems and our severe -- additional more severe outcomes may also be studied such as all-cause mortality depending upon the nature of the pandemic. Obviously, it will be much easier to study such an outcome in a severe 1918-style pandemic than in a '68-'69 pandemic.

And of course, observational studies outside the context of randomized trials must collect data on possible confounding factors between receipt of vaccine and outcomes. Selection bias, for example, is likely but we can't assume the direction. In older individuals, if those with more severe out -- more severe underlying diseases are prioritized for receipt of vaccine, they are likely to be more ill than the underlying population of people that age.

On the other hand, if the vaccine goes to very narrow groups of younger people such as firefighters, they may be more likely to be healthier than the underlying population in that age group.

And of course, we will need to link existing individual health data to vaccination and outcome data to control for these possible confounders. And our plans will continue to evolve as vaccine priorities develop. For example, again, our existing systems cover children well but specifically in the context of this vaccine that's being spoken about this morning, community based studies may not be as very efficient if initial vaccine is prioritized to a few critical infrastructure sectors and we'll have to take other study designs to get at those individual, small populations.

Also, we need to think about vaccine distribution and tracking methods. State and regional registries may be used to identify vaccinated individuals if all available pandemic vaccine comes through government sources. But again, there will be a need to link pandemic vaccine receipt back to the medical home, if you will, such that medical and demographic data are able to be collected and used in analysis of effectiveness.

We also have plans to expand our existing systems. For example, in the future, hopefully, we could study effectiveness among adults, hospitalized adults in the EIP system. We would also be interested in expanding the rapid method used by the Marshfield Clinic to other sites that have electronic medical records. There's also the potential for new systems. For instance, consideration of using our sentinel provider system and some of the point of care diagnostic tests that Robin Robinson referred to that are being developed under HHS contract.

And finally, CDC is eager to work with our governmental and other partners to make sure that we're able to provide effectiveness data that meets the needs of the nation.

Thank you. And I'd just like to acknowledge those people who contributed to this presentation.

DR. KARRON: Next we'll hear from Dr. Caubel about the Sanofi Plan for Pharmacovigilance.

DR. CAUBEL: Good morning. My name is Patrick Caubel. I am Head of Pharmacovigilance for Sanofi Pasteur in North America. Planning for the prospect of pandemic influenza is one of the most effective steps to mitigate the impact of such an event. Preparing for the next influenza pandemic requires support and collaboration from multiple partners at the state, national and international levels.

Vaccination remains a critical defense against a pandemic influenza. Vaccine safety monitoring is critical and should be part of a comprehensive plan, public health surveillance program in which we are committed to take part.

Pharmacovigilance plan objective -- the objective of the pharmocovigilance plan should be to detect, to evaluate and to minimize the potential risk due to the pandemic influenza vaccine. It should contribute to the benefit risk evaluation in a pandemic situation. There should be an agreement on several objectives. Number one, the objective for the post marketing safety surveillance; number two, a collaborative plan with the key stakeholders; and number three, we have to establish a system which is going to function in a pandemic situation.

Pharmacovigilance planning will be critical in a pandemic situation. Pharmacovigilance activities have to be designed considering the following constraints: Number one, there will be limited clinical data available prior to the onset of a mass vaccination. Number two, a high volume of safety data, mostly spontaneous reports, is anticipated during a very short timeframe.

Number three, an increased public anxiety with adverse events reported is expected regardless of the causality. Any adverse event reported of any size is going to increase -- most likely are going to increase the level of public anxiety.

Number four, limited qualified personnel will be available in the industry and regulatory agencies. And finally, the pharmacovigilance stamps are very likely to be disrupted. In addition, there is a need for an ongoing safety signal detection and evaluation in order to enable appropriate decision with respect to the vaccination campaign.

And finally, the feasibility and effectiveness of appropriate actions and measures need to be tested prior to the onset of a pandemic.

I'm going to try to distinguish between what may happen during the pre-pandemic period and during the pandemic period itself. So during the pre-pandemic period, the usual routine pharmacovigilance practices will apply. Start out spontaneous report will be collected routinely. Aggregate reports will be produced at different time intervals. Signal detection and analysis will be conducted at also regular intervals. Safety surveillance studies could be initiated and one can think of possible cohort study in the first responders who are going to be vaccinated prior to the onset of a pandemic.

Passive collection of vaccine federal reports are going to take place as it is usual for any vaccines. The objective of this plan in the pre-pandemic period is to develop a better understanding of a vaccine safety profile that could impact the pandemic safety monitoring.

In a pandemic situation, we propose that some changes to the usual pharmacovigilance practices are considered. Number one, we would like to proceed a more focused spontaneous reporting on adverse events of high safety importance. We would like also to consider simplified aggregate reports focusing on the issue of real public health interests.

Number three, we feel that the real time signal detection analysis is necessary to allow quick decision making on the vaccination campaign.

Number four, we have the need for a safety surveillance study in earlier recipients after the pandemic is declared. And number five, the passive collection of vaccine federal reports will continue as usual.

The objective of this proposed revised pharmacovigilance practice is not to diminish the level of safety surveillance but more to allocate the available on tasks critical for understanding the evolving benefit-risk profile in the pandemic situation. We think that we need to focus on the information on the analyses which are going to provide the most relevant information to -- in order for the authority to make the appropriate decisions.

I would like now to examine the critical steps that could be impacted -- critical pharmacovigilance steps that could be impacted in a pandemic situation. Number one, the spontaneous reporting -- data collection -- well, spontaneous reporting will remain the basis for safety evaluation. We think that one, common, simplified and targeted collection form could be used by all parties when the vaccination process begins. It should help to focus on the collection of the most important adverse events and for safety monitoring of pandemic flu vaccine.

Healthcare professionals and patients were very likely to be the primary source of information and should also be encouraged to report primarily serious adverse events, life threatening adverse events, adverse events of special interest. And I will come back to that later --

Adverse events of special interest -- all parties, in fact, must consider a list of adverse events of special interest for which a common case definition will be used in order to ensure harmonized safety analysis of cases. Europe has already proposed a list of adverse events of special interest for pandemic flu vaccines survey and we propose that the key stakeholders in the U.S. who are on a similar list of adverse events have special interest.

Focusing our safety analysis on these terms without neglecting, and I want to be sure it's quite understood, we weren't neglecting the rest of it, that declaration could lead to a quicker identification of a potential safety signal.

The safety database is the repository in which key safety analyses are going to be conducted during the pandemic period. It is important that all stakeholders seed the database with all safety information available. The rapid and open communication and information sharing between Sanofi Pasteur, other vaccine manufacturers and authority, public health and public health services is absolutely essential, and electrical communication also should be established prior to the pandemic period. One single safety database dedicated to flu pandemic vaccines could be used and shared by all parties and, for example, a subset of a VAERS database could meet these goals.

Aggregate report or period reporting -- Periodic Safety Update Reports are prepared at define time intervals. However, during the pandemic period, due to the limited resource, preparation and submission of PSUR may not be feasible. So we think that several options might be considered. The first one could be to have some what we call simplified PSUR focusing on serious adverse, even death, life-threatening events and adverse events of special interest.

Another option could be to have a PSUR prepared on ad-hoc bases upon request from authority if any suspicion of potential signal or potential issue emerges. Of course, an aggregated PSUR will be prepared and submitted with the pandemic is declared finished.

Signal detection is a critical step for identification of safety issues with vaccine and with any pharmaceutical product. The crude inspection of single indicators and line listing is not any more an adequate method to detect a safety signal. We need to consider quantitative and automated data mining methods, for example, using different statistical scores like proportional reporting rates, Bayesian methods to enhance the efficacy of signal detection.

The modification of the standard method might be required, like some specific stratification by sub-population age group. We need also to identify the appropriate comparator in the pandemic situation. In fact, this data mining may detect an increase in the incidence of the adverse events of special interest and also help in the detection of unexpected adverse events. The signal detection tools and practices should be tested with seasonal vaccine prior to the onset of a flu pandemic.

Monitoring for vaccine effectiveness -- and we just spent some time reviewing this issue -- well, as you know, there is no vaccine which is 100 percent effective, and this applies in particular to the vaccine we are reviewing today. Vaccine failure evaluation done through pharmacovigilance monitoring should not be used to assess vaccine effectiveness, and if you want, I can come back to that later, but the case we are collecting during the pharmacovigilance process are uncontrolled by nature and not eligible for effectiveness assessment.

Safety surveillance studies are powerful tools to assess the safety profile of a newly licensed vaccine, and this applies as well to the flu pandemic vaccine. The safety profile would remain unknown in numerous populations prior to the vaccination campaign due to the lack of clinical data in sub-populations. So some consideration should be given to initiating cohort study, either pre or prior to the pandemic, for example, in first responders and critical works who, or after the pandemic is declared, like, for example, in the earlier recipients of a vaccine.

Case control study using large population-based databases like Vaccine Safety Datalink may be useful for the analysis of rare adverse events.

These studies, given the complexity, and in particular in the pandemic situation, should be coordinated by national and international public health agencies.

So in summary, streamlining and prioritizing is essential for early detection and communication of potential risk, and consequently for a good and rational decision making. The pharmacovigilance plan and information system must be tested and harmonized during the forthcoming and subsequent season in order to be sure that it will be fully effective during the pandemic.

The proposed pharmacovigilance actions are part of an evolving plan to be refined with key stakeholders together with a better definition of roles and responsibilities. Of course, you know, Sanofi Pasteur is coming to global pandemic preparedness and last point -- conclusion is safety is, for our vaccine, is of paramount importance for Sanofi Pasteur, and Sanofi Pasteur is prepared to work with the Government on efficient methods of collecting safety and effectiveness data. Thank you.

DR. KARRON: Thank you, Dr. Caubel. Dr. Ball?

DR. BALL: Good morning. My name's Bob Ball. I'm Chief of the Vaccine Safety Branch in CBER, and I'm going to be talking this morning about two topics, the first is pandemic influenza vaccine safety and effectiveness monitoring. I'll be including some information provided on CDC's plans for vaccine safety monitoring systems by Dr. John Iskander of the CDC Immunization Safety Office. And then I'll be providing some comments on Sanofi Pasteur's H5N1 vaccine pharmacovigilance plan.

So first, some general considerations for why it's important to do post-marketing safety monitoring of pandemic flu vaccines. There is limited safety and effectiveness data available for these vaccines prior to use. Robust safety and effectiveness monitoring is essential for, really, three reasons. Morbidity due to adverse events may be severe, and the best historical example of that that's relevant here is probably Guillain Barre Syndrome that occurred after the 1976-77 swine influenza pandemic vaccine use. It's also important to alleviate unwarranted fears to strengthen competence among the public in the safety and effectiveness of the vaccine. And finally, this type of data will facilitate benefit-risk analysis.

In recognizing this, Homeland Security Pandemic Influenza Plan has tasked the FDA with, among other things, tracking adverse events following vaccine administration and coordinating the definition of protocols for conducting vaccine effectiveness studies during a pandemic. We've heard already a little bi this morning about some of the factors that will affect vaccine safety and effectiveness monitoring. Those includes the stage of the pandemic that the vaccine is used, the population receiving the vaccine, and strategy for vaccine distribution. I only want to point out that because of the variety of possibilities, it's important that a robust and flexible system be in place for vaccine safety and effectiveness monitoring.

And I'll shift a little bit and talk about key systems that are currently available to study vaccine safety. First, there is the Vaccine Adverse Event Reporting System, or VAERS, which is the early warning system of vaccine safety surveillance. It's a national passive surveillance system that's jointly operated by CDC and FDA. It's been in place since 1990 and accepts reports from physicians, other healthcare providers , and the public. It's a hypothesis-generating system that seeks signals of potential concern.

Advantages of VAERS are that it is national in scope, covers diverse populations, and is able to detect rare events in a cost-effective manner. Rapid detection of possible signals is possible through VAERS, and these generate hypotheses that can be tested in other systems. It's also possible to assess lot-specific vaccine safety.

Disadvantages of VAERS include the presence of reporting biases. It's known that if there's under reporting, those serious events are more likely to be reported than non-serious events. And there's also over reporting since many reports that are not causally related to vaccination are also reported to VAERS. VAERS does not provide information on the number of persons vaccinated or the background incidents of conditions in the general population, so this information has to be obtained elsewhere.

The vaccine safety data link at the CDC was developed to account for some of these limitations of the VAERS system. There are eight geographically diverse health maintenance organizations that participate in a large linked database which tracks vaccination, outpatient, emergency department, hospital, and laboratory data to measure health outcomes, contains demographic variables which can be confounders and covers about three percent of the U.S. population. The VSD can be used to test the hypotheses that are generated by VAERS or other sources.

Advantages of analyses in the Vaccine Safety Datalink include that all medical encounters are available at most of the sites. It allows calculation of background rates of adverse events that can be compared to reporting rates in VAERS. Medical chart review for diagnostic validation is possible, and it's rapidly available for urgent studies.

Some of the limitations of VSD analyses include that the sample size, although very large, still may not be adequate very rare events such as Guillain Bare Syndrome with an incidence of about 1 to 20 per 100,000 per year is background. Vaccines administered outside of the HMO setting is not captured by the VSD database, and there is limited demographic and socioeconomic diversity in the eight HMO practices. The unvaccinated population may be small and so, therefore, may require special methods for analyzing outcomes.

The CDC also has other resources available for study of vaccine safety. The CDC collaborates with the Brighton collaboration which developed standardized case definitions of adverse events following immunization for use in clinical trials or epidemiological. CDC also sponsors the Clinical Immunization Safety Assessment Centers, or CISA, which developed standardized patient evaluations for adverse events and can provide clinical guidelines for providers in managing adverse events after vaccination.

Additional data available from the CDC comes from the biologic surveillance system which provides vaccine dose distribution for a calculation of reporting rates of adverse events, though it's important to note that this is not doses administered, simply doses distributed.

There are also a number of surveys from which important information can be gathered including the Nationally Representative Coverage Surveys, National Health Interview Survey, National Immunization Survey, and the Behavioral Risk Factor Surveillance System. CDC also has extensive relationships with state and local health departments and immunization registries. Hospital discharge and mortality data sets are also available which can be used to calculate background rate of adverse events.

In preparation for pandemic, the FDA and CDC have undertaken some additional activities for vaccine safety. These include pilot projects to assess the use during a pandemic of the VAERS system beginning with the 2006-7 influenza season reports. Also, in collaboration with Harvard, there's a planned expansion of the Vaccine Safety Datalink to another site to expand the number of people under study and evaluation of other large automated databases of encounter and/or claims data for similar use.

The FDA has also begun a pilot project with the Center for Medicare and Medicaid Services to obtain rapid access to data on influenza vaccine and treatment claims since Medicare beneficiaries are a large group that receives annual flu vaccine. FDA has also begun discussions with DoD and the Veterans Administration on their plans for a pandemic influenza vaccine safety and effectiveness monitoring.

So in expanding the existing systems, a number of principles are important to keep in mind. Complementarity, coordination and minimized overlap between government agencies and vaccine manufacturers to ensure that each is contributing valuable information to safety and effectiveness monitoring is important.

To that end, the FDA has initiated pharmacovigilance planning, and since 2005, has requested that vaccine manufacturers submit a pharmacovigilance plan with their Biologics License Application. These pharmacovigilance plans should follow FDA and International Conference on Harmonization E2E guidelines on pharmacovigilance planning. And in addition to reporting of adverse events to VAERS as required by regulation, it's often important to have enhanced safety surveillance and/or observational studies as part of these plans.

And it's highly recommended that sponsors work closely with the FDA and CDC to develop and conduct studies to monitory safety after licensure.

So you heard just before me the presentation of Sanofi's pharmacovigilance plan. I'm just going to make two general comments about two aspects of the plan. First, Sanofi has proposed changes to adverse reporting during a pandemic that are not consistent with current regulations include less frequent or simplified submission of periodic update and/or other reports and use of simplified reporting forms.

The FDA has not yet made any decisions about whether or not changes to adverse event reporting will be needed and what they might be during a pandemic.

Sanofi also does not propose to conduct additional safety or effectiveness studies of the H5N1 vaccine.

As we go forward, there are a number of issues that require clarification for safety monitoring for your consideration. Should specific adverse event reporting requirements be increased or decreased? Do we need to monitor for particular adverse events of interest, and if so, what are they? Who, FDA, CDC, Sanofi, other groups, should be responsible for what aspects of safety monitoring of the H5N1 vaccine? And how might these above considerations vary according to pandemic stage?

Similarly, there are a number of issues requiring clarification for effectiveness monitoring. What outcomes should be assessed to evaluate effectiveness, that is how should influenza be defined? We heard from David Shay talking about laboratory-confirmed influenza, but it might also be necessary to evaluate off-course mortality in large claims databases. What study design should be used to evaluate effectiveness to account for some of the issues that David mentioned about biases in certain study settings? And then who should be responsible for what aspects of effectiveness monitoring of the H5N1 vaccine, and how might this vary according to pandemic stage?

So finally, a robust and flexible safety and effectiveness monitoring system is needed to address the range of possibilities during an influenza pandemic. Epidemiological studies will likely be important, and close coordination between Government agencies an Sanofi Pasteur would be beneficial. And it's desirable for Sanofi to commit to working with the FDA and CDC to fill the gaps in collection analyses of safety and effectiveness data if the H5N1 vaccine is used.

And I'd just like to acknowledge those who helped with this presentation. Thank you.

DR. KARRON: Thank you. Questions or comments for any of the previous three speakers? Dr. Wharton?

DR. WHARTON: I think in thinking about how to monitor safety and effectiveness, it's really important to keep in mind that we don't know how the vaccine is going to be distributed. And this was implied by a number of speakers, but just to make it clear, if we are using a vaccine distribution system similar to what we use for seasonal influenza where vaccine is distributed through multiple providers, many of them healthcare providers that are involved in our existing networks, then our existing infrastructure may be very helpful in looking at both effectiveness and safety, because we're likely to capture those populations that are being immunized.

But for some of the possibilities for how a severe influenza pandemic might play out, we could be dealing with a quite different distribution system. So if, for example, we were targeting first responders, those are unlikely to be captured by the VSD, and these critical infrastructure workers are not the usual target group. So just to make that really clear as we're thinking about these things, that we may need quite different systems than we currently have to answer some of the questions that will be important to answer.

DR. KARRON: Dr. Hetherington?

DR. HETHERINGTON: I wonder if we have any clarity on how first responders are identified, and you obviously will have a limited number of doses of vaccine available, so how will people be designated as receiving? If, in fact, they are heavily concentrated within the governmental agencies, wouldn't there already be a database available that would be able to be mined for safety follow-up in a sense?

In follow-up to previous question, in fact, there may be a structure that's available if we know who the first responders are and we know where the links to their healthcare reside, whether they're in a governmental database or some specific HMO. So that -- I wonder if that's been given any thought by the FDA or anybody else?

DR. GELLIN: Let me comment on that. And as David Shay presented in his description of the ongoing process, there is currently a revisiting of the prioritization for vaccine in a pandemic. And the biggest difference is really the incorporation of the input by the National Infrastructure Advisory Council, I think it's called, which is a DHS Homeland Security Advisory Council, to get a better sense of who in the critical infrastructure is critical. But you raise an important point because, as Melinda highlighted, it's likely particularly early on to be distributed differently than seasonal vaccine.

So I think that is important -- so I don't have a clear answer for you other than I think you highlighted that recognizing that there are different systems of care for these people that we should look at whatever existing databases might be able to capture that care, whether they're in the Defense Department, whether they are in Occupational Health or whatever. But I think that that's an important consideration.

