FOOD AND DRUG ADMINISTRATION

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

(8:12 a.m.)

DR. KARRON: If everyone would please take their seats. We're going to begin this mornings session.

Yesterday we heard mostly about pandemic influenza and today we're going to hear mostly about seasonal influenza.

Christine?

MS. WALSH: Thank you, Dr. Karron.

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's session 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 now like to read into the public record the conflict of interest statement for today's meeting.

"This brief announcement is in addition to the conflict of interest statement read at the beginning of the meeting on February 27 and will be part of the public record for the Vaccines and Related Biological Products Advisory Committee Meeting on February 28, 2007.

This announcement addresses conflicts of interest for the discussion of Topic 3, Discussion and Recommendation on Strain Selection for the Influenza Virus for the 2007-2008 Season, and Topic 4, a Discussion on Circulating Lineages of Influenza B Virus.

In accordance with 18 U.S.C. Section 208(b)(3), waivers have been granted to: Dr. Robert Couch, Dr. Lisa Jackson, Dr. Ruth Karron, and Dr. John Modlin.

Dr. Seth Hetherington is serving as the industry representative, acting on behalf of all related industry and is employed by Icagen, Incorporated. Industry representatives are not special government employees and do not vote.

In addition, there may be regulated industry or 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.

With regard to FDA's guest speaker for Topic 3, the Agency has determined that the following information is essential:

The following information is being made public to allow the audience to objectively evaluate any presence and/or comments.

Mr. Albert Thomas is employed as Director, Bio-Manufacturing, Sanofi Pasteur in Swithwater, PA.

This conflict of interest 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 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 any firms, its products and, if known, its direct competitors.

Thank you. And Dr. Karron, I turn the meeting over to you.

DR. KARRON: Thank you, Christine.

Our first speaker is Dr. Rakesh Pandey from the FDA.

DR. PANDEY: Thank you, Dr. Karron.

Good morning, everyone. I welcome you all to this mornings meeting.

I am Dr. Rakesh Pandey from the Division of Vaccines Applications at CBER's Office of Vaccines where I have been a scientific reviewer and have been involved with the review of influenza files for the last 12 years.

I'm going to introduce a topic for today's discussion on the composition of 2007-2008 season influenza vaccines.

This meeting has been an annual activity for years. And around this time, this Advisory Committee meeting is convened to get its recommendation on the composition of the influenza virus vaccine for the next season in the United States.

So why do we change influenza vaccines annually? Influenza vaccine is probably the most widely used human vaccine in the United States and millions of doses are produced and used year after year.

Influenza vaccines do not give long lasting immunity and generally do not protect well against the strains that are not included in the vaccine. So as listed on this slide, efficacy of the influenza vaccine is related to two things:

The efficacy of the influenza vaccine is considered to be related to vaccine potency. That's the amount of hemagglutinin antigen present in the inactivated vaccine, which is measured by SRID, or a single radial immunodiffusion acid. The antigen content relates to the immune-SRID, which is measured by the immune response seen in HAI assay.

Also, the efficacy is related to the match of HA and NA antigens to those of the circulating strains. The HA and NA antigens keep on changing continuously, and that is why influenza virus is considered to be a moving target as far as the vaccines are concerned.

Influenza vaccines were licensed in 1945 for the first time in the United States. And, in fact, within two years of their use, evidence came up for reduced vaccine effectiveness because of the antigenic drift.

In order to ensure the effectiveness of influenza vaccine from season to season, we review the antigenic composition of the vaccine every year and change one or more strains as need, as and if needed.

However, in order to come to a conclusion that a change in vaccine composition is warranted, these are the four questions that we need to answer.

First, we need to know if there are new influenza vaccines out there that are antigenically different from the ones in the vaccines. And for this purpose, the WHO, CDC, and other agencies are involved in global surveillance, which is a collaborative effort to monitor the emergence of new influenza viruses that might be showing an antigenic drift or shift.

Then, we need to know if these new viruses are actually circulating and spreading from one geographical location to another in human populations. Are such viruses confined