DR. KARRON: Dr. Farley?

DR. FARLEY: Given all the complexities of not knowing in advance, really, how this will be introduced -- I mean having some general concepts -- and it will be somewhat unprecedented in terms of if we launch a major campaign against pandemic flu or at least it's been many -- it's been a generation or so since we've had to do anything like this, and we will potentially have larger stockpiles, have actual access to vaccines in a more timely fashion -- I mean how practical would it be to force a registry process into this so that each and every individual who receives a pandemic flu vaccine go into a national registry regardless of age and whether it be building upon old systems or having a dedicated system to this event? I don't know who might be able to answer that but maybe Bruce.

DR. GELLIN: Let me start and others may add to that, but among the investments that are being made of this in a pandemic budget is looking at registries. And I think the question you raise is whether or not existing registries or some alternates. Because I think that here is the opportunity to do many things, not only to monitor both safety and potentially effectiveness, but it also is the complexity of this is a two-dose schedule. You want to ensure that those who get the first dose get the second dose as well.

So I think that -- I don't know if anybody wants to comment on that specifically other than there is a -- that's recognized as one way that there might be a system either building on existing systems or creation of a slightly different system to be able to accomplish those three goals.

DR. WHARTON: Yes. Just to extend what Bruce said, at least what the immunization program grantees tell us is that they are planning -- the majority of them are planning on using their existing immunization registries as part of their activities.

Now at this point, state programs are going asked to make pandemic plans and I think this is on the assumption that there will be a state health department directed public health activity that will be I charge of the initial states of vaccination. So it certainly is possible but the existing immunization registries can help with this. I do have to say that there are concerns about the capacity of those registries as they currently exist to actually fulfil this function, but at least the majority of state programs have told us that is what they plan to do.

DR. KARRON: Dr. Couch?

DR. COUCH: Just thinking back a little bit about 1976. When you start vaccinating everybody, you're going to need a frame of reference for a likelihood of events that will be occurring. Because you see, Guillain Barre was unanticipated. The deaths that occurred in Pittsburgh, you stopped that, you know, and that campaign was a little bit different. But you have to say that's not -- that's expected, you see, and be able to respond to that without a control group, because everybody's going to be in line to get their vaccine. And one of the options for that might be rapid response teams to something like the VAERS reports that can trace these, because they'll end up in the headlines of the local newspaper, and the local newspaper will cause all the talk programs and everybody else to start questioning safety unless you're prepared to respond.

DR. KARRON: Dr. Jackson?

DR. JACKSON: I just had a question for Dr. Wharton. When you say existing registries, what do you mean?

DR. WHARTON: Well, I'm not sure what the grantees meant. What they have reported to us is that they planned on using their immunization registries. Some of these are state. Some of them are local. But these All right. considered to be public health programs that are largely run by the public health infrastructure, so I assume these would be the governmental registries as opposed to private sector registries.

DR. JACKSON: So primarily childhood vaccination registries?

DR. WHARTON: Yes. And I think that's one of the capacity issues is that the registries were originally developed for childhood immunization although some states have now extended them to adults and there's no a priori reason that other states couldn't do that, but that is one of those capacity of infrastructure issues that raises some issues about whether or not these registries have the capacity to do that.

DR. KARRON: I think we'll have more discussion of this topic after the break. We have a fair amount of time budgeted for post open hearing discussion. So we'll take a break now and reconvene at 11:00 o'clock. Thank you.

(Whereupon, off the record at 10:45 a.m. and back on the record at 11:12 a.m.)

MS. WALSH: I think we're ready to begin. I'd like to ask everyone to please take their seats. Next on the agenda is the open public hearing. As part of the FDA Advisory Committee meeting procedure, we are required to hold an open public hearing for those members of the present who are not on the agenda and would like to make a statement concerning matters pending before the committee. I've not received any request at this time. Is there anyone in the room who would like to address the committee?

(No response.)

MS. WALSH: Dr. Karron, I see no response and I will turn the meeting back over to you.

DR. KARRON: Thank you, Christine. At this time, we will have the FDA presentation of questions by Dr. James.

DR. JAMES: Okay. I've already given you the proposed indication. I will repeat it once again. Sanofi's proposed indication is that H5N1 Influenza Virus Vaccine A/Vietnam/1203/2004 (Clade 1) 90 micrograms per milliliter is an influenza viral vaccine indicated for active immunization against influenza disease caused by H5N1 A/Vietnam/1203/2004 (Clade 1) influenza virus and primary vaccination of healthy adults 18 through 64 years of age.

The first question to the committee is are the data sufficient to support the effectiveness of this product for use during a pandemic or in situations of potential high risk exposure?

Second question. Are the data sufficient to support the safety of this product for use during a pandemic or in situations of potential high risk exposure?

And the last question is please comment on studies to collect additional information about the effectiveness and safety following this vaccine's use.

DR. KARRON: Thank you, Dr. James. I'd like to open this up for discussion, but I actually will lead off with perhaps a question for you, Dr. James, if you don't mind. And that is I know that there's been a lot of discussion during these proceedings this morning about how and when and in whom this vaccine might be used. Am I to understand from the way this question is written that the vaccine would be used during a pandemic but not pre-pandemic, for example, in first responders? Is that -- am I understanding that correctly?

DR. JAMES: You are understanding that correctly. Sanofi did not phrase their indication, did not propose their indication that way, but we're specifically, the FDA is specifically asking you to consider the data that we've presented for use during a pandemic or in situations that may occur prior to a pandemic but that are potentially high risk exposure situations.

DR. KARRON: Ms. Province, did you want that clarified that last statement? Is that --

MS. PROVINCE: Yes, please. I'm sorry, it just seemed that you answered it one way and then answered it another way or maybe I misunderstood you.

DR. KARRON: I think that what maybe you were asking for is a clarification of what high risk exposure constitutes. Could you give an example of that.

MS. PROVINCE: You phrased that better than I did. Yes, what would constitute the high risk exposure? Would that be the first responders prior to pandemic?

DR. JAMES: First responders, military who may be deployed to Indonesia for whatever reason. Yes, high risk basically would be first responders and the like.

DR. KARRON: Dr. Self?

DR. SELF: Thank you. I work better when things are really concrete, so I'm trying to imagine, you know, what might actually happen. There are some transmission chains that happened in Indonesia or somewhere in Southeast Asia, a couple of cases start showing up around our airports, and then this -- those sorts of events would trigger the use of this vaccine among first responders followed by the 600 million dose reference 6 months later of some other vaccine or of this vaccine? Is that a scenario that we're talking about?

DR. JAMES: I actually would ask -- well, it looks like Norman wants to respond first.

DR. BAYLOR: I'll respond. I mean first off, I mean how this vaccine will be used other than during a pandemic, I mean these are policy decisions. But you may have examples where you start getting spread to human transmission from human-to-human, and a decision may be made at that time that, yes, we should start, you know, deploying this vaccine. Or you may have individuals going into a region where there is a human-to-human spread or even avian, there's a high level of avian influenza in the area. So these are decisions that will -- you know, they're policy decisions but thee are opportunities where this vaccine might be deployed.

DR. SELF: And then what's the relationship of this vaccine to the vaccine referred here, the 600 million --

DR. BAYLOR: Whose slides are you looking at?

DR. SELF: Six hundred million doses -- well, I ask because, you know, part of our charge is to think about what information we need to get about the characteristics of this vaccine. If this is a stopgap vaccine that bears little, if any, relationship to the vaccines that are going to really carry the load in a pandemic situation, then that's one thing. If it's very closely -- if it's the same vaccine, then the answers to that question are perhaps different, so it's --

DR. BAYLOR: Well, I think we, and Robin from the Department can answer as well, but when you look at this, you have to look at where we are in time. The 600 million doses, I mean it depends on where you are. I mean we know that there are numerous vaccines under development that are potentially better, if you will, than this vaccine. This is an interim vaccine. Depending on where those vaccines are in development, those vaccines -- some of those newer vaccines, the adjuvanted vaccines or what have you, those may be the vaccines that we use to hit the magic number of the 600 million doses.

But again, it just depends on how fast those vaccines are developed, how fast those clinical trials are done and what the data suggests or support for those vaccines.

So again, today we're -- what we're faced with is this vaccine that requires two doses and at 90 micrograms. That's where we are today. A month from now, a year from now, five years from now, we could have -- we probably will have additional products that we hope that are going to be better. You're still not clear?

DR. SELF: Well, to get information about effectiveness of this vaccine, we're probably going to be in the early stages of the pandemic? I mean we're not going to get that probably from the high risk groups? I'm just guessing because it's still not clear to me who those high-risk non-pandemic vaccinees would be. And that implies a certain timing of events then.

There is, you know, only a handful of months between the first opportunity to get effectiveness information about this vaccine and the 600 million doses or, you know, whatever the next wave of response is. And so, again, I'm -- the charge here is what studies, you know, should be done to characterize the effectiveness of this vaccine. And I'm wondering -- there are limited opportunities for that, and I'm trying to figure out what information would be critical and could be used in how this unfolds so I could, you know, answer that question.

DR. GOODMAN: Yes. I think that the likely -- you know, this is an interim preparedness measure. There's also -- you know, although we're all very optimistic about reports and information we receive about potentially more immunogenic vaccines, we always have to see that data. And particularly with things that haven't been used in hundreds of millions of people already, we have to be concerned that the safety database is adequate before a completely new technology is widely used.

Now all that said, what HHS has been encouraging is rapid development of that information from a number of manufacturers, and FDA is encouraging that we get that data as it becomes available so that in an emergency -- we're going to be in a situation -- we are in a very dynamic situation right now. Every six months, we learn a lot more. That's one of the good things about this, because there's a lot of information about flu vaccines that is being developed that is informing us. But it's also one of the big challenges and it's a challenge in making a static risk-benefit decision.

But what I was going to say is that so let's say there were a pandemic tomorrow, six months, two years. What we're going to do is look at all the data that are available out there, probably seek advice also about that data from people like yourselves -- where are we at that moment, what is the snapshot, and then what are the alternatives for trying to provide protection, you know, to the American people.

Right now today where this vaccine is and the use that's being proposed is that in a situation perhaps where human-to-human transmission is beginning to occur or we see people who may have high risk of exposure to H5N1 that this vaccine would be available for use under license, and getting -- you know, and obviously the amounts of this vaccine because of the technology are going to be somewhat limited. It's going to be what's in the stockpile. I think getting to your original question, what's important, and to CDC's presentation, is it will be good to be able to evaluate field efficacy of that vaccine. It may be more or less efficacious than expected, and that early information may inform decisions about further production, etcetera.

I think certainly there's a likelihood there could be a drifted strain or a different clade as discussed and that probably, you know, you wouldn't want to produce that even if it was with this technology.

So the real question is, and this gets back to, I think, your question about the modeling, and in my presentation later, I have a slide about that, but I think you don't want to make too much of models, but they have many, many assumptions. But part of what drives this is the historical experience with flu which is that some immunity can have an affect, even limited immunity. And then in these models, for example, suggest that if you achieve 30 percent protection even and even with 1 dose of such a vaccine, that there may be, in certain circumstances, combined with other measures, significant affects on a pandemic.

So the real question is here's a vaccine based on a known technology -- it is somewhat different; it's a higher antigen does; it's a unique antigen, but can this provide at present the potential for benefit in this situation, and it's a fluid situation.

But I think input on how CDC and FDA should help evaluate efficacy early in a pandemic would be helpful. And I also take Dr. Couch's point. And you heard from both the CDC and FDA presentations, there's a lot of concern for how do we improve monitoring the safety and communication about safety early in a pandemic, and we're trying to exercise those systems with annual influenza vaccine as well.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I think I'm struggling with what several others have alluded to or directly mentioned, but I guess for Dr. James and perhaps Sanofi Pasteur, I have one question and one comment. How can we come up with a better definition of high risk group? I think it seems very vague and unclear, and that's going to be a key issue in a stopgap measure. If it's beginning of pandemic, if it's pre-pandemic, if it's -- how are these definitions made? And I think they should be made.

And secondly, post deployment monitoring, I think that's one of our charges is to address that. It seems to me that there really have to be plans in place to compare not only unvaccinated and vaccinated individuals but also comparison of the people who are vaccinated with seasonal vaccine as a comparison group. And I think that's something that should be looked at for cross-protection or priming for future vaccines.

DR. KARRON: Dr. Hetherington?

DR. HETHERINGTON: Dr. Webster earlier said this is an historic vaccine in more than one way -- sorry, my microphone doesn't seem to be working --

Dr. Webster earlier said this is an historic vaccine in more than way, and I think what I sense we're all struggling is the roadmap to where this vaccine is going. We're at a starting point, but it's not clear what the evolution is in front of us. For instance, one of the questions to us is what additional studies should be done to assess the efficacy of this vaccine. Well, if the roadmap is to replace this with a new clade vaccine in the next 6 to 12 months, I'm not certain that we can justify any resource going into any additional efficacy studies, while safety certainly should be continued to be monitored. So it's not really clear to me that that has been laid out to the committee as a whole as to where this fits into the entire evolution of vaccines for this specific purpose.

We're told that this is going to be for first responders, but there is also an indication up there for primary vaccination of healthy adults which speaks to a much broader population, and it's not -- wasn't clear to me, at least, in the presentation that that was really something for which we should be considering this vaccine. It looked like there was more of an evolution to it. I wonder if somebody might address that?

DR. HACHEY: One group that's likely to receive this vaccine, in part because DoD does have a smaller stockpile than HHS, but nonetheless we do have a small stockpile of the 1203 (clade 1) vaccine. And we are probably likely to be included eventually in the national strategy as far as the larger stockpile that will be evolving over time.

DoD has some unique attributes that we have built into our immunization programs already that all of our active duty members which represent that healthy kind of middle-age-younger-age group, every vaccine that they do receive, whether it's influenza, anthrax, or in this case, a pre-pandemic or a pandemic vaccine, is already monitored. So we do have a tracking system already in place as well as a tracking system to monitor for adverse events. We have also established a system that could be easily adaptable to monitor, again, ongoing adverse events and efficacy of the vaccine.

So DoD is in somewhat of a unique niche in that we're likely to use the vaccine should a pandemic be imminent. We have some already in hand, and we do have tracking mechanisms that are either already in place or easily adaptable to monitor the outcomes that are currently of interest.

DR. KARRON: Dr. James, did you also want to respond?

DR. JAMES: Yes. I just wanted to clarify the third question that we are asking for is really on if the vaccine is licensed post use. So post licensing use of the vaccine, what sort of effectiveness and safety data you would like to see, you would like collected. Okay? So it's not necessarily unless you believe that the data presented are not sufficient to license it, then, of course, you can speak on additional studies that you believe need to occur. But that question is specifically for if the vaccine is licensed and it is used, how do we go about collecting effectiveness and safety data.

DR. KARRON: If I can maybe clarify or amplify your question, Dr. Hetherington. I think you were asking not so much that as perhaps where does this fit into the pipeline of vaccines that are going to come before the FDA to help us understand given that everybody admits that this is a stopgap vaccine, to help us plan what kinds of post licensure test, if this vaccine were to be licensed, should be done. Is that -- did I say that correctly?

DR. HETHERINGTON: I think the FDA spokesperson was correct in that I did misread the question, but I think the larger question was exactly as you said. What is the overall roadmap for this approach.

DR. KARRON: Dr. McInnes?

DR. McINNES: Thank you, Ruth. I'm struck by these conversations that we don't seem to have a real advocate for this vaccine, which is an unusual circumstance coming before the FDA Advisory Committee. We don't have a manufacturer of a product that intends to market this commercially, and all those incentives that go with that are not in place. And so I think we have to take a little bit more pragmatic view here with all of this uncertainty. And I think we do have a vaccine here. It is immunogenic. It's not as immunogenic as we'd perhaps like to see. It uses a lot of antigen. It's a -- you have to have two doses which is very inconvenient. But it is a vaccine.

And the safety profile in the small numbers that have been put before us, I find somewhat comforting. I think it's acceptable. Of course, we'd like more data but I think it looks quite good at this particular point in the small numbers.

I think all of the uncertainty tracks around the concern that, you know, a recommendation to license, and we don't understand how the product will be used, by whom, when, where. We normally have a vaccine manufacturer who is manufacturing bulk lots of vaccine, not just one, who is marketing product, it is getting used, we're continuing to gather data, which then provides the basis for a community assessment of whether this is safe and efficacious, and we're not going to have any of these readouts. We're not going to have any of these signals coming in this situation.

And so -- yet the option on the table is to look at licensure for a product, and we understand licensure and use in a completely different framework normally. So I am left -- I am struck -- you know, Sanofi does not propose to conduct further studies to gather data, and I understand that they were a contract manufacturing in this situation. They have no commercial market. I'm not sure I totally understand why there isn't a potential commercial market, but that's apparently not their plan.

So I'd like to understand a little bit more from the Department about a very clear articulation of use of this product, when, where, how, decision making. And I'd like ask for some guidance from the FDA about what options they have to manage not only the license but the use of the vaccine, because I think once it's licensed, it sits there with that. And I'm trying to understand the options for management of this package which is very unusual.

DR. KARRON: Bruce?

DR. GELLIN: Pam's laid out some important principles and questions. There's also been a discussion about 20 million, 600 million, where the vaccine sits, so I think that we need to remember that this was created for the stockpile.

And it's also important to remember that when we started doing this, the idea of stockpiling an influenza vaccine, if you, you know, rewind the clock and look at what the doctrine was at the time, nobody was stockpiling influenza vaccine because you know the virus would change and you had to keep up with it.

So the idea that you would go into creating stockpiles knowing that the virus would now do what it has shown it is going to do was where we got into this, but the idea was that a stockpile could provide some protection.

I would think, importantly, early on we also recognized that going through all this was going to provide a lot of experience to a lot of people that we didn't want to learn in the time of an emergency. That's separate from the discussions here, but what we have now is vaccine that's created for a stockpile.

As you've seen from Robin's slide, accumulating enough to hit the national target of enough vaccine for 20 million people is not easy given that your B essentially we refer to this as B you know, we're asking the manufacturers to use every available minute of their off season production. So the slid that Robin showed shows how you can only get incremental amounts of that to be able to put that vaccine into a stockpile. And then we have the issues of time that -- of what happens to this vaccine over time, what happens to the virus over time. So we have -- so the 20 million goal was set as a construct to have something at the beginning of a pandemic that you might use that would provide some protection to people on the front lines while you were then creating the better vaccine that was tailored to the circulating virus.

So the piece that was in -- so I'll draw on two different slides -- so I think that Norman ended with his final bullet was the benefit of having a licensed vaccine against a potential influenza virus strain weighed against the risk of having no vaccine. So that's an important principle.

The other piece was embedded within David Shay's slides that while he talked a lot about the process for revisiting the pandemic vaccine prioritization, there's a separate process done by the same interagency group to take a hard look at that if we had to use it today, how we would use the existing vaccine and then up to 20 million doses -- 20 million people, what kinds of people would those be. So I don't have a clear answer now. There is a pretty vigorous interagency process that's defining that at the same time its defining the priority list broadly, but that's the construct here.

Again, I think it's been signaled in many of these slides. It is front line, critical infrastructure that constitutes those 20 million, but I can't give you more precision until there is clear guidance on who those people are.

DR. KARRON: I do want to follow-up on Pamela's question, though. So does that suggest then that the 20 million doses in the stockpile, it would be used for first responders, who those people are is to be defined, but it would not be used for populations other than first responders?

DR. GELLIN: That's the goal of the stockpile recognizing it was a finite amount and that you would then -- we have the larger goal that Robin can talk more about of creating a production capacity so you would have enough for the larger population.

DR. KARRON: A production capacity with this vaccine?

DR. GELLIN: For the vaccine that you needed, that you have to determine at the time. I mean, again, I think that the general idea was that they -- this quote pre-pandemic vaccine was the best you could make based on the viruses that were circulating at the time that then might give you somewhat of a match with the idea being that you would likely have to have a different vaccine for the pandemic --

DR. KARRON: I just have a particular question for the FDA that has to do with their draft guidance, because obviously this stopgap vaccine, if you will, does not meet the criteria -- some of the criteria set forth in the draft guidance. Will licensure of this vaccine, if it is licensed, have any impact on the draft guidance or not?'

DR. BAYLOR: I'll answer that. I mean, in essence, no, this vaccine was -- these clinical trials were done prior to these guidances. The guidance documents will be applied to forthcoming vaccines, but I think we have to look at the type of vaccine we're dealing with. Here we're looking at a vaccine that's been manufactured by a licensed process. If another vaccine comes forward with a U.S. licensed process, more than like, it will be evaluated the same way. That guidance document will apply to vaccines, in particular the pandemic guidance will be applied to the vaccines that are coming henceforth.

DR. KARRON: Dr. Wharton?

DR. WHARTON: Understanding the need to get these doses manufactured as quickly as possible and the need to get as many doses as possible out of the antigen that could be made, I'm assuming that this is preservative-containing vaccine in multi-dose vials. My question regards that formulation. Given that the 90 microgram dose, I believe, is a 1-mil dose, from what was said early, should I -- is it correct that those two doses administered would contain 100 microgram of mercury thimerosal preservative in those two doses?

DR. JAMES: I'm being told we -- yes, we believe that that's accurate, but if Sanofi can confirm that.

DR. LEE: Hi. I'm Dr. Sam Lee representing Industrial Operations for Sanofi Pasteur. Yes, the 1-mL vaccine does contain 100 micrograms per mL of thimerosal and the vaccine would contain that 100 micrograms.

MS. KRIVACIC: Given that it contains the thimerosal and we're looking at annual flu vaccinations, what is the risk of exposure of thimerosal from your annual vaccinations as well as this, you know, stockpile? If you can kind of comment on that in terms of the exposure of thimerosal?

DR. LEE: Right. I'm not sure I'm the right person to answer that.

MR. HOSBACH: Hi. I'm Phil Hasbach, Government Policy and Government Relations for Sanofi Pasteur. In terms of our seasonal flu vaccine, we have a variety of formulations available, some of it unpreserved with no thimerosal at all, and of course others with multi-dose vial. It's really constrained by our filling and finished capacity for single-dose syringes and single-dose vials. So for the traditional multi-dose vial vaccine, it has 25 micrograms of preservative -- of preservative in it, thimerosal.

In this instance, we're looking to produce as much vaccine as possible to get it into the arms of citizens as quickly as possible, and right now with the fill and finish capacity that manufacturers have, especially Sanofi Pasteur, it's optimal to do it with multi-dose vials and using a preservative.

MS. KRIVACIC: I had a question for Dr. James. On the studies that you were looking at with the pediatric population, was there any indication of those that were previously vaccinated with the annual flu vaccination, the seasonal flue vaccination that had thimerosal and then this particular flu vaccine?

DR. JAMES: In terms of the pediatric studies, I will ask the NIH to address that. The pediatric data has not been submitted to the BLA.

MS. LOWERY: So with the pediatric trial in which we evaluated two doses of the 45 microgram vaccine or an optional third dose of the vaccine, we did collect information on children who had previously received the TIV or the inactivated vaccine and also on if they had received FluMist, but we did not specify if the inactivated vaccine that they had received that, trivalent inactivated vaccine did or did not contain thimerosal.

DR. KARRON: Dr. Gellin?

DR. GELLIN: I think this is a manufacturing question. We often hear about that the switch to multi-dose vials to single-dose vials can translate into numbers of doses that can be lost because of, I don't know exactly why, some of it sticks to the side or whatever, but you have to do some overfilling of each vial so, therefore, that adds up. Given the slide that Robin showed of how difficult it is to even accumulate the targets we're going for, can you give us a sense of -- do you know what the math is, what the calculation would look like if you went from the number of multi-dose vials, how many single-dose vials would you have?

DR. LEE: Switching from multi-dose to uni-dose vials, actually, there's probably a twofold answer to that. It's problematic in two ways. One is certainly, as you said, the filling capacity. The actually filling rates are significantly different for a multi-dose versus a uni-dose vial because just the number of vials that you're dealing with. In this case, we're talking about a five dose per vial, 5 mL's in a single vial verus a uni-dose where there's 1 mL per vial. So the filling rates right there, you're taking five times as long to fill the same number of doses.

The second is just the sheer practicality of needing to handle so many vials for distribution and use. In a pandemic situation, you're talking about distributing 300 million, 600 million doses. You're talking about 300 million, 600 million vials versus. If this is a 5 mL dose or I'm sorry a five dose vial, you're talking one-fifth the number of vials. So those -- kind of a twofold aspect.

In terms of the actual overfill and what you have addressed there, there is a slight difference in terms of need to overfill of uni-dose vials, and it could be fairly significant on the order of 10 to 20 percent.

DR. KARRON: Dr. Krivacic?

MS. KRIVACIC: I just had one other question or comment. I guess I understand we don't have a lot of time here with regard to, you know, dealing with a pandemic, but I think it would be very important to understand what the effects of an annualized flu vaccine with thimerosal plus the avian flu vaccine, what kind of effect that would have on a woman of childbearing potential. That would be my concern and I think that's a concern you may get from the general population down the road so just a sort of an FYI.

DR. KARRON: Dr. Jackson?

DR. JACKSON: I guess this is a question for FDA. I'm struggling with the obvious limitation of having a study data set of 100 people who've received the recommended dose regimen and it being asked to go forward for licensure which is a very, obviously, unusual situation. With that number of people, I mean we're not able to exclude the possibility there are, in fact, quite common adverse events that we're not detecting in the original trial. And this is a vaccine produced by established methods.

However, it's got 12 times the amount of strain-specific antigen as the usual seasonal influenza vaccine course, so I think that leaves open a possibility that there could possibly be something unexpected when 90 micrograms is give twice.

In addition, we know very little about, as I said before, any sort of subsets, potentially important subsets, such as age within that group, and we probably can't learn a lot more given the small sample size.

So it would seem like a relatively straightforward and limited scope study could be attempted along the lines of a, you know, traditional immunogenicity safety study that would provide more information to reassure us about this particular vaccine and perhaps also to provide some suggestion of where the vaccine should be targeted given that the supply will be limited and we want to avoid targeting groups for whom the vaccine was unlikely to be very effective at all if there was heterogeneity in the response by personal characteristics.

So I wonder if the FDA could just tell us more about sort of why we are where we are now in regard to that.

DR. BAYLOR: When you say where we are now in?

DR. JACKSON: Why does it -- I mean I'm just curious as to why we seem to be restricted to looking at a data set of 100 people without an option for a more expansive assessment of immunogenicity and safety prior to moving forward with the important step of licensure.

DR. BAYLOR: Again, I'd go back to my point, and I understand your point about the manufacturing process. And if we go back and look at the process, as I mentioned earlier today, for strain change, and we would not require clinical data, we do not require clinical data for changing the annual strain. This vaccine is manufactured by a licensed process.

Granted there is more antigen here, although you are receiving 45 micrograms for -- although it's a single dose, you do receive 45 micrograms in the seasonal, so it's a 15. It's a 45 versus a 90 times 2.

And we believe that the -- you know, that we admit -- we recognize that the data are limited, but I think we really have to keep a perspective what the aim here. We're saying -- there are -- there is no licensed U.S. vaccine for an H5N1 strain. We have a limited amount of data but with a product that is manufactured by a licensed process. And so we believe this data would be sufficient to be submitted for the evaluation for licensure in the case of, as we said, during a pandemic or, as we've said, individuals who may likely be exposed to an H5N1 or go into a region with an H5N1 as we've heard from our colleagues from the DoD.

So, yes, the data are limited but we -- the data are supportive -- I believe the data is supportive of at least demonstrating that this vaccine, based on the licensed manufacturing process, doesn't elicit any undue concerns. But that's what we're asking you. We're asking you based on this limited data set, what are your recommendations.

I mean we really want to know your comfort level in using this vaccine that's been studied in a limited population but considering how this vaccine will be used and when it will be sued and the fact that this vaccine is -- will not be commercialized. This vaccine will not, more than likely, be distributed to the entire population in the country. So within that realm, we're asking you for your opinion and recommendations and advice.

DR. KARRON: Dr. Couch first, then Dr. Farley.

DR. COUCH: I guess I just generally want to speak for supporting this, and a lot of what we would like to have, I'd say, I'm not sure that I would strongly recommend using the resources that would be necessary to get it. I think -- I don't know a whole lot about the H flu vaccine -- H5 vaccines that are out there, but there are others coming along. And if you read the press reports, they're going to be better than this.

But this is the vaccine we have before us for licensing, and if you're a practicing physician, there's a whole lot of difference in using a licensed preparation and using an unlicensed preparation and even unapproved use. And so this is -- I would think of this almost as a step one. This is not the solution to the H5 vaccine problem, but a step one in moving that direction.

And I would also like to put in the point that I would strongly support keeping the five-dose vials, because that gives us flexibility. You see? And if -- I would not be unhappy at all as the physician if I was required to make this my first priming dose to be followed by the clade 2 vaccine if that's the supply required me to make that kind of decision and use it in that way so that we've got an approved preparation that is less than desirable to what you're hearing a lot about around this table, and nobody would differ with that. There are other preparations that are coming along that the press reports say are better, and there's actually some of it published.

But this is a step one to move us in that direction. And Dr. Baylor has been saying, you see, this data is with a preparation that is using a manufactured processed, and it is an immunogenic preparation, and we don't know how good it -- we know it'll work. We just don't know how good it will be.

DR. KARRON: Dr. Farley?

DR. FARLEY: Well, I just wanted to comment that, in response to Dr. Couch's statement, that I think physicians are much -- they have a comfort zone of using a licensed vaccine because it has gone through a rigorous standardized process that has, you know, set, you know, a high standard. And I think that we would probably all agree that this is a special circumstance quite different from our usual rigorous high standards and high expectations. And that's sort of the issue here, and that is, you know, is there any mechanism for licensure that is, you know, has some sort of qualifier that it is licensed for non-commercial use in special circumstances. I mean are there different levels or circumstances. Or in our pandemic planning, should we be setting some policy that would allow for -- you know, that we are -- we would be designating this?

I mean otherwise, our concern is that we're licensing it and it's just licensed like every other vaccine and we've given it a seal of approval that says it met, you know, a standard that I'm not sure we're able to say here but we understand why. And I'm supportive of the idea of having the stockpile and being prepared, but the standard of licensure is this question of are there gradations.

DR. KARRON: Dr. Webster, did the FDA want to say anything in response to that or? No. Okay. Dr. Goodman?

DR. GOODMAN: Yes. I -- you know, one comment I would make is that the original concept here is this vaccine was developed. I mean the thing that is different about this vaccine, every year a new influenza vaccine is made using this process by this manufacturer. And, you know, tens of millions of doses are used. And there's a very well-established record there.

So what's -- in fact , the original intent is that if somebody used a licensed manufacturing process -- and in the original draft guidance -- that we would require some clinical data to be -- establish the dose and immunogenicity but it wouldn't necessarily require a new license. Okay. The issues here are that dose is somewhat higher, but there was not an anticipation that this is as if this is an entirely new vaccine So I think it's important to put that out there. Because it is intended for this specific use, because it could potentially be available when an annual vaccine is available, and it's a different indication, it's gone down the pathway of having -- being proposed for a separate license.

But, you know, just to sort of share our view of what this all came from. So it's not that this is sort of this new product in a vacuum and we're saying just consider it based upon 100 patients of something like this, but this is really a modification of a well-characterized product.

But that said, you know, I'd also like to point out that on -- I mean I think we're here to hear these concerns, and if there's a feeling that additional patients would add to the comfort level here, we should hear that.

But one comment I would make is that, for example, in 1976 with swine flu occurring with the unexpected complication of GBS occurring, there actually were many substantive studies of those vaccines in larger numbers of patients. You would have only detected a rare adverse event like that if you had done pre-licensure studies, let's say, of a million individuals. So the question is what do you get for additional studies and the investment?

We're not here in any way to say they shouldn't be done. We do want input about that. But I just think it's important to give the context that this is not -- that while the dose is twice the level of antigen in the single -- total antigen in the single inoculation and it is less immunogenic, essentially this is the same vaccine. So we welcome the comments, but I think it is, as several people have said, an unusual situation, but --

In terms of your other question, there are other options for how vaccines could be used. And under licensure, there is a so-called emergency use authorization if the national emergency is declared, etcetera, etcetera. But our goal here isn't to lower standards for licensure. What we look at in licensure is does the potential benefit of the vaccine in this situation outweigh the risk, and what you're hearing here is that an indication is for not going out and immunizing the entire population tomorrow but being prepared if there were a pandemic to immunize people or to immunize people who potentially are at a high risk o exposure to the avian virus or a virus that's transmitting among humans.

But again, to go back to it, we do every year with a vaccine that is only minimally different from this under licensure -- the annual vaccine, tens or fifty million people are immunized annually. So there's a context with the existing vaccine that gives us information that's applicable to this. Let's say that it's information that we would consider, and then there's a context with the proposed use and the risk situation that needs to be considered.

DR. KARRON: Dr. Webster?

DR. WEBSTER: Thank you. I would like to follow-up the comment to Bob Couch and support his attitude regarding acceptance of licensure of this vaccine. It's an interim vaccine. We have to look upon this as an interim vaccine. It does not meet the standards of the seasonal vaccine.

But this is an avian vaccine, and we don't know what the correlates of protection are. We will not know what they are until the pandemic comes. And with the numbers that we have, this is a very serious situation. The modelists tell us that this virus, if it does acquire human-to-human transmissability, the first wave will go through in three months. When are we going to prepare the 600 million doses of vaccine? We need this pre-pandemic stockpile, and we need to use it to determine whether we can prime people.

There are many additional things to do with this. We're only at the beginning, and it worries me that if we don't like this one, there are more better ones in the pipeline. What are the consequences -- my question is what are the consequences if we don't license this one, acknowledging it's the best we've got and it's not all that great?

DR. COUCH: Just a quick comment that I think you should say we'd rather not use this vaccine, but if we have to use it, a vaccine for H5, we'd rather have a better vaccine. But that's all out in front of us.

DR. MODLIN: I just wanted to say I concur with both Dr. Couch and Dr. Webster on the issue at hand. And secondly, I want to both acknowledge and express my appreciation to both the agency and the sponsor for bringing this forward in a pubic way. I think it's -- we all are agonizing a bit over the uncertainties and clearly agonizing over what we would also consider, I think, a disappointing immunogenicity for this vaccine.

But nonetheless, I think it's very, very important that the press and the public hear this and this open and transparent way, not only to understand the uncertainties and the anxiety but also to understand the progress has been made and look at the progress we intend to make over the next few years. So again, I want to express my thanks.

I -- even though the immunogenicity is disappointing, I'd just like to point out that it's in a range that's not a whole lot different than the efficacy that we already recognize from inactivated vaccines in very young children, even with two doses. And these are vaccines that are already licensed down to age 6 and so that there is precedent for using licensed vaccines in a certain population already.

And then finally, I'd just like to say that I think that the consequences of any other action would provide -- would be far worse. I think the last thing that we'd want to do would be to discourage manufacturers from collaborating with public health authorities in all of this. And if we were to set a roadblock at this point in time, I think there would also be a considerable concern that we would complicate distribution of a experimental vaccine should it need to be used. So I would be very supportive.

In terms of safety, there may very well be that we can learn a little bit from our experience from just a few years ago with smallpox. Smallpox vaccinia, when we were considering how the vaccine would be used, we were identifying certain groups that would be prioritized to receive vaccine much in the same way that the influenza plan is unrolling. And it was pretty clear that we weren't going to be able to anticipate all of the adverse events that may arise.

At the time, there was aa group formed by the ACIP. The CDC and the Department of Defense, the Armed Forces Epidemiology Board collaborated on basically a safety monitoring board that was put -- that was established at the time that the vaccine program, the vaccinia vaccine program was being rolled. And it actually turned out to be a very effective step in that this was a group that was able to establish thresholds for concern for certain adverse reactions, to set up surveillance mechanisms for monitoring for unanticipated adverse reactions. And indeed, they did come along such as the myocarditis that occurred with vaccinia that was completely unanticipated. And these were dealt with in a rather effective way.

And so maybe some sort of a similar mechanism case would be very useful as we're planning to deal with these pandemic influenza vaccines.

DR. KARRON: Thank you. Actually, I think that's a good segue also into what I'd like to do now which is o have a larger discussion on if this vaccine were to be licensed, plans for post licensure monitoring of safety, immunogenicity, effectiveness. In listening to the earlier presentations this morning, I was a bit struck with the probably exception of the DoD, there's a bit of a mismatch between our existing monitoring systems and the population most likely to get this vaccine, so that our monitoring systems are largely targeted at children and we're talking about vaccination of first responders. So maybe we could have some discussion of what kind of information we would want to collect and what kind of systems might help us collect that kind of information. Dr. Wharton?

DR. WHARTON: Following up on Dr. Modlin's comment about the smallpox vaccination program, I do think there may be some lessons learned from that program in terms of under those kinds of extraordinary circumstances what kind of safety monitoring systems, in fact, work.

And I -- at least my impression is that, really, the enhanced passive surveillance system that was implemented as part of that program was effective at identifying the unexpected severe adverse events which occurred as part of that program, because this is not -- this vaccine is not going to be administered, in all likelihood, in anything remotely the same as the seasonal vaccine program. It's likely to be administered in special clinics. There's an opportunity to provide some special guidance information, facts sheets and so forth.

So I do think it's possible to do an enhanced passive surveillance system that I expect would be much more effective than our usual approaches to passive surveillance as part of the influenza vaccination -- as part of our seasonal influenza vaccination program.

DR. KARRON: I'd actually like to hear from committee members about their thoughts about monitoring immunogenicity of this vaccine in a larger population, if they think that that would be useful post vaccination?

DR. SELF: I'll pile in first. Yes. So given the limited immunogenicity data that's available to date, having some program to monitor, at least under random sample, immunogenicity in the roll out just would absolutely be key in my opinion. In terms of effectiveness, I think it's going to be hard, if not impossible, to really define in anything but the most crude way what the effectiveness of this vaccine is. And ordinarily, that would trouble me greatly, but the way this is being characterized as a stopgap and as a vaccine that's not likely to have legs for the future, that actually doesn't worry me so much.

The information, though, that I think would be -- that could be obtained that would be really critical would be the relationship between these assays and clinical outcomes. And if there could be a program for -- among the first responders or the high risk of storing a sample so that then one could go back and relate those outcomes and define, at least in some way, correlate protection. That seems to me to be the most critical information around effectiveness that could be obtained from this type of vaccine. I guess I'll stop there.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: I'm going to what Dr. Self just said. Yes. Immunogenicity data would be useful but not as an end to itself, only as supporting clinical efficacy by HAI titer. That much is, I think, a bare minimum in terms of efficacy.

In terms of safety, I don't have a clear idea -- I mean the issues in safety are not so much what we've heard about today in the couple of hundred volunteers because that's a given -- unusual neurologic events and totally unanticipated events such as carditis in smallpox vaccine. I'm not exactly sure how to go about setting it up, and it would require extended conversations with folks from CDC and Sanofi about how to go about this. Because I don't see it clearly right at the moment.

But those two issues, I think, would be important in following licensure and use of this vaccine.

DR. KARRON: Dr. Goodman?

DR. GOODMAN: A couple of comments. I think we totally agree that trying to better define correlates of immunity would be really good, and so I think that's a good suggestion and we should think about if we can practically do some of that. For example, if the vaccine is used for a high risk individual, storing serum to then allow correlate with efficacy.

I would say we shouldn't give up on the idea that I think it will be extraordinarily difficult and, as you said, it will be only gross measures, like does a vaccine protect you from hospitalization or death or something like that. But I think we should think about how we can measure that during a pandemic, because I think a couple of things.

I think, yes, this vaccine could be less effective than we think. It could also be more effective than we think. You know, essentially, if you prime the immune system, it may be that you'll get substantial protection irrespective of antibody levels. You know? It's a race between the immune system and the -- but we just don't know.

It could also be not effective. And I think that could also be true even of next generation vaccines, which -- so we need to be able to evaluate those in a pandemic. I mean -- so I think irrespective, we would welcome, and I'm sure our colleagues, how we'd do that.

And then on the safety thing, I think per Dr. Eickhoff, you know, this is a challenge that the entire medical product industry, the FDA, CDC is facing now, which is how do we detect rare events, how do we determine whether they're really due to a product. And again, I think the challenge, and Melinda Wharton has really defined it, is how do we set that up ahead of time for the populations that are likely to get this early in its use.

Now again, some of that, if there are people who are going to go out and get exposed to human clusters or in these areas, some of that data may be obtained before a pandemic potentially. But I think, again, in a pandemic, how do we get robust detection signals that are oriented towards the first people likely to get the vaccine. And, you know, the Defense Department's is one very good example of where there may be potential to do that. But I'm not sure there aren't other approaches we could take.

DR. SELF: So just to clarify. I wasn't suggesting that the cohort or the case control studies to try to estimate effectiveness shouldn't be done, but just from a practical point of view, if the response rate of this vaccine is 40 or 50 percent, and because this is built a priori, if there is a mismatch with the emerging strain, you know, cut that down by another half or two-thirds.

You know, the prospects of enough efficacy to distinguish from the sorts of selection biases that those study designs are going t have seems just very low. So, yes, you should them, but in my opinion, I think the real value over the long run is going to be in those correlate studies.

DR. KARRON: Dr. Hachey?

DR. HACHEY: Just one point of clarification as far as DoD. Our current concept is not to start providing the vaccine to our active duty personnel as soon as it's licensed.

What our concept is is when the pandemic appears to be imminent, so we have a robust, let's say, WHO base for -- that's -- the writing is clearly on the wall that we're going to be evolving to Phase V and then to VI to a true pandemic, that's the kind of trigger that we're looking at as far as providing the vaccine to our personnel.

So, yes, there would be a window of opportunity to get some of that data, but right now for our active duty members who are, let's say, stationed in high risk areas where there's avian disease but still WHO Phase III, at least at today, our plan is not to provide the vaccine to those folks. I mean that may change as vaccine supplies change.

But right now out concept is when the pandemic appears to be quite imminent, that is our trigger. And with the amount of vaccine that is likely to be available to us, clearly we won't be immunizing DoD totally but fairly select groups. With our current vaccine supply, we have probably enough for about -- I think it's about 700,000 personnel. So a decent end but we're not talking millions.

The other point is as far as looking at efficacy, one thing to keep in mind is that I think a lot of the first responders are probably also going to have access to antivirals. So I don't know how that would kind of cloud the efficacy question.

And just one additional advantage to actually using this vaccine is that many of the folks who are likely to receive it are also likely to be in that top tier for the pandemic specific strain vaccine. If it does turn out to be a decent primer, than that does decompress that top tier and allows you to fill that top tier with the pandemic specific strain much quicker.

DR. KARRON: Dr. Word, did you have a comment?

DR. WORD: It workshop just related to when you were talking about how to gather more immunogenicity as well as safety data and just made me think back of one of the -- the first time that we had a delay in the production of seasonal influenza vaccine. And one of the first things we recognized was that the Government had no control over distribution.

And I think we're in a different situation if this is approved whereas the Government will have it. And I guess I'm looking at it in terms of you have a designated administrative cite. You're sending people there. You can collect the data on them. You'll be able to obtain it and store it. And whereas the Government, you have more control to gain that data.

So the gap may not be as great as or as challenging as we may see and just say this is the first time we will have something nationalized as opposed to looking at what distributor will provide it for us.

DR. KARRON: Dr. Jackson?

DR. JACKSON: My recommendation would be that any program of actual use of this vaccine would incorporate a method to obtain at least a post second dose blood sample for storage on vaccine recipients, because I, of course, think we need to know a lot more about the immunogenicity of this vaccine, plus I think we'd really want to know how response to this vaccine would predict subsequent boosting response to a different pandemic-formulated vaccine. And I don't think we'd want to lose that opportunity by failing to collect blood specimens that might prove to be extremely useful later on.

DR. KARRON: I think if there are no other comments or questions from the committee, we are probably ready for our vote, and I'd like to ask that the first question be productive.

Okay. The first question is are the data sufficient to support these effectiveness of this product for use during a pandemic or in situations of potential high risk exposure. Dr. Modlin, we're going to start with you.

DR. MODLIN: I've read this question over a number of times and recognize that it actually is very well-worded. Somebody spend a lot of time working on that I'm certain, Dr. Baylor, Dr. James. Obviously, the data are not sufficient to give us any confidence with a degree of effectiveness, but the data are sufficient to support the effectiveness, so I will vote yes.

DR. KARRON: Dr. Couch?

DR. COUCH: Despite the fact that Dr. Modlin said he spent a lot of time on the language, I would change it. Are the data sufficient to support a degree of effectiveness for this product? By all means, yes. And I would say yes to that and to the question.

DR. KARRON: Okay. Dr. Cox, I know you're not a voting member for these proceedings. I don't know if you would like to comment at all.

DR. COX: Sure. As a person who sort of lives and breathes influenza and has been involved in H5 preparedness since 1997 and having seen a lot of data over the years and having been in the meeting in Geneva last week or the week before where some additional vaccine data were presented and sort of taking into consideration the real risk that we see for this virus to evolve into a strain that could become transmissible, I would vote yes if I were able to vote.

DR. KARRON: Dr. Farley?

DR. FARLEY: I will vote yes as well. I think the question, as it's worded, really gives us the sense of the special circumstances that we're dealing with that I am very much in support of.

DR. KARRON: Dr. Self?

DR. SELF: I would vote yes as well, although I would also say that the clarification about this being a stopgap, that there is a vaccine and the intended use is as that, although it's not perhaps reflected in the nuance of the wording of the question, with that understanding, I would vote yes. I'd also just go back to the slide 31 from Dr. James' presentation which was the litany of unknown efficacy, unknown correlative protection and all that. I suppose that slide is in there just to give me heartburn.

(Laughter.)

DR. KARRON: Dr. Eickhoff.

DR. EICKHOFF: I vote yes without further comment.

DR. KARRON: Dr. Wharton?

DR. WHARTON: I hope we never have to use it, and I hope if we have to use a vaccine, we have a better one, but this is the vaccine we have now. I vote yes.

DR. KARRON: Ms. Krivacic?

MS. KRIVACIC: I vote yes. I agree with Dr. Wharton as well.

DR. KARRON: Dr. Hetherington, I know you're not a voting member, but would you care to comment?

DR. HETHERINGTON: Just highlight a couple of things that have already been mentioned. This is a vaccine with limited immunogenicity and the response itself reflects some limitation on protection and with a clade change that sounds like it's imminent, if not here, you even have a lower likelihood of success.

However, let's now forget among human cases to date, we've got a 60 percent mortality, and we don't know what that translates to our society, but clearly it's going to be -- this is not a deliberation on a seasonal vaccine, and I think that it really behooves us to approve this vaccine to have something available. So I agree with all the prior comments.

DR. KARRON: Dr. Word?

DR. WORD: I think that as struggled with this initially, I kept on thinking about it as a seasonal vaccine. And then as the more I read, I realized this isn't a seasonal vaccine. So based on that, and particularly the way it's worded, during a pandemic, then I would definitely vote yes.

DR. KARRON: Dr. Jackson?

DR. JACKSON: Well, I don't think the data are sufficient, but given that it's our only alternative, I think it should be made available. And if that requires an answer of yes to question one, then I'll vote that way.

DR. KARRON: So that's a yes. Okay. Dr. Gellin, you're also not voting. Would you care to comment?

DR. GELLIN: Well, I mean as Nancy has taught me in this, that there's nothing about any of this is easy, and every step forward reveals the next, you know, cascade f complexity, and we've heard some of that this morning. So I think that that's -- you know, so I think this is an important step but recognizing that, as has been discussed here, there's a lot about this that is going to require a lot further work in addition to assessing further future vaccines.

But I think, as John Modlin highlighted, I think that the importance of having this meeting can't be under estimated, that if this was just a seasonal vaccine, we wouldn't be here talking about it in this way. But because everybody's got a stake at this, the opportunity to have a public discussion about this, and to have that reported on so other people can consider what we did today is really critically important. So I'm glad that John brought that up, but I think that is probably the most important thing we're doing here today.

DR. KARRON: Ms. Province?

MS. PROVINCE: I am also going to vote yes. I concur with all the previous comments. I think that it is extremely difficult to make what's been called a static evaluation of a risk-benefit analysis in this atmosphere of extreme uncertainty. But -- and part of which, a huge part of which, of course, is just the extreme limitations of the data. And so I echo the concerns that have been expressed here, but I think as a stopgap measure, we really -- you know, the answer to the question must be yes. So I do vote yes.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I concur with the previous comments, and I do think that we can use the data on immunogenicity to say that it does support effectiveness, although it's a limited titer or the amount of immunogenicity is poor, but it's better than nothing. And given the stopgap, we have no choice but to say yes.

DR. KARRON: Dr. Hachey, again, I know you're not voting. I didn't know if you'd like to make a comment.

DR. HACHEY: I concur with the previous comments. I still think it's an important step towards combating what will be a pandemic sooner or later. And if the animal data is somewhat predictive of what we can expect in human models, then this vaccine may not be quite as bad as we think it is.

DR. KARRON: Dr. Webster?

DR. WEBSTER: This is like a child taking its first step, very tentative and necessary to do, and I see this as the very first important step, and the vote is yes. There's a long way to go yet, though.

DR. KARRON: Dr. McInnes?

DR. McINNES: I have confidence in the process whereby the vaccine is made. I think this manufacturing process is tried and tested. I think the NIAID trial showed a dose response curve, so I think the -- it is immunogenic at the 2 to 90 microgram -- 2 doses of 90 micrograms. I don't think this is a bad vaccine. And so I pragmatically accept this package as a measure of a degree of effectiveness. So I vote yes.

DR. KARRON: Thank you. And I would also echo everyone else's votes and say that I think this is an important first step in the development of pandemic influenza vaccines.

We're now going to move to the second question which is are the data sufficient to support the safety of this product for use during a pandemic or in situations of potential high risk exposure. And this time, Dr. McInnes, we're going to start with you.

DR. McINNES: The only safety data we have on hand is essentially local and systemic reactogenicity as measured in the particular NIAID trial. If that is all we have on which to base this and given the confining around the circumstances on which this would be used, then I accept this as yes. When that use becomes broader, I become much more uncomfortable about it.

DR. KARRON: Dr. Webster?

DR. WEBSTER: The data available is extremely limited, but in the face of a pandemic, the answer would be yes at this time.

DR. KARRON: Comment, Dr. Hachey?

DR. HACHEY: Just that I agree with the previous comments and that because of DoD being the way DoD is, we are fairly well-positioned to monitor the safety of this vaccine.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I concur that the data are very limited and difficult, and from that sense, to draw conclusions from, but based on its relationship to the current seasonal vaccines and the manufacturing process, I am comfortable with saying yes.

DR. KARRON: Ms. Province?

MS. PROVINCE: Again, in the context in which we find ourselves with the limited data and the scenario that's been presented to us, or one of many possible scenarios, I do vote yes on the question.

DR. KARRON: Comment, Dr. Gellin?

DR. GELLIN: Thank you. My comment -- again, I won't be voting -- is that the data on this specific vaccine are limited as has been highlighted, but this rests on I don't know how many years of experience with seasonal vaccine for which this is the same exact process. So I think that Lisa's highlighted there are some specific differences here in terms of antigen content that raise some issues, but I think we can't forget the fact that this is built on a large experience of safety information of a vaccine prepared this way.

DR. KARRON: Dr. Jackson?

DR. JACKSON: Yes. I've expressed my opinions about the safety data before, but I would vote yes on this.

DR. KARRON: Dr. Word?

DR. WORD: I'm sorry. I would also vote yes on this question.

DR. KARRON: Comment, Dr. Hetherington?

DR. HETHERINGTON: I have nothing to add.

DR. KARRON: Okay. Ms. Krivacic?

MS. KRIVACIC: I'm having a real difficult time with this one, and I think, you know, part of it is the issue of safety and the fact that this is going to be going into first responders who are healthcare workers, and those people are not necessarily volunteers as we have been testing them in this 452-volunteer trial. So I don't know. This is a tough one for me, and I think I'm going to abstain.

DR. KARRON: Okay. Dr. Wharton?

DR. WHARTON: For the question we are asked, I would say yes, but that's with the understanding that we will make provisions when we're actually using the vaccine to collect additional safety data.

DR. KARRON: Okay. Dr. Eickhoff?

DR. EICKHOFF: Similarly, I vote yes subject to some of the questions that will be addressed under question three.

DR. KARRON: Dr. Self?

DR. SELF: I vote yes as well on this. The -- you know, the balance of risk for use during a pandemic, I think, are -- you know, that's pretty easy to balance out even with the limited data. For the high risk exposure, that application, that causes me a little more of a problem. The case fatality rate is so high, though, that that high risk exposure would have to be awfully low to counterbalance the safety concerns, even given the limited amount of data. So I guess, with that sort of thinking, I would vote yes.

DR. KARRON: Dr. Farley?

DR. FARLEY: And I would vote yes as well given the question and in the setting of the high risk exposure and pandemic.

DR. KARRON: Dr. Cox, an opinion?

DR. COX: Nothing to add.

DR. KARRON: Okay. Dr. Couch?

DR. COUCH: Yes.

DR. KARRON: Dr. Modlin?

DR. MODLIN: Well, again, focusing on the wording here, the data support safety -- are they sufficient? Obviously, no. I guess like -- Bruce Gellin brought this up and I think it's a concern we have to recognize that we're using a lot more antigen than we currently use with the seasonal vaccine. Therefore, it's at least biologically plausible that this vaccine could be associated with a higher risk of adverse events, even though we don't even recognize occurring with a current vaccine. So I think that does raise the issue, not just for this vaccine but also for its successors, if they require high antigen contents, that we need to keep that in mind in terms of designing studies to ultimately try to assess safety issues that we don't fully understand now.

I think it will be critically important to set up an adequate monitoring system, but that's what -- I guess, we'll be talking about that when we discuss the next question.

DR. KARRON: Okay. And I would also vote yes. I think there's a -- the safety data are adequate to support licensure for use in first responders during a pandemic, but as for -- I think about to talk about -- it will be important to establish safety monitoring programs for the target population for this particular vaccine.

Okay. The third is really just asking for comments, and we've had many of these already, on studies to collect additional information about the effectiveness and safety following this vaccine's use. Just to remind the committee of what we've discussed already, I've heard discussion of an enhanced passive surveillance system. I've heard discussion of a monitoring system similar to what was set up with smallpox vaccine use. I've also heard discussions about a need for collecting immunogenicity data, particularly as it relates perhaps to being able to get some sense of correlative protection. And I'd just like to ask the committee at this point in an open way if there's anything that they'd like to add. Dr. Wharton?

DR. WHARTON: I think that the Department of Defense is uniquely situated to provide information in a timely way once the vaccine begins to be used. I understand it won't be used initially, but at the point it is used, it seems that DoD does have some infrastructure and capacity in place that can provide really important information for the whole country. And in this particular circumstance, giving the likely initial vaccinees, I think the DoD population is less different than it sometimes is for other vaccines from the population and the civilian sector that will be a target of the vaccination program. So I think there -- DoD has the potential to teach all of us a lot about this vaccine in its early use.

DR. KARRON: Okay. Yes, Dr. Webster?

DR. WEBSTER: The other topic that was raised was the question of prime-boost. It's, I think, very important to consider additional work on prime-boosting, particularly with the development of clades and new sub-clades, whether the clade 1 will prime sufficiently. I think it's a very important question and to go ahead and boost or prime a substantial number of people to find -- to answer that question.

DR. KARRON: We'll have some of that discussion this afternoon. Other comments? Dr. Farley?

DR. FARLEY: We didn't spend much time on this, but I -- given the low rate of Guillain Barre Syndrome and other unusual neurologic complications, and the point being made that you couldn't do clinical trials of size to pull out, that is a safety issue.

But -- so I would just like to encourage the fine-tuning of use of large population-based data sets, electronic data sets that can pick up signals on, you know, where you have a large denominator and validating the use of those before the onset of the pandemic, so we can say that we are able to pick up trends and changes in low incidence diseases that might be associated, so that we're in a good position to know what the baseline is, and then it could be used in a setting of the onset of a large scale use of these vaccines in the setting of a pandemic.

DR. KARRON: Dr. McInnes?

DR. McINNES: It seems like this setting up the surveillance system is going to be intricately linked with actually specificity around who are these high risk -- these populations and these first responders. And so I would urge that that move forward in a very active way and that that be clearly articulated. And then the plans can be more specifically more tailored.

And in fact, one might even be able to hierarchically rank where you're going to lead with this from and not have everybody simultaneously -- but I can imagine that if these are not put in place, then you're going to be playing catch up all the time and not have an adequate system in place. But until the people are defined, you can't really design a system.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: I would simply caution that these kinds of post use studies, desirable though they be, are going to be conducted in the early stages of a pandemic in all likelihood. And whoever plans these trials or these follow-up surveillance studies needs to be very much aware of that, because it could be pretty horrendous, I think, carrying out some of these surveillance studies.

DR. KARRON: Dr. Self?

DR. SELF: So that point, to the extent that some of the characterization of immunogenicity and safety could be done in studies before we get to that chaotic situation, I think that would be very, very useful. It also might help in defining some strata that could be helpful in looking at measures of effectiveness as well.

DR. KARRON: If there are no other comments, I also would like to echo Dr. Gellin and Dr. Modlin's earlier comments and actually thank the FDA for bringing this topic before the committee. I think having an open public discussion of this issue is really important. We're going to adjourn for lunch. We will reconvene at 2 p.m. for the afternoon session. Thank you.

(Whereupon, off the record at 12:43 and back on the record at 2:07 p.m.)

DR. KARRON: I'd like to call the afternoon session to order, if people would please take their seats. Our first speaker this afternoon is going to be Dr. Jesse Goodman from the FDA who will introduce the topic of Clinical Development of Influenza Vaccines for Pre-pandemic Uses.

DR. GOODMAN: Okay. Good afternoon. My purpose here is to frame a discussion that is sort of, to some degree, an opposite place of where we were this morning in talking about what would one do in evolving emergency, etcetera, to what are some of the issues involved in potential pre-pandemic use of pandemic vaccines and to get input from the committee on issues like priming and how to do studies, etcetera. And I'll just -- we realize this is a huge issue. It requires much more time than there is here, but the point is to begin to get your input and to begin to have people thinking about it and just to say this -- we are having sponsors, etcetera, now consider some of these issues, so this informs our dialogue with them.

Now where are we right now? Well, it's very important to re-emphasize, and I think this is a huge issue for pandemic preparedness, emergency preparedness in general, that we live in a world of uncertainty. And I think it is very important -- you know, I -- we get asked by colleagues, family, reporters, etcetera -- it's very hard to calibrate the message somewhere between the sky is falling and there's no problem. We don't deal with those calibrated messages. But, in fact, here the probability, timing, severity, and identity of a future pandemic are unknown. But the reality is that I think H5N1 starting, as Nancy put back in '97, is a bit of a wake-up call. This is out there. It persists. There are more deaths. And there are other sera types out there that could emerge.

There has been -- there's the possibility that we would observe evidence of increased human-to-human transmission, perhaps with relevant genetic or antigenic in the virus. There's a possibility that we would observe that before a pandemic, but waiting until such evidence occurs may leave very limited time to have a vaccine produced and available. Okay. And certainly these uncertainties complicate our planning.

Now why even consider the possibility of immunization strategies that are prior or early in a pandemic, and I think the prior is a harder one for people to get their arms around, but early is the biological relevance is similar. Well, I think everybody knows that with current vaccine technologies, production times are fairly long. We've done a lot. Dr. Webster mentioned reverse genetics can speed this a little bit. The companies are very efficient at this. But it's still talking tree to six months at the lower end, absolutely everything goes perfectly with current methods. And that's for the first vaccine to come out. And then, of course, capacity to ramp up for the population, that magic, hundreds of millions of doses, is limited. And then when you consider global vaccine needs as has been so poignantly pointed out recently by countries who have no vaccine capacity, this is a huge problem.

Now the stockpiling you've heard about has provided at least the potential flexibility to consider early use, and we heard this morning this structured around well, if there started to be evidence of human-to-human transmission, etcetera.

Okay. There is evidence, and we're not going to have time to review it today, but there is emerging evidence, and I'll talk a little more about it, that priming and cross-protection can occur just like with annual flu strains among so-called heterologous H5 strains, in other words isolates from Vietnam versus Hong Kong that are H5N1 or even some of the more diverse isolates like among the clades. And then as we discussed a little this morning, modeling suggests benefits to the early use of a vaccine, even one with fairly limited efficacy and potentially even in single doses, and I'll show you a little bit about this.

Well, what are the things we should be doing to remediate the situation, and some of this fits in with what Robin Robinson talked about this morning. So we're taking efforts along with out colleagues globally and at CDC and industry to make strains and reagents and testing processes which, really, you know, many of which are very old methodologies. The reverse genetics is one example. But to make this happen faster, we are all exploring dose-sparing strategies which not only might help overcome this rather poor immunogenicity of this antigen but could obviously make more doses be available more quickly.

Could there be ways to more rapidly induce immunity? Well, as for that, enhanced cross protective properties, the next line, there is some suggestion that perhaps live vaccines could be helpful in this respect. There's a lot of early studies about conserved genes. So these are all things that can make us get ready faster. Scalable rapid production methods -- there's a lot of U.S. Government and industry investment in cell culture technology that doesn't get it done a lot faster, probably not faster, but it may have certain advantages and scalability as would, of course, recombinant strategies. And then there is a lot of investment which the world, not just the U.S. needs, but the world needs in manufacturing capacity, and in this country, in stockpiling.

And I will say I always like to take the opportunity of the Bully Pulpit in a sense to say I think we should not only think about that it's not just H5N1, it could be another strain, could be H7, could be H1, whatever, H2, but we think about it may not be influenza either. So our public health response capacity, we should think about how we leverage this in general.

Okay. So what are the approaches to sing a pandemic vaccine? What are the different timings? Well, we talked most about, and I think there's the highest comfort level with during a pandemic so there's a very clear benefit risk. So we heard this in the opinions ventured this morning. But frankly, as was indicated in 1976, even with a proven vaccine, a strain could have an uncommon or unforseen adverse event, and we need to be very transparent with the public about that. You know, you could do the best studies in the world, but if something occurs in 1 in 50,000 people, we have to be prepared to deal with that. And the biggest con here is that it's just simply too little too late, and I'll get to that.

Then in an emerging pandemic, this is sort of what we talked about again this morning, vaccination could begin if you had a stockpiled vaccine. You could target individuals such as were discussed or geographic areas such as a country if human-to-human transmission began to emerge somewhere. And these may be effective strategies both in saving lives but potentially in pandemic control. And as I mentioned, even at reduced efficacy, models predict benefit. And this could be a temporizing strategy, again, as discussed today, until a matched vaccine was available.

The benefit-risk ratio is clearer than in pre-pandemic use, although, again, as we saw in 1976, sometimes our ability, even as events unfold in front of us, to predict whether a virus will become pandemic is not -- you know, it's not a fine-tuned ability. And of course, ths stockpiling is quite expensive, and there is the potential need, based on either stability or change in the virus, to replace or rotate stockpiles.

Now what about pre-pandemic immunization? This could be potentially considered as an option separately from or as part of annual immunization program. And this could either be done to individuals who are perceived as having increased risk either of bad outcomes or potentially, again, as discussed this morning, in an emerging pandemic, of early on exposure.

If successful -- if you have what makes a pandemic a pandemic, because you don't have population immunity and individual immunity so it's successful, such strategies could potentially blunt or maybe even prevent a pandemic. And there are obviously human and economic benefits. It requires less search capacity and could reduce the need for a number of emergency measures and stockpiles.

But the problems are ones that everybody is familiar with. You could immunize people and have a completely different strain emerge, so if you don't have strong cross-protection, there may be a mismatch and limited efficacy. And of course the biggest one is the uncertainty of whether a pandemic will occur and, if so, what it would occur with, so you're measuring a potentially small risk from a vaccine safety issue against what essentially is a hard to predict benefit or unknown benefit. It's the unknown risk of a pandemic.

I'm trying to find -- I had another slide that I thought I had in here -- oh, there it is, but it -- look at that. See, there's a slide that refuses -- I'm going to -- well, I bet you when I put it up, it won't -- yes, I've encountered this once before. There's some kind of control that changes the slide, but I'm just going to show this. I think this is sufficient.

But this was the point I had meant to make early on which is that if you look at vaccine production capability and this is preparing the seed, making the monovalent, filling and testing it, and then under this rather aggressive scenario, let's say in four months your vaccine begins to become available, you can see here that in this crude presentation of the first wave of a pandemic, basically that first vaccine is becoming available as the first wave is receding. So this is not a highly effective immunization strategy in dealing with that first wave. And then you consider that, you know, it's probably going to take you a week or two to get meaningful immunity, even limited immunity, from a first dose much less to then come and administer a second dose a month later. So this was the background that I wanted to share before that.

And I knew I had that there. I once had that happen to me in front of about a thousand people, and I had about eight slides like that that wouldn't let me show them.

So that's the pre-pandemic issue. So the big issue here is the uncertainty of the risk of a pandemic. But the big benefit is if people -- you know, this is a viable strategy to get immunity into the population as opposed to six months after a pandemic starts.

Okay. So what is the background in terms of priming and cross-protection? Well, we know that natural infection provides long-term protection against that strain, invariable but sometimes surprising degrees of protection against related strains. We know that inactivated vaccine provides some protection also beyond one flu season, even though we see, as you saw today, the way the antibody levels tend to fall off in the 6 to 12-month period after immunization and also against related strains. And there are some recent randomized controlled studies that, in fact, show protective effects against fairly drifted strains of annual vaccine. And this may be increased with live attenuate vaccines because of the nature of immunity that they induce and probably also because of the presence of additional conserved antigens, etcetera.

Now what really is moving this field and I think will inform it considerably is that, for example, as Rob mentioned this morning, there's preliminary animal serologic and clinical studies of H5 that do provide evidence of variable degrees of cross-protection between heterologous strains. And there's also evidence that some of the novel adjuvants may boost that cross-protection. We don't know whether that's because they induce a different immune response or just a better, more robust one. But there is some evidence for many of them, for example, that they may prime cellular immunity, etcetera. And there's also a suggestion that priming, and Dr. Treanor, I think is going to present some of his data, may in fact be possible and durable with these H5 antigens.

The big caveat, though, here is that the predictive science, again, is not great here. It's not clear how well-matched strains need to be, to what degree serologic studies or molecular studies could predict this. But I think the science is advancing there, again, as we get a lot more sequence data and start to correlate that with serologic data and animal data, and Nancy has done a lot of work in this area.

A very important point, and came up this morning, is certainly that the surrogates for protection are also not well defined and the assays are highly variable. So when we go to a meeting hearing about multiple different vaccine candidates but they've all been studied with different assays, you have to really take that with a grain of salt. And at WHO we had a recent discussion. There was general agreement of the desirability for standardizing these antibody assays so we can compare one vaccine to another better. But until then, everybody needs to take this with a grain of salt.

Well, there were questions about the modeling, and as I mentioned this morning, to me, a model is just a model. There are many, many assumptions, but it's worth looking at this a little. So this is -- there are two groups that have actually had surprisingly similar results looking at it form slightly different directions. But this is the one group of Ferguson. There's some more things from Longini's group. This is very complicated but what I would say is that over here it shows, for example, the effects of a delay in vaccine availability.

Now if you look at that backwards, it shows you what are the effects of having vaccine ready right at the beginning of a pandemic. And here is where vaccination at a certain rate -- I'm not remember it offhand -- I think it might be a million doses a day -- begins within days of the pandemic versus 30 days versus -- no excuse me -- day 30, 60, blue or 90.

Okay. So if you begin right away, cumulative attack rate of 1 percent. If you wait even 30 days -- second bar is -- beginning on day 30, 1 percent; 60 days, 13 percent; 90 days, 31 percent. And this is a high transmissible virus and a medium transmissible virus. But you can see the dramatic increase in disease or conversely the dramatic decrease through immunization becoming rapidly available to the population.

Okay. This is a pre-immunization strategy. Again, this is very complex. But this shows if you could pre-vaccinate 20 percent of the population with a low efficacy vaccine, in this case 30 percent reduction in susceptibility, so this would be half as effective as we might expect an annual vaccine to conservatively be, and just give one dose here, added to some other measures like household quarantine and some antiviral measures. But the addition of this vaccine policy, pre-immunization of 20 percent, you can see results in a fairly dramatic reduction when added to these other strategies and also overall.

Now, again, I'd like to stress that this is modeling. It's very dependent on the infectivity of the virus. It's very dependent on multiple other assumptions. But when you think about this, if people -- people cannot only potentially benefit from being personally protected, but if they then transmit virus less, this is how you begin to have population impacts of immunization.

However, there are many, many unanswered questions, and that's why we're beginning this dialogue. Certainly, how can we better measure and predict both protection and cross-protection. We discussed that this morning, and it's very important to mention that for a pandemic vaccine, hospitalization, death, in fact, infectiousness may be much more relevant and even more immunologically achievable measures than just infection.

Animal models were beyond our scope but we heard how we can't tell from these possible correlates, you know, how effective a vaccine -- can we develop animal models that tell us more? Can we use them and correlate them with immunologic correlates? What dose and dose intervals are need for priming or boosting, and Dr. Toerner is going to discuss this some. How durable will priming be, and Dr. Treanor's results may help address this. Are specific levels of antibody needed? Must they be maintained.

And then we come down to -- and again, we heard some data suggesting this at WHO recently is people are looking at some cross-serologic cross-reactivity data, that this may be very not just sera type specific or clade specific but even virus specific, so we're seeing some viruses behave differently. In general, the more genetically-related you are the better you cross protect, and I think that's safe to say. But there are some things that may defy prediction, although Nancy may have more to say about this.

So what we have here is data needs. We have a possibility for something potentially effective, but we have data needs. And these pertain to either early pandemic use such as the possibility we discussed this morning with the vaccine under discussion, but it also pertains to potential pre-pandemic use. How should we design clinical trials to evaluate immunogenicity? And there's an opportunity here now. I think this is a very important point. Lots of companies are studying lots of vaccines. There's an opportunity to build into these studies cross-protection studies, immunization with one vaccine followed by another. And these are studies that take a long time and cost a lot of money, so getting them right and getting your input and getting FDA's input and companies' input, there's a lot data that could benefit a lot of people.

What should our standards be? You saw how you wrestled with that this morning. These are very arbitrary numbers. Beyond such arbitrary numbers, what should we be thinking about? And certainly, these scientific data, public discussion are important in considering any of these types of vaccine use. And certainly a substantial safety database will be needed in considering pre-pandemic approaches. And I think that drove a number of the questions today, even about a vaccine using a very well-characterized manufacturing method. And certainly when you consider novel vaccines, this is something we're not here really to discuss in detail today, because it's a whole discussion, but we wanted to bring up.

And so we both want to further the data-gathering process and begin these kinds of discussions, because I think the successes in vaccine development , actually, the good news is that they're going to bring these questions to us. And the sooner we start trying to effectively get the data, the less likely we'll be to be scratching our heads quite as much as we did this morning.

So thank you very much. And now I guess we'll leave discussion for later, but Dr. Toerner's going to follow-up with this.

DR. TOERNER: My name is Joe Toerner. I'm a Medical Officer in the Vaccine Clinical Trials branch at the Division of Vaccines and Related Product Applications. My goal today is to enhance what Dr. Goodman had just mentioned to you and provide you with a bit more summary data on the rational for why we might consider an influenza vaccine to be used in the pre-pandemic setting, also provide for you hypothetical clinical development scenarios that would help your discussion this afternoon on a determination of efficacy and safety.

We're all familiar with the current situation with influenza H5N1. The host has expanded outside the avian species and to date, there 278 cases that have been reported to the World Health Organization of human infection, and the case fatality rate is greater than 60 percent which highlight the nee for urgent vaccine development.

Dr. Baylor this morning had mentioned that the U.S. Government is involved in influenza pandemic preparedness, and as our part of helping develop a plan for use of an influenza vaccine during a pandemic, we issue draft guidance industry on the clinical development of an influenza vaccine to be used during a pandemic or in situations of potential high risk exposure.

And in the guidance document, we had outlined immune response criteria that are reasonably likely to predict clinical benefit, and that is the hemagglutination inhibition antibody assay in different proportions, a fourfold increase versus the proportion greater than or equal to 1 to 40. We're in the process of revising that to further clarify that it would be either or to demonstrate an immune response reasonably likely to predict clinical benefit.

As well, our guidance documents outline safety database requirements, and those differ based upon whether or not a sponsor has a long-term manufacturing experience with a seasonal influenza vaccine.

However, the discussion this afternoon, again, as Dr. Goodman pointed out, there are many limitations to the production and ultimate availability of a pandemic vaccine in a pandemic situation. So what we're interested in is your feedback on a different strategy of use of a vaccine prior to a pandemic or the so-called prime-boost or cross-protection that might be demonstrated in adequate and well-controlled studies.

And so for discussion today, we'd like some feedback on the adequate and well-controlled studies of clinical trial design endpoints and the duration of those clinical trials as well as some discussion of what we might require as a size of the safety database.

So the issue of priming is illustrated in the pediatric population where children are felt to be naive to the seasonal influenza antigens in circulation. And it's for this reason that children below nine years of age who are receiving influenza vaccine for the first time, two administrations of vaccine approximately one month apart are recommended for adequate immune response.

Two recent studies had evaluated the prime and the boost dose that were given approximately six months apart and whether -- and these were seasonal influenza vaccines --and whether the seasonal influenza vaccine remain the same or whether the seasonal influenza vaccine had differed in the antigen content, there appeared to be similar immune responses when these children had vaccine administration more widely separated in time.

And following my talk, Dr. Treanor's going to provide an overview of immune responses that were observed among study participants who had the remote administration of an H5 antigen that also illustrate this concept of priming.

Data from observational studies indicate that prior antigenic experiences protect or ameliorate influenza illness, and this appears to be true even in individuals who have had antigenic experience with the same influenza virus in circulation or whether the influenza virus in circulation represents an antigenic drift.

And as well, seasonal influenza vaccines appear to offer cross protection against antigenically drifted influenza strains, and this has been demonstrated in two culture confirmed studies, one more recently published in the New England Journal of Medicine where vaccine efficacy appeared to be greater than 70 percent for -- even when influenza virus in circulation differed from the vaccine strain.

When considering pre-pandemic vaccination, it's been mentioned earlier in today's session that we'll have to consider the risks associated with administration of a vaccine that may or may not have potential benefit. The risks and benefits for a seasonal influenza vaccine are known. For example, the Institute of Medicine performed an exhaustive review looking at neurological adverse events following administration of contemporary seasonal influenza vaccines and found that data do not support an association between administration of seasonal influenza vaccine and the development of Guillain Barre Syndrome with the exception of one observational study in one year in the early 1990's where it was estimated that the risk of Guillain Barre Syndrome appeared to be 1 additional case per 1 million person vaccinated.

So currently, the Advisory Committee on Immunization Practice, in their publications, indicate that if an association exists, it's estimated to be a risk of 1 additional case of Guillain Barre Syndrome per 1 million persons vaccinated, and therefore the risks versus benefits of a seasonal influenza vaccine are well balanced.

However, we have the historical experience with the swine flu vaccine in 1976 where the Institute of Medicine found that the data did support an association with administration of that particular vaccine and Guillain Barre Syndrome where the risk of Guillain Barre Syndrome was 1 case per 100,000 persons vaccinated.

And so for consideration of rare serious adverse events, they become highlighted when there is the unknown potential benefit.

And now I'd like to move on to a hypothetical clinical development scenario, and this is a straightforward slide that demonstrates clinical development that might occur for demonstration of adequate immune responses to be used during a pandemic or a high risk situation. However, for pre-pandemic use, in order to demonstrate this issue of priming, administration of the vaccine more widely separated in time might provide adequate data that would begin to support the concept of homologous immune protection over time. And similarly, in order to gather data on the issue of cross protection, we outlined a clinical development scenario here where individual cohorts would receive a monovalent influenza vaccine that represented a different clade. And immune responses following the administration of subsequent different vaccine would be evaluated for evidence of cross protection.

And so for discussion this afternoon, we would like for you to comment on the use of immune responses in order to determine efficacy of a vaccine to be used for pre-pandemic use or for priming and is the immune response assay following the prime alone adequate to support demonstration of pre-pandemic use. Should we require that immune response assays be obtained following a boost at future time points, or in particular, for administration of a heterologous antigen in order to gather information on cross-protection? Or can we rely on immune response measurements following the prime using assays that involve the heterologous antigen? And if there's time, to comment on the use of hemagglutination inhibition antibody as an immune response endpoint versus other immune response assays such as microneutralization.

And related to this is clinical trial design, if you could discuss the feasibility of requiring longer term clinical studies of the prime and the boost in pre-pandemic vaccine development and what would the duration of such a study be -- six months out to one year or perhaps longer than one year. And we recognize that this might require collaboration among different sponsors who may or may not be developing monovalent influenza vaccine strains with a different clade.

And for a safety database discussion, we wanted you to keep in mind that for a pre-licensure safety database, it is not likely that we will be able to detect rare or serious adverse events, for example, at a rate of 1 per 100,000. So keep that in mind in your discussions when we ask you to comment on the safety considerations for licensure.

As well, we believe that safety database requirements may differ for a sponsor who is seeking licensure of a vaccine with a novel manufacturing process or the use of a novel adjuvant. And so that will conclude my talk, and I'd like to turn the podium over to Dr. Treanor.

DR. TREANOR: Okay. Thanks. What I'm going to present is really the same presentation that one of our infectious disease made at IDSA in October, and it's basically an analysis of the immune responses of healthy subjects who received a single dose of the Vietnam virus after they had previously received vaccination with an H5 vaccine for the A Hong Kong/156/97. And this was presented at IDSA by one of our ID fellows, Nega Gogi.

So as we heard this morning, the non-adjuvanted, inactivated, subvirion vaccine required two 90 microgram doses to elicit a neutralizing antibody, and we knew that in a previous study that was done by Iain Stevenson and Karl Nicholson, they looked at a vaccine for H5N1 based A/Duck/Singapore, which is a low-pathogenicity avian virus that is antigenically similar to A/Hong Kong/97 that was used to make a subvirion vaccine back in 1997 and which was then evaluated either with or without MF-59 in a study that was published in early 2000-2001. That study showed, as you know, that adding MF-59 had a very significant dose-sparing effect on that vaccine, but that responses to the unadjuvanted vaccine were relatively low.

They took a subset of those same subjects and revaccinated them 16 months later with those who had received unadjuvanted getting unadjuvanted vaccine and those who had received vaccine with MF-59 getting MF-59 adjuvanted vaccine. And what they found was that in both groups, after a third dose administered 16 months later, there was a significant enhancement of the antibody response so that individuals after that third dose achieved levels of neutralizing antibody that were higher than they had achieved after the first two doses.

So pre-priming is a strategy that might generate better immunity. It would potentially allow a single dose in the face of an emerging pandemic which would be logistically, I would think, more feasible, but in reality, using that strategy for an emerging pandemic would probably represent boosting people with an antigenic variant because of continued antigenic evolution of H5 and other avian viruses.

Now we had done a study back in 1997 using a baculovirus expressed recombinant H5 of the A/Hong Kong/156/97. Now despite the chronology of these viruses, the 1997 viruses are referred to as clade 3, the 2004 viruses are referred to as clade 1, 2005 being clade 2. So we took advantage of the fact that there were still many people around who had been in this study evaluating the recombinant H5 clade 3 virus vaccine to bring them back and give them a single dose of the vaccine we were evaluating, the Vietnam/1203/04 (clade 1), to see whether or not there was, in fact, evidence that that previous vaccination with the baculovirus-derived Hong Kong vaccine had primed them to respond to the Vietnam vaccine.

Now just to refresh your memory, when we had done the study with the baculovirus-derived vaccine, we looked a neutralizing antibody which was measured by Jackie Katz at CDC, and we looked a two-dose schedule. We did several time points to look at the kinetics of antibody, but the two doses were separated basically by 28 days. And we found results which actually were very, very similar to the ones that we just presented with the subvirion vaccine in that there was a very strong dose-dependent effect and the best responses were seen when subjects received two doses of 90 micrograms of baculovirus-expressed recombinant hemagglutinin where they generated antibody with a neutralizing GMT on the order of about a 160.

So the objectives of this study were to determine the ability of the clade 3 H5 recombinant vaccine to prime for immune responses to a subsequent clade 1 H5 subvirion vaccine in health adults. Now to do this comparison, it's important to understand that we did not randomize people to be primed or unprimed. People were primed because they had previously been in the study. We simply gave a dose of the subvirion H5 vaccine and compared their responses to the ones that we had seen in unprimed subjects in the 063 study. We also wanted to determine the safety of revaccination.

The subjects then were participants in the previous study which was conducted in late 1997 and early 1998 who had received a clade 3 recombinant H5 baculovirus-derived vaccine at any dose. And these subjects were administered in open label fashion a single 90 microgram dose of the subvirion recombinant A/Vietnam (clade 1) vaccine. They completed the same kind of diary card that was used in study 063. Adverse events were recorded over 56 days, and we tested both serum HAI and microneutralizing antibody against the Vietnam virus on days 0, 28 and 56.

Now the primary analysis here was to compare the results of a single dose in the primed population versus the results of a single dose in an unprimed population. That was our primary evidence of whether these individuals had been primed. If they were primed, they should respond to a single dose with significantly better responses than seen in an unprimed population. As a secondary analysis, we also compared the responses to that seen after two doses of 90 micrograms in 90 subjects.

So this lays the study out sort of diametrically. Here's the original study in 1998. Individuals in that study received the recombinant clade 3 vaccine at a variety of different doses. Any individual who had received vaccine at any does, not the placebo recipients but vaccine recipients, were eligible to participate in the open label study. All of these individuals, and there were 37 of them, received 90 micrograms as a single dose in open label fashion in study 0043 conducted in 2005. So approximately a seven year interval between these two studies. Their results are compared to those which have already been presented in H5 naive subjects who received two doses of the same clade 1 vaccine at 90 micrograms.

Here are the demographics of the two population. The results are compared to, again, that same 90 microgram group we talked about before. This is the demographics of the 37 subjects from the previous study who came back to participate in 043. You can see it's a largely Caucasian population. Slightly more than half of them are female, and the age of these people who are correspondingly about seven years older than the volunteers we typically have in our studies is slightly older than the median age of the people who were in the 063 study.

These are the rates of solicited adverse events within seven days of receiving the vaccine. And it simply shows that the rate of local pain, tenderness and other side effects in those who had previously been primed and received a single dose of 90 micrograms are not different than those which were seen in naive subjects who received 90 micrograms. The rates of all of these effects after one or two doses in naive subjects or after what amounts to a third dose in primes subjects are all essentially the same.

This is the results of the serum hemagglutination inhibition assays, titers following either two doses of H5 vaccine in the naive subjects or one dose in the primed subjects. Now again, these analyses use the 1 to 10 definition as the starting dilution, so everything is on the same frame. And you can see this is the same data that I presented earlier in the 063 study that after two doses of 90 micrograms on day 56, the GMT of HAI antibody is 27.7. This is the 95 percent confidence limits. This is the result in the open label study in primed subjects. And you can see a significantly enhanced response to a single dose and that after a single dose, these subjects actually have higher levels of antibody than those individuals in the 063 study after two doses.

Similar results are seen with the microneutralizing anybody. Again, this is the 063 study. These are the neutralizing anybody responses. Again, using 1 to 10 as the definition of the starting dilution, after two doses of 90 micrograms, the GMT of neutralizing antibody on day 56 is approximately 23. Individuals who had been primed by previous exposure achieved a GMT of 94 after a single dose of 90 micrograms. This is clearly higher than seen after a single dose in unprimed subjects showing that the subjects are, in fact, primed and is actually higher than we're seeing after two doses in unprimed subjects.

And this is a summary of response rates looking at the percent either who responded or the percent who achieved a titer of 1 to 40 by either HAI or neutralizing antibody by day 28 after the second dose in the 063 study or day 28 after a single dose in the 043 study. And again, you can see that the proportion of subjects who respond with a fourfold response is higher after a single dose in the primed subjects than it is after two doses in the unprimed subjects.

And similarly, the percentage of individuals who achieve a titer of 1 to 40 is higher in those who had previously been primed even after just a single booster dose than were seen after two doses in those who were naive at the beginning.

Now one of the things we tried to look at is whether or not it made a difference what regimen a recombinant baculovirus vaccine was received back in 1998. This gets into very small numbers, but it did not look like ti made a substantial difference whether you received 25, 45, 90 or a total of 100 micrograms of baculovirus vaccine back in 1998. Both the GMT of HAI and neutralizing antibody as well as the response rates are fairly similar with the proviso that it's a very small number of subjects in each subgroup.

It did look like perhaps it made a difference whether you responded back in 1998. If we divided those 37 subjects into those who had a neutralizing antibody response against the Hong Kong virus, you could see that those who had responded against the Hong Kong virus in 1998 were somewhat more likely to respond when boosted with the Vietnam virus with a response to Vietnam. Again, none of these differences are statistically significant because of the very small numbers.

Now this is some preliminary data from Dave Topham who's been looking at memory B cell responses. This is an assay that was developed at Emory by Shane Crotty and Rafi Amed and looks at the proportion of all immunoglobulin-secreting cells that are making ELISPOTS specifically against the antigen of interest. And you can see that after vaccination, there is an increase on day seven in the numbers of memory B cells that recognize either the recombinant hemagglutinin of the Vietnam virus or the recombinant hemagglutinin of the A/Hong Kong so that of these go up immediately after vaccination in the 043 vaccinated subjects.

Unfortunately, we don't have the comparison data from naive subjects, so all I can say is that these go up, but we don't know what we would have seen in naive individuals.

So in conclusion, the antibody responses to a single dose of unadjuvanted vaccine support the hypothesis that previous vaccination with clade 3 primes for a response to a clade 1 H5 vaccine. The antibody responses did exceed those that we saw after two doses in naive subjects and actually were somewhat better than those that were seen in the original study. And the reasons for these vigorous responses are not clear. Whether or not individuals have memory B cells that are sitting around ready to respond or whether there's something else, it's just not clear. Revaccination was well tolerated. The side effect profile is similar to that of vaccination of naive subjects.

And clearly, further studies, which I think were nicely outlined earlier, to evaluate different schedules to really verify these results in larger populations are needed. But if you were confident that this would happen, then you could consider pre-pandemic vaccination programs, at least for some populations which we've talked about, potentially healthcare workers, first responders or the military or individuals who would be likely to be exposed and need to stay on the job early in a pandemic.

So I'd just like to thank the individuals who collaborated in the study, particularly Tom Rowe who did all the serology; Mark Wolff and Heath Hill who were responsible for the data analysis; and my collaborators at DMID and at the University of Rochester. Thanks.

DR. KARRON: Thank you. At this point, we'll take questions for Drs. Treanor, Toerner and Goodman. And actually, I'd like to lead off with one question for you, John, which is I was wondering did you see cross-boosting at all? In other words, if you looked at HA responses back to the '97 strain? Did you look at that?

DR. TREANOR: We haven't looked at that and that's partially because we don't have a similarly, at least to my knowledge, have a similarly reverse genetically- engineered low pathogenicity variant of the H5 from 1997 that's easily available to work on. So although that's in the cue of things to do, to my knowledge, that hasn't happened yet.

DR. KARRON: Dr. Jackson?

DR. JACKSON: I just wondered -- I mean it sort of raises the question on how much antigen you need to produce the boosting response, so I was wondering if that was something you're exploring?

DR. TREANOR: It's very possible that it requires less of a dose to prime someone than it does to generate antibody, and so some of the lower doses might be effective for boosting, and that's something we're eager to look at.

DR. KARRON: Dr. Modlin?

DR. MODLIN: John, I know your study didn't look at this, but I think in terms of future studies, it might be real interesting to look ant see what the kinetics of the response is to that third dose. We're talking about trying to prevent a disease that can kill within days after exposure with a short incubation period and so that knowing how quickly you induce protective antibody with that third dose, all the studies that have been done very recently with meningococcal disease and meningococcal conjugate vaccines for that, but the timing of that response to a booster dose appears to be critically important in preventing disease. So in terms of future studies, I think that would be something that would be very interesting to look at.

DR. TREANOR: I agree. That would be very important.

DR. KARRON: Dr. Webster?

DR. WEBSTER: Really the same question that Ruth asked, did you find any evidence of original sin?

DR. TREANOR: Well, unfortunately, we have not assessed any type of immune responses to the 1997 Hong Kong except for the memory B cell responses which do appear to recognize the Hong Kong virus.

DR. KARRON: Dr. Cox?

DR. COX: Yes, John. I think I know the anser to this, but I just wanted to be absolutely sure. Were the serologic tests done at the same time for the two groups --

DR. TREANOR: No.

DR. COX: -- or were they done at different times?

DR. TREANOR: The data that -- the comparison in the naive subjects is exactly the same data we talked about this morning. It's the published -- well, not the published data, but it's the data from the 90 microgram group in the original study compared directly to the assays done on the open label study.

DR. COX: Right. So theoretically, if you have enough serum left, you could go back, test those two sets of sera at the same time with clade 1, clade 2 and clade 3 viruses and rally get quite a bit of interesting information about --

DR. TREANOR: That absolutely could be done.

DR. KARRON: I think if there are no more questions at this point, we'll proceed to the open public hearing.

MS. WALSH: As part of the FDA Advisory Committee Meeting Procedure, we are required to hold an open public hearing for those members of the public who are not on the agenda and would like to make a statement concerning matters pending before the committee. Dr. Karron, would you please read the open public hearing statement?

DR. KARRON: Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making. To ensure such transparency at the open public hearing session of the Advisory Committee meeting, FDA believes that it is important to understand the context of an individual's presentation.

For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the committee of any financial relationship that you may have with any company or any group that is likely to be impacted by the topic of this meeting, for example, the financial information may include the company's or group's payment of your travel, lodging or other expenses in connection with your attendance at the meeting. Likewise, FDA encourages you, at the beginning of your statement, to advise the committee if you do not have any such financial relationships.

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

MS. WALSH: I have received a request form Dr. Bruce Innes representing GlaxoSmithKline. Dr. Innes? You can come up to the podium for your slides.

DR. INNIS: Good afternoon. My name is Bruce Innis. I'm an employee of GlaxoSmithKline. We're a manufacturer of licensed vaccines, licensed influenza vaccines in the United States and in many countries around the world. I'd like to make a public statement regarding our ongoing development of a novel influenza vaccine against pandemic strains.

This morning we've heard an awful lot about the challenges of using vaccines to reduce the risk of pandemic influenza. GSK has a strategy to confront these challenges, and that's what I want to describe. So we'll skip that slide and go right to GSK's position, which is that advanced production and stockpiling are the foundation, the foundation of pandemic preparedness. And the ideal vaccine to support this approach has three attributes. It should be effective against drift variants and elicit immunological memory against them. It should be antigen-sparing, and if it's going to be stockpiled, it needs to have shelf life.

We are developing pandemic vaccines with just these attributes, and I'd like to share a little bit of data with you. Now here you see immune responses from a dose-ranging study in which healthy adults 18 to 64 years of age were randomly allocated to 8 formulations of Vietnam/1194/H5N1. They received 2 doses of the vaccine 21 days apart with or without a novel adjuvant system. The doses ranged from 30 micrograms down to 3.8 micrograms. The endpoint that we're showing here is hemagglutination inhibiting antibody response, the proportion of volunteers that achieved a post vaccination titer of greater than or equal to 1 to 40. And I'm showing an orange reference line which indicates the 70 percent response threshold that CEBR has articulated in their draft guidance as evidence of an adequate response to reasonably infer clinical benefit.

So if you look on the blue, the dark blue bar, that's the 30 microgram dose without adjuvant. That group failed to meet the criterion. But if you look on the other end of the graph, the light yellow bar, that's the 3.8 microgram group with adjuvant, and they exceeded the CEBR criterion. So these results show that the adjuvant system confers a very marked antigen-sparing affect.

Next week our colleagues will be presenting new data in an international scientific meeting in Hong Kong regarding the ability of this vaccine to immunize against H5N1 drift variants. That's a second important attribute of an ideal pre-pandemic vaccine. Now you may ask is there a regulatory pathway for pre-pandemic vaccines. And the answer is yes, there is. The EMEA in Europe issued comprehensive guidance on the 24th of January, and I'd like to go over its salient features. Their guidance says that applicants for licensure should evaluate as the primary efficacy surrogate the homologous HI anybody response but that also they should characterize vaccine cross-protection. This is considered important, and three types of evidence are requested: cross-reactive neutralizing anybody responses; cross-protection of ferrets; and the ability of vaccination to support a booster response to a subsequently administered drifted strain.

GSK has been and is continuing to generate these data for its pre-pandemic vaccine candidate that's manufactured in its facility in Dresden, Germany. For that vaccine, we have already filed an European Union license application, but currently we are discussing with FDA the evidence that would be required to support a U.S. license application for our pre-pandemic vaccine in our facility in Quebec, Canada.

The plan that's under discussion with FDA is divided into three parts: the conduct of pivotal trials of our vaccine under an IND in 2007 and then immediate application for a BLA for the use of the vaccine in the face of an imminent threat; and while the application is under review, and our understanding is that it would be reviewed under the accelerated approval regulations, we would extend the development by evaluating the vaccine in children and generating data regarding the ability of the vaccine to prime subjects for a subsequent heterologous booster response; and lastly, we would propose post licensure to conduct large, large safety trials in adults.

Now ultimately GSK is interested in developing a product that's suitable for general use prophylaxis, not only stockpiling, against various forms of pandemic influenza. This type of vaccine conceivably could be used in a routine national immunization program, and this could be the best possible form of risk reduction against any future subsequent pandemic.

Let me close by saying that GSK has committed to developing a new generation of influenza vaccines against both seasonal flue and pandemic flu and the use of the adjuvant system that I've briefly mentioned to day is certainly central to that vision. Thank you very much.

MS. WALSH: Thank you, Dr. Innis. I have also received a request to speak for Ms. Manon Cox representing Protein Sciences. Ms. Cox?

MS. COX: Okay. I would like to take the opportunity to update the committee and the public today on the recombinant hemagglutinin vaccine that was so nicely introduced by John Treanor before, because you may wonder what happened in the last eight years. In principle, it became clear to Protein Sciences that it was important to first develop a vaccine for inter-pandemic use, so we embarked on the development of a trivalent recombinant hemagglutinin-only vaccine which is produced in vitro using insect cell culture technology. Every year we will clone the HA's from the WHO-CDC recommended strains and since this is a recombinant DNA approach in principle, you do not need eggs. It is easier to produce. And you also do not need to produce live viruses prior to inactivation. There's no bio containment required and we also do not plan to use preservatives.

The hemagglutinin antigens are then highly purified and they have the correct three dimensional structure as is demonstrated by their biological activity, hemagglutinin activity and by the fact that they produce protective immune responses.

I would also like to outline our approach to watch a potential pandemic. As John indicated, in 1998, we very rapidly, after NAID approached us to produce a pandemic vaccine produced vaccine that could be used in the clinical trials that were described earlier. It took us six weeks to get from gene to product and eight weeks to get from genes into humans. We also demonstrated that we could fully protect chickens against a lethal challenge. And in a very short period, this vaccine was given in 1998 already to 200 healthcare workers of which a little over 50 percent reached titers that were greater than 1 to 80, titers that were found in convalescent sera. This was the first pandemic vaccine in clinical trials, and it didn't really help the development of Protein Sciences' vaccine, because the fact that you needed two doses of a relatively high amount of hemagglutinin left people to believe that a recombinant vaccine would not be as efficacious.

But there are some very important considerations to keep in mind. In principle, if you want to produce a recombinant hemagglutinin, you do not need to grow or handle a live virus. You can use a well-defined cell line versus an undefined egg production. There's also an enormous search capacity, because these products can be produced in a mammalian or monoclonal antibody production facilities.

And if you think about it, the worldwide production capacity for mammalian cell culture is about 2.5 million liters. And if then imagine that you could produce 1 million doses of 135 micrograms of vaccine per 10,000 liter in a 5-day production cycle, you could imagine that it is feasible to produce billions of doses in matters of weeks.

It's also important to point out that the antigen that you make using a recombinant baculovirus approach is an exact match to the H5N1 that is naturally appearing, so you do not need to make a reverse genetics modified strain. I also want to point out, as was pointed out by Dr. Couch before, that a potential influenza pandemic doesn't focus around age 5 alone. If we look at the last about 10 years, there have been many different avian viruses circulating with various impacts on humans. And this slide, the only purpose of this slide is that there's not just H5 but there's also H9 and H7, and as has been pointed out, H2 might been present a greater threat.

So what we have done at Protein Sciences is we have produced four additional hemagglutinins. They have been cloned from strains with pandemic potential, and they were produced using the general hemagglutinin purification process that we also use in the development of our inter-pandemic vaccine. I believe strongly that a recombinant protein-based influenza vaccine is the most vital, proactive approach in fighting against a potential influenza pandemic. As was indicated earlier, it will take time for antibodies to develop, and if you can prime the immune system, that can have a major advantage.

We plan to produce and market a prophylactic pandemic vaccine after FluBlok has been approved. It is clear that you need a large safety database before you can develop such a vaccine, and therefore we have been and are conducting quite a number of clinical trials, and I want to very quickly highlight those trials.

We are conducting at this very moment a trial in young children age 6 months to 59 months. We expect results in the second quarter. We are also conducting an immunogenicity but also efficacy study in elderly or people 65 years and older. The immunogenicity results of this study are expected in the second quarter. And a couple of months later, we will also have efficacy results. We earlier found I a field study that the commercial dose that we selected, and it's important to point out that this commercial dose will contain three times the antigen content of the regular influenza vaccine, so it will contain 45 microgram of each of the hemagglutinins, was 100 percent effective against cell culture confirmed influenza in subjects that presented with influenza-like illness. CDC defined influenza-like illness, even against drifted strains. We characterized all the strains, all the viruses from this study, and they all represented drifted strains.

As far as effectiveness goes, there was a 54 percent reduction in subjects that presented with CDC-ILI versus placebo. So we demonstrated in this study that a hemagglutinin-only vaccine can be efficacious and effective without neuraminidase, and we also showed that the vaccine was highly immunogenic. More antigen leads to better immune responses as has been demonstrated or written in the literature for quite a while. We were able to show protective antibody levels for all antigens for at least 6 months.

We also earlier in collaboration with NAIAD conducted a study in the elderly where we particularly defined our endpoints against around the H3 antigen. These studies were published by Treanor et al in 2006. And as you can see here, there is a clear dose response effect. When given more hemagglutinin, you will get a greater sera conversion.

We also were asked by the Journal to perform a subset analysis because, as we know, as you grow older, your immune system may become more senescent, and therefore the right part of this graph is also quite interesting. And if you keep I mind that we selected the purple dose as our commercial dose, that gives you some reference.

Now the baculovirus technology provides speed, cost and safety. It also provides a rapid response to emerging strains. So in other words, if tomorrow a new virus is identified and we know what antigen could convert protection, we would be very rapidly able to develop a vaccine.

There is not need to handle a live virus. And if you keep in mind that the latest outbreaks around SARS all came from laboratory or places where they were working with viruses, you can imagine that this is a risk that you cannot underestimate. As I mentioned before, we use the same antigen that is actually circulating.

Our next steps are that we are going to further evaluate the two clinical studies that are currently ongoing. We will initiate an efficacy study in the 2007-2008 season in a very large group. We expect to initiate our BLA filing in the fourth quarter of this year. And subsequent to that, we will initiate the development of a prophylaxis vaccine.

And in case it wasn't clear, I am an employee of Protein Sciences, so that will be my conflict of interest. Thank you.

DR. KARRON: Is there anyone else who would like to speak during the open public hearing?

(No response.)

DR. KARRON: Seeing noone, we will take a break until 3:30 when we will reconvene for a discussion.

(Whereupon, off the record at 3:17 p.m. and back on the record at 3:41 pm.)

DR. KARRON: Okay. I think we're going to go ahead and begin if people in the back of the room would please take their seats, it would be very much appreciated. And I think the best way to begin this discussion is to really go through each of the slides, each of the items that Dr. Toerner put up for our consideration. So we'll begin with this first one which has to do with the issue of assessment of immune responses, both the kinds of assays used and assessing responses following prime and following boost. So at this point, I'd like to open those issues for discussion. Comments?

I think maybe what we can do is start with one of the last items first and then perhaps move up, which has to do with the use of HI antibody assays versus microneutralization assays as a measurement of immune response. Would anyone like to comment on that?

DR. COX: Thanks. I think that there's a growing body of evidence that indicates that HI assays using horse red blood cells are really good assays for detecting antibody to H5. But in our experience at CDC, and I can just speak to that and to some experience elsewhere that I've heard about, the microneutralization assay, although it's a lot more labor-intensive is still the gold standard, and there are some unusual effects that you can sometimes see with horse red blood cells depending on the animal and the test and the antigen and so on. So I think that it's really fantastic that we have now the ability to screen using the horse red blood assay, but I think the microneutralization test actually is better in reliably detecting antibody to H5.

DR. KARRON: So would you perhaps advocate using both tests as measurements of immunogenicity?

DR. COX: I think that at this point in time, it would be a good idea to use both assays, and as more and more data are developed, perhaps it would be possible to move to the horse red blood cell assay because it is so much easier. But right now I think we're still at a stage where we need to do more assays and more comparisons and really get the cutoffs right.

DR. KARRON: Dr. Toerner, I was actually wondering if you could elaborate a little bit on your first two points there, immune response assays following prime and boost and what you wanted to elicit from the committee, comments you wanted to elicit on those points?

DR. TOERNER: The point that I wanted to make is regarding the first bullet point, the immune response assay following a prime. What I would be interested to hear is the heterologous immune response following a prime, would that be supportive evidence of demonstration of appropriate priming across protection in contrast to following subjects over time and administering a heterologous antigen to the subjects and then measuring the immune response to the subjects following a boost administered a future time point. Does that --

DR. KARRON: So if I'm understanding this correctly, I wonder if maybe Dr. Treanor has data that could bear on this point at all, and that is to say from your original studies where you immunized with Hong Kong/97, did you then go back and look at those -- before those individuals were boosted, did you ever test their sera and look antibody responses to 2005? Is that the kind of thing that you're asking?

DR. TOERNER: Yes, that's correct.

DR. TREANOR: That would be a really good idea but we didn't do that. It's another thing on the list of things that would be good to do. The sera are available but they haven't been tested.

DR. KARRON: Dr. Self?

DR. SELF: I don't know too much about this system, but I can compare the nature of this discussion to what we've had in HIV vaccines. And there are very well-standardized, broad panels of reagents, of pseudo variants in this case, and a system of labs and assays that have been highly standardized, validated, proficiency panel tested so that endpoints taken at standard times from the last boost can be compared across, you know, many different studies with, you know, some reliability. I'm not hearing anything of such a system here. Maybe there is something like that but if there is, exercising that kind of a system and having those sorts of reagents and standardized assays sounds like it would be a very good thing.

DR. COX: I think that you're absolutely right and there's a lot of thought going into that type of system so that you have standard sera. And I think that the NIH is involved in some trials that will produce some standard sera. And also having a standardized panels of antigens that could be used to test so that you can actually compare from study-to-study. And I would certainly advocate for a lot more harmonization among the studies that are being conducted so that we can -- and a lot more head-to-head comparisons so that we can really understand what is going on in terms of cross-protection, how much greater cross-protection you get with adjuvants using a specific antigen, and a whole variety of other things. So these panels are -- discussions are occurring about how to get these panels put together correctly.

DR. SELF: So in HIV, these panels are also tiered that begin with the homologous virus and then sort of expand, not going to the next tier unless you see a good breadth in magnitude in the current tier. And the assays also span both antibody as well as cellular immune response. I heard earlier that there's some interest in the role of cellular response here, and so maybe there's also something --

DR. COX: Yes. There's not a lot done but John may want to speak to more studies that will be done to look at the --

DR. TREANOR: Well, in the study that I presented, a large proportion of the subjects had peripheral blood mononuclear cells obtained. The laboratory that's going to be assaying those has really spent quite a bit of time validating their cellular assays and showing reproducibility and reproducibility of thawing cells and all that kind of stuff. And I think they're just beginning now to start actually doing the assays on the PBMC that will provide another way of looking at immune responses beyond antibody.

DR. SELF: So the reagents there, the antigens there to reflect, you know, variability in the targeted virus population raises a whole another series of problems, so that's another issue to -- that you'll have to address at some point.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: John, correct me if I'm wrong, but I thought one of the things that I heard you had on your list of things to do, which must be very long by this time, was to take the sera from the Sanofi vaccine that we just looked at this morning that you'd carried out and to test those sera against clade 2 and perhaps clade 3 viruses, is that correct, as a measure of heterologous response?

DR. TREANOR: Yes. I think that, you know, it would be fascinating to know what kind of response revaccination generates against the original antigenic exposure and it would be a very great practical interest to know whether these individuals are also responding to clade 2. And so, you know, this is a assay development sort of issue, but those will be done at some point by the reference lab.

DR. EICKHOFF: Well, for that reason, I think heterologous or testing against heterologous antigen would be very useful as outlined in that slide, simply because it may provide some -- it may correlate with a level of boost that you get with a heterologous virus or will it correlate with the level of boost following boosting with a heterologous virus.

DR. KARRON: Dr. Couch?

DR. COUCH: Well, just a couple of comments for the discussion. One was that John made essentially and that is what we're talking about when we're looking at antibody and boosting antibody, that's an operational definition of prime and boost. If you really want to know whether a person's primed or not, you're looking at whether those lymphocytes are recognizing that antigen or not, and we've got th technology to do that. See, I was unaware until he told me a little earlier that Jim Crowe had been trying to do that specifically for H5. If that's really a goal of pre-pandemic vaccinations, then the priming assay should be out there being looked at right now and know the differences in dose and age and maybe underlying disease in terms of what the variables are that determine priming, because that will define your response at a future time. And that's a far better and more accurate way to define priming than the way we're talking about it with operational definitions prime and boost.

Second is that there are a lot of different ways to do neutralization tests. When we use the term microneutralization, we're usually talking about the test that was described originally by Maurie Harmon that Jackie Katz has picked up using at CDC, and I think that the British version is essentially the same thing. And even the way we do it, which is somewhat different -- essentially all of them -- they don't have to be that way, but essentially all of them are another way of measuring antihemagglutinin and antibody, so I think it's important that you keep that in mind. Because if they don't correlate, well, then you've got to raise a question as to exactly what your neutralization assay is measuring and antihemagglutinin and antibody is the antibody we've all been focused on as the desirable immune response, not the only one that might be useful but the desirable one and the one that we're still using as a standard for making decisions on influenza and influenza vaccines.

So the major value of neutralization, we've been doing a version of it for four years. In our hands, it has greater sensitivity than the HI test. I didn't make that comment earlier this morning, but it's perhaps useful for the way some of the thinking that's got on here that the HI test is really a fairly crude and relatively insensitive test for antibody. So you need to think about that when we talk about how we're going to try to use it.

DR. KARRON: I think that probably we should -- I think we may be able to move on to the next slide. I think if I can summarize what I think I'm hearing for the consensus -- I think the consensus is you probably want to measure heterologous protection both at the time of prime and at the time of boost using modern technology as well as using -- as just using at conventional antibody responses. Okay. Yes? Sorry.

DR. GELLIN: A question really for Bob. You got into it a little bit, but we're talking mostly about hemagglutinin, and I guess the question is given that the neuraminidase may be less variable, how do we use this as an opportunity to get a better understanding of what neuraminidase immunity buys us?

DR. COUCH: I'm sorry Rob Webster's gone, but you might say that we fall into two camps, the neuraminidase proponents and the neuraminidase is not so significant. I'm on the first camp that neuraminidase antibody is a highly desirable antibody. And what we know about the neuraminidase for at least the H1 and H3 -- that's Nancy's territory -- is it's less variable than the hemagglutinin. And of course, Ed Kilbourne's not here now but that would be his basis for proposing very strongly that the neuraminidase does need to be evaluated and is -- I mean this is H5N1, see, we're talking about and in H1N1 is an N1 neuraminidase, you see.

Maybe I better make a question out of it to Nancy then. Do we have the data to say that there is no cross relationship between those and that the N1 that we're currently vaccinating with on an annual basis would have no benefit for H5N1 as we know that neuraminidase. I don't know any of these questions and/or answers.

DR. COX: Unfortunately, we don't have the answer to these questions. I think there was a recent publication out of Rob's group which indicated in an animal model there was some cross-protection and there -- I think it remains to be seen. Of course, we know that because of the ages of a number of the individuals who have died of H5N1, they surely were exposed to H1N1 viruses in their lifetimes and that certainly didn't protect them. So I think the jury is still out and we have a lot more to learn about the role of neuraminidase.

DR. KARRON: John?

DR. COUCH: I might say that's a separate comment from the fact that neuraminidase is a useful antigen and immune responses to neuraminidase does indeed convey protection.

DR. TREANOR: I was just going to add that in that study that Rich Webby did, we did send sera, and there is a low level of recognition of the avian N1 in a panel of human sera from people who had received conventional vaccine, so the levels of neuraminidase-inhibiting activity are substantially lower against the avian N1 than they are against the human N1, but there is recognition of the avian N1 by human sera.

DR. COUCH: I suppose you might say then they are primed.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I'd like to ask a question of the flu people. Also like Steven, I'm from a different background. But does the prime boost suggest that there are T-cell epitopes that are linked to the B-cell epitopes, and if so, have those been mapped at all? And if not, that would seem to be something that should be done.

DR. KARRON: There has been substantial sequence analysis of those viruses, and I know that a number of epitopes have been potentially identified. The Hong Kong and Vietnam viruses are actually about, I think it's, 90 to 95 percent similar on an amino acid level. The differences are all in the antibody epitopes and more or less. So there probably would be potential cross-recognition, I would think anyway.

DR. KARRON: Okay. I think we'll move on to the second discussion point which was to discuss the feasibility of long-term clinical studies of prime and boost, whether studies should be six months, a year, greater than a year and also this issue of collaboration among different sponsors. And I think the intent there was if one sponsor might have had a clade 1 virus and another had a clade 2 virus, how might that occur. Yes, Dr. Robinson?

DR. ROBINSON: I just want to acquaint you with some contracts that we let out in January for antigen-sparing of pandemic influenza vaccines. One of the characteristics of those three contracts is that the contractors who are known publicly as GSK, Novartis and Iomai will submit to HHS their adjuvants for evaluation with the same antigen or other antigens for human influenza to inform public health decision makers as to if there can be a mix and matching that, during an imminent pandemic, if those adjuvants would work with antigens that we have in our stockpiles, because most of the stockpile is in a bulk form. So we are encouraging that and we're in the planning stages right now with the manufacturers of the adjuvants and also the antigens.

And so what we're looking for now is guidance from CEBR to help us come up with the proper study designs that would be acceptable and also the regulatory pathways as we move forward to that so that we can present a suitable case to VRBPAC in the future years.

DR. KARRON: Other comments? Yes?

DR. FARLEY: This is a little bit off the main focus of this discussion right now, but one of the things that I noticed about the modeling is that -- and what I think I'm aware of with Longini's as well -- is that the focus on prioritization to children seems to be sort of important and kind of drives, in some ways, these models of getting it out there early and having it work well, and I wondered if there needs to be any modification of the route with which this phase of development -- I mean should we be working more on or increasing the emphasis on pediatric trials and understanding their role in the prime and boost and whether they're going to tolerate the adjuvants and those sorts of things?

DR. KARRON: Jesse?

DR. GOODMAN: Well, I think that's a good question. You know, we have encouraged pediatric trials of pandemic vaccines, but we've done that cognizant of sort of the special status of children in how we ought to have some safety data to support that before such studies are done. But yes, you notice even with the nonadjuvanted Sanofi vaccine, there is a small pediatric study that's been done. And then I think this is an important point -- and then I know that others who are developing new vaccine, once they have substantive evidence of safety and immunogenicity in adults are planning pediatric studies, and we're encouraging them.

Now I would say the -- you know, maybe Dr. Couch or others might want to comment -- but I think the notion of children as sort of hyper spreaders and important -- both important to protect in a pandemic and also potentially important in transmission, I think that the first one everybody would agree. The latter one, I know it's not as well documented as it could be, but I think that we should get that data. You know? But I don't think I'd want to go prematurely with novel compounds into children.

DR. KARRON: Dr. McInnes?

DR. McINNES: I want to follow-up on Robin's introduction there. We have not taken a position that an adjuvant can be thought of as a stand-alone project. I mean there's no adjuvant licensed. It's a product that has antigen in combination with an adjuvant that comes forward for licensure. So I'm trying to -- maybe you could explain a little bit what the plans are for the product characterization of these, what essentially are, off the shelf mix and matches and the characterization of the product, the pre-clinical safety evaluation of that product given that you might be looking at varying concentrations of either component and then what you're thinking about in terms of the Phase I? I mean you would have a characterization piece, a pre-clinical safety piece, an immunogenicity piece and moving to human studies.

So if you look at that sort of, you know, could really be a 20 by 20 box or a 10 by 10 box or a 50 by 50 box. I mean I have no idea what the plethora of combinations could be. So could you tell us a little bit about that strategy and who's going to take this one and how is that going to work when you don't have manufacturers necessarily envisioning a commercial project here?

DR. ROBINSON: Well, I'll leave it to the manufacturers whether or not they consider their products with adjuvants to be commercially viable or not on that count. But essentially, we do recognize exactly what you just said is that adjuvants are not stand-alone products in our world. What we would like to know is if it is possible to actually develop formulations of vaccines that actually can be filled into antigen concentrations that can be tested in pre-clinical animal models, preferably in a fair challenged model with the adjuvant either pre-formulated with it or admixed prior to giving the vaccine to the animals and then challenging them.

From that data, two things will pop out. One is that we may see that some formulations, some antigens, the way that they are actually formulated as a bulk product and then finally into a final container product, are contraindicated for other adjuvants. That's a possibility we would find that out so that we can strike that one out.

If they are compatible and they do afford in animals a reasonable, and in these cases both homologous and heterologous cross-protection, then the data supported in the toxicity studies also, then we would envision a subset of those going into Phase I clinical studies for safety, immunogenicity, and cross-protection as far as cross-reactivity for serological samples. So that in a nutshell would be what we think.

But you're exactly right. I mean one of the things that we do want to understand is is there compatibility of these different products, and we would understand that in the pre-clinical setting to move forward.

The desire for this, the reason for this is that if we have stockpiles of bulk antigen that are there and then we can increase the amount of doses that we can actually put into people's arms, then it behooves us to at least look at, in pre-clinical studies, you know, are these compatible and then, as I said earlier, what can the CEBR give us as a pathway to move these forward and presumably to use these under emergency uses authorization. I don't foresee these mixing and matching of antigen and adjuvants as a licensed product except for the homologous systems where the company has developed and has moved forward with the licensure of a particular antigen with that adjuvant.

DR. KARRON: Dr. Toerner?

DR. TOERNER: Just to provide some additional clarification with the bullet point of the collaboration among different sponsors, I think what our goal was to emphasize the value of following subjects wh o were enrolled in studies. If you go back a few slides or look at your handout, the cohort A and cohort B were those study subjects who've received one dose or two doses of vaccine. We think there's value in following those subjects over time in order to administer a different vaccine or a different clade, so that's the point that we wanted to make about the collaboration among different sponsors.

DR. KARRON: Essentially, to be able to replicate studies like Dr. Treanor's over time?

DR. TOERNER: Yes. For example, if a sponsor is pursuing development of a vaccine for use during a pandemic and they have those data, they have those immune response data from study subjects in cohort 1 and cohort 2, that there perhaps could be a mechanism to follow the study subjects out in order then to demonstrate the possibility of priming.

DR. KARRON: Dr. Goodman?

DR. GOODMAN: One thing we'd appreciate input on -- it may just be that this is very straightforward, but if we can go back to Dr. Toerner's slide about heterologous prime and boost, you know, are these basic outlines of, you know, the 1 or 2 doses then followed by a boosting dose at 6 or 12 months, you know, do people looking at those think that they're reasonable approaches or have other suggestions? We all agree we should get as many useful assays and samples at the different time points to understand whether we can predict a good boosting effect, etcetera, what heterologous immunity might have existed before the boost, etcetera.

But, you know, the real question is does this kind of approach where you'd get, for example, one, or I suppose this doesn't have two followed by boost, but one or two doses? And, you know, there are many, many variables and approaches to doing this, but, you know, for example, a six-month boost might be comparable to being immunized in a sort of pre-pandemic-emerging pandemic followed by a new vaccine. A year or greater might be similar to just a population being pre-immunized. Are these sort of reasonable approaches.

And the other issue I heard raised to both Dr. Treanor, and I think is a very -- again, you can't do everything at once. You have to start with the simple stuff and get your principles, but this issue of might these approaches -- there might be different approaches to dosing. You know? And then I know you're not going to be able to anser it, because we couldn't even answer it this morning for a much simpler question. But what should be thinking about in terms of what is a meaningful heterologous response?

So I want to frame first of all is this a reasonable structure, because manufacturers are, as you heard, may be starting to do or even doing some of these studies, and we can encourage, as I said, that is if these are just additional arms to ongoing studies, we could get data a year or two before you might do it if you did these studies just sequentially. So we do want input is this the right track.

And then as Joe said, it would be wise to look at not just with one manufacturer's vaccine against another's but in the real world, in an emergency, the boosting could occur with -- you know, you're not going to be able to say did you get manufacturer x and now we're going to just give x. So I hope that's helpful in terms of framing some of our questions.

DR. KARRON: Dr. McInnes?

DR. McINNES: In some senses, this takes me back, to the haemophilus influenzae days of thinking about -- and it's back to this operational issue of priming and boosting and, in fact, whether you are really testing boosting with a full concentration or whether you should be really looking at -- do you need one dose to prime or do you need more than one dose to prime I think is one question. And then the boosting piece, in fact, I'm not sure of that with 45 micrograms you're actually testing boosting. You may actually need to go with a much lower dose concentration to evaluate boosting.

But I think you have -- we have to be very careful what the question really is on the table. And there are so many questions that could be asked here. So you could -- I mean theoretically, you could take group A and wonder why -- you could look at waning immunity with time and look at what they look like boosting them at -- you know, giving a second dose at 15 months. It just sort of never ends. So I mean I think what you've put up on the table is as reasonable as -- for a starting point. And I think maybe it's better to actually raise the principles around which you want the studies to go rather than trying to define specifically timing.

DR. KARRON: Dr. Couch?

DR. COUCH: Yes. You made me -- probably like a lot of them, I didn't think about it when you were saying it, so you made us start thinking about it. And if you really focus on, as you said, pre-pandemic use, I guess there are a whole lot of variables that you'd like included in these studies and then they might well become prohibitive. This is not bad. If I added one besides that, I would add Vietnam at each of those boosting sites as comparison to that heterotypic boost, the homotypic boost. That would be the scientific data I'd like to see.

DR. KARRON: I'd actually just like to -- not that I can begin to answer this question -- but the issue of what kind of a heterologous would be considered adequate? I have no idea but I would say that just sort of to echo something that Dr. Couch said, I think one of the other reasons to use more modern methods to look at priming besides the fact that it actually looks at what is we think of when we biologically of what priming is is that it may, in fact, be much more sensitive, that you might be able to detect responses that you cannot detect using conventional antibody assays. And so I think that should be kept in mind.

DR. COUCH: Part of what that homotypic would give you that I wanted to put in there was the question of whether you would need then still another boost of Indonesia because you'd have the Vietnam.

DR. McINNES: And then depending on when you're going to do your serology, I"m not sure the timing on whether you will really be able to characterize the kinetics of the response. Maybe it is a very early response, and I think we have to bear in mind that you may have to actually be pulling blood much more frequently maybe as an initial study to characterize the kinetics.

DR. KARRON: Dr. Cox?

DR. COX: Yes. Given a lot of the discussions recently about the need for developing countries to also have access to H5 vaccines, I think the looking at what the quantity of antigen that it really takes to prime and to boost is incredibly important, and so you've got to have some dosing components in there, because we really do want to be able to conserve antigen as much as possible.

DR. KARRON: Dr. Treanor?

DR. TREANOR: Well, I was just going to say two things. Just bear in mind that all the experience that we have with H5 that relates to this sort of prime boost idea involves a two-dose schedule for the priming. So we don't have any information about a single dose of one thing followed by a dose of another. And the pediatric data suggests, as I understand it, that that's not quite as good when there's a strain change. So that's the thing.

And the other thing I would say is that I think -- bear in mind that the assays for neutralizing and hemagglutination-inhibiting antibody which we're using now have been the subject of quite a bit of effort to standardize. And even so, it's well-recognized while within a lab, they can be very reproducible, the absolute titers that two different laboratories on the same sera are not always in complete agreement. I would think that you would be faced with similar or perhaps more challenging issues trying to really standardize and validate the types of flow cytometry and gamma interferon ELISPOT assays you might be using to look at some of these other cellular responses. And I think that will be a major challenge for looking at some of these other questions related to priming and boosting.

DR. KARRON: Dr. Goodman?

DR. GOODMAN: Something somewhat encouraging -- it doesn't shed light on this but is that certainly both in the literature and that for those of us who were at the WHO meeting, we did hear several vaccine developers of some of the newer vaccines show, you know, heterologous immunity without boost, so there may be some background there to work with. And also, on Nancy's point about antigen content, of course, you just heard a presentation suggesting that potentially with certain adjuvants, that may greatly reduce the antigen content.

But, you know, we're going to be stuck with the challenge of trying to answer some of these questions without the studies becoming, you know, impossible or overly complex. And, you know, I see this sort of 12-month framework that's up there, and maybe with the additional of the homologous control, you know, as a good one, I think if we were to having to use high amounts of antigen, then the idea may be at some of these boost points, you know, titering what you boost.

But we're not going to be able to answer all these questions at once. And as I said, I also think there are companies, for example, doing studies or thinking about studies that may inform -- you know, again, just like we said this morning, in three months we could have additional data from one place or another, but we do appreciate these suggestions.

I didn't hear a comment on, you know, what -- you know, should we just be happy if there is heterologous immunity? Do people have comments on tying that as was suggested in the discussion of how the Europeans are looking at it to efficacy in the ferret? What do people think of those issues? Because we will at some point bring back to you probably one of these, hopefully, wildly successful vaccines and say does this evidence for priming or heterologous protection, you know, merit either a claim or an indication.

DR. KARRON: Nancy?

DR. COX: I did notice the sort of question implicit in the presentation about the use of animal models, and I do think that it is informative to use the ferret model. It's perhaps the best model that we have right now, best defined for H5. There are some clear endpoints that can be obtained using challenges and so I think that it would be very useful to include the ferret model for looking at cross-protective antibody.

DR. KARRON: Okay. I think we should probably go on to the last question -- discussion point I should say. And this has to do with -- actually, one more -- Christine, sorry. This has to do with issues related to safety considerations, pre-licensure safety database, and the issues related to novel manufacturing processes or adjuvants.

One comment that I would make that, really, Dr. Robinson's comments made me think about with the sort of mix and match issues is I think that raises real -- there are efficacy considerations. There are also safety considerations, because every adjuvant behaves differently with regard to inducing immune responses when combined with different antigens. It also may have different safety profiles when combined with different antigens. And so I think that's something that has to be considered. Dr. McInnes?

DR. McINNES: I'll put something out. So in a pre-pandemic setting, I see a relatively high bar to demonstrate safety, a requirement for safety, and I would see that in needing to be in many thousands of people. And I don't know how many thousands. One would have to sort of give that some consideration. But I certainly don't see it on the same scale of what we talked about this morning. So I think this is really no different than the way we would think about licensure of other vaccines, that when you're going to be introducing into a broad population, we have a responsibility to be documenting and characterizing the safety profile. And maybe this gets done in a staged way, in a stacked fashion, but I think that we are looking at large studies.

DR. KARRON: Comments, reactions? I'll say that I actually agree with Dr. McInnes in terms of pre-pandemic and pandemic use and that, you know, my sense is the bar would be very different in terms of safety profile in those two settings. Dr. Wharton?

DR. McINNES: Yes. I would agree with your comments and in the sub-bullet, there were serious adverse events, those at a frequency of 1 per 100,000 not likely to be detected in a typical pre-licensure database, I think we're unlikely to be able to detect those in atypical pre-licensure database either. It's hard to imagine a study large enough to do that, so I think that one is left with having to come up with plans where those can be identified in the post licensure setting.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: One issue thought of as you were talking actually, Ruth, and as with an adjuvanted vaccine down the pike when it's coming up for review, since there can be differences in reactivity and immunologic response to different antigens, it will be important, I think, as different clades come out, even though those are fairly subtle changes structurally, that even clade to clade evaluation will be different, and it's not going to probably be like our current system I would think but I'm not sure as far as being able to go one year to the next. And I guess I'd be interested in what the influenza experts on the other side of the table and the top table if they agree that's going to be an issue or not.

So the question is do you think with an adjuvanted vaccine, because you can have quite different responses based on formulations with different antigens, that the year to year variability will need to be assessed more diligently than the current system with the non-adjuvant vaccine?

DR. COUCH: Well, I think you can only comment on that in a general way, and I don't want you to -- you're looking at me, but you're not looking at an expert on this subject. Let me say that.

But adjuvants, if they do what we want them to do, increase the immune response, and the greater the immune response the greater the cross-reactivity if there's something like an antigenic drift that we're talking about -- so I think the general concept we would all accept -- now when you start talking about an individual antigens -- adjuvants, there are TH1 adjuvants and there are TH2 adjuvants and their immune response is different, I'm sure the cross-reactivity would differ. And that's data that we don't yet have for a lot of antigens including just a little bit of data, I'd say, on influenza.

But in general, I would say that if an adjuvant does what we want it to do and expect it to do, it should increase the cross-reactivity if that's what you were driving at, Jack.

DR. STAPLETON: Not exactly.

DR. COUCH: Not exactly? Try me again.

DR. STAPLETON: No. I just was curious from a regulatory standpoint, safety standpoint if that's going to require a different level of diligence and study annually on a year-to-year basis. And I throw this out as a rhetorical question.

DR. COUCH: Oh, safety questions are a whole new subject. When you bring up safety questions for adjuvant and most people know the experience with incomplete Freund's adjuvant in the 1950's. You know, there were probably 10 to 100,000 vaccinated, and the responses, they were reported in the military primarily, in the use of this vaccine were excellent. A lot of it was used civilian wise as well, and the adjuvant was shown very clearly to be dose sparing. The general figure I carry in my mind from that data is you could get the same antibody response with about 25 percent as much antigen if you used incomplete Freund's adjuvant along with your vaccine. And that was considered highly desirable in 1957.

But if it hadn't been shown inappropriately to produce polyoma tumors in mice, and which it was an artifact in retrospect and hadn't been shown to produce sterile abscesses, although they were very rare, it might still be around today as an adjuvant that we knew a great deal about and could be considered quite useful. But that's not where we are. So I think that really you're asking for an FDA comment, but I'm not sure to a considerable extent adjuvants need to be evaluated, that we're almost starting over again with relation to flu vaccine despite the fact that there was a good bit of experience in the 50's.

DR. KARRON: So I think what Jack is asking, probably of the FDA, is if you were to license an adjuvanted vaccine for a particular pandemic strain, then would the laws of sort of strain change apply? Or given that you have an adjuvant, would you need to reassess in the context of that, say, a new clade? Do I have that correct?

DR. STAPLETON: You said that better than I did. Thank you.

DR. BAYLOR: I think -- and let me make sure I have the question right -- you're saying if you have an adjuvanted antigen and then we change the strain the next year, then would we require additional data for that? And I think, you know, the manufacturing process is going to drive this. The first product, we're going to look extensively at that adjuvant and the safety of that adjuvant with the antigen. We would have to really think about whether adding a different antigen would really change sort of the profile of that. And, you know, in the absence of data, I can't say that.

But sort of on the normal under just general principles, you would not think that that would be the case, that changing that antigen would elicit some kind of safety issue that you didn't pick up in the previous. But -- well, we don't know that. And so we would have to think about that really seriously, to think about whether that new antigen would add some higher concern. But again, it's sort of in the once you've approved the adjuvanted whatever that is that year, it almost falls into the strain change paradigm. But again, it's -- we're dealing with the unknown.

DR. KARRON: Dr. Hetherington, did you have a comment before?

DR. HETHERINGTON: No. I think it's been answered.

DR. KARRON: Dr. Treanor?

DR. TREANOR: This is a more theoretical -- just a hypothetical issue that's brought up by something that Robin mentioned and this may be naive, but as I understand it, there's an issue with the possibility that we would have difficulty supplying the number of doses we need for a pandemic vaccine even if all the manufacturers who are capable of making the vaccine were operating at full strength. And in this process of evaluation of adjuvants, we might go through, you know, five, ten different adjuvants and find out that a particular adjuvant, adjuvant x, is absolutely ideal for a pandemic vaccine.

So the question would be under what circumstances can now manufacturers a to z use adjuvant x for their pandemic vaccine so that we would have an adequate supply? And this is where the mixing-matching idea comes from. And I don't know what the pathway for it is for should a very, very important adjuvant be discovered but only be made by one company, how would this be able to be used by other manufacturers to improve the supply?

DR. KARRON: Jesse?

DR. GOODMAN: Yes. I think that is what Robin was trying to address. I think there's a whole number of issues which are both scientific and then intellectual property-business relationships, etcetera. But I think the scientific one, and this is I think what triggered Jack's question and what Robin was commenting on, is we have a lot -- there probably are more concerns about an antigen made with one manufacturing process and an antigen made with another manufacturing process and whether when those are mixed with ideal adjuvant x in potentially different circumstances or time points, that could raise a bunch of issues about formulation, stability, immunogenicity, safety.

So I think there would be scientific issues that would need to be addressed, you know, probably through clinical studies. And how you would address those would be different perhaps in an emergency versus for a routine product. I think what Robin was saying is that at least in HHS's contracting, they've tried to preserve some ability to at least do some of the studies that would answer those questions.

But, you know, I think what we would hope is that as data emerges, if there are ideal candidates, they'd become -- you know, or if some -- I mean there may be more than one candidate that works very well and more than one approach, and there's always value to having that. But if there aren't, then I think this would sort of be both the national public health issue as to how you brought the best technology to bear and, to some degree, a business issue.

But I think the big question here and we heard several knowledgeable people comment at WHO that it -- you know, there are things about the chemistry of the manufacturing process and we're aware of these, too, but may not always line themselves to this being a simple matter of, you know, taking x and throwing y in it and add simple things like pH, polarity, water content, etcetera. All these kinds of things may affect behavior with an adjuvant. So -- but this is another area where we need more science.

DR. KARRON: Dr. McInnes?

DR. McINNES: I'd like to ask the card-carrying influenza accolades for some input on -- I was recalling the data from VTEU studies where there was a very clear dose response curve, immunogenicity wise dose response cure, and then thinking -- and I don't recall the slide as clearly -- but I was struck with the GSK immunogenicity of really a very pretty much flat immunogenicity response and not much of a dose response curve based on this is the adjuvanted product, that they really pretty much looked the same across the spectrum of those concentrations of antigen. And is that typical for -- who has recollections of adjuvanted flu vaccines and whether you have a dose response curve or whether you get almost an all or none, there's some critical level that it -- you get as an equivalent immune response?

DR. TREANOR: Pam, that's more or less identical to the pattern that's seen with MF-59 as well.

DR. COUCH: The same with MF-59 with H9, too, with the adjuvant, but the dose responses there are for non-adjuvant.

DR. STAPLETON: Yes. Have you done studies where you've actually reduced the antigen down to levels where you can say that you don't have a dose response? It's maybe that you just get such a good response in that slide, 3.8 micrograms, maybe they just need to go down to .038 micrograms?

DR. TREANOR: I haven't actually done these studies directly, so I don't know what the whole dose range is that's been studied.

DR. COUCH: I think -- well, you want to -- we'd like to see the lower anchor for dose responses. That's part of what you're saying. We didn't see that one there. But one of the more important aspects -- or maybe one of the important aspects of those, what look like, comparable responses to lower doses is the duration of that response and the pattern of that response which out to be dissected. And I'm probably sure GSK is doing that.

DR. KARRON: Dr. Innis, would you like to make a comment?

DR. INNIS: I would about the dose response. What we've seen is that there is a very slight dose response. I didn't show you the GMT's and so that distorts things a little bit, but there is a slight dose response. But I expect that we're way up on the shoulder of need or maximum response. And the operation -- of course, we would like to test much lower hemagglutinin concentrations. The issue is that the SRID assay is qualified down to formulate as low as about 2.5 micrograms. And so if we wanted to do less, we could, but we need to come up with ways to do dilutions that everyone would have confidence in.

And an even larger question is if you found that these lower doses were, in fact, effective, potentially effective, let's say immunogenic, how would you actually then be able to formulate commercial product and release it into the marketplace. So we're hamstrung right now by the limited quantitation of the SRID assay as it's specially formulated.

And this is a very, very serious issue. So if you have thoughts about that, we sure would like to hear about them.

DR. COUCH: My comment on the dose response is more of a scientific one than it is a manufacturing one, Bruce, but just scientifically, you'd like to know and understand what's going on with those responses in that adjuvant group, that's one. Manufacturing is quite different. Nobody questions your concern there.

The single radial immunodiffusion has problems, and it is an old but has been a reliable assay. But I'm not alone and you're not alone in saying that better and newer assays need to be developed for this purpose, for standardizing vaccines, and pandemic flu may be part of the stimulus to be doing that. And I can't comment on -- for some of the people who are beginning to look at some of these things, too. So that concern about the radial diffusion assay, hopefully, is only a temporary one.

DR. KARRON: Yes. Would you like to make a comment?

DR. WILLIAMS: Yes. My name is Mike Williams. I worked in the flu lab at the FDA from 1976 to 1996, now a consultant to the pharmaceutical industry. I'd like to add in 1978, when SRID was instituted as the potency assay for flu vaccines after the extensive clinical studies, we could not release final vaccine at the potency level of 7.5 micrograms, and potency of vaccine was released on monovalent concentrates as it was really up until really recent years. So there is a mechanism to do this. I think the FDA needs to get creative in working with the manufacturers. I would say if you can make a pandemic vaccine down in the .1 or less microgram range, then you really ought to be doing it, and there are mechanisms to do that and there is historical precedence to do it.

DR. KARRON: Thank you.

DR. WILLIAMS: Any questions?

DR. KARRON: Yes. Dr. Goodman?

DR. GOODMAN: I would just say we're certainly open to creative solutions to solve the problem. I also do think, you know, with modern chemistry and analytic tools, we can probably maybe do both, find ways to deal with the present situation and also probably find analytic methods that might improve on what exists, might even be better.

DR. KARRON: Any other comments that any committee members would like to make or members of the audience?

(No response.)

DR. KARRON: Okay. In that case, I think our day on pandemic influenza is concluded. Tomorrow, we will begin again with discussions of seasonal influenza vaccine. Thank you all.

(Whereupon, at 4:40 p.m., the foregoing matter was concluded.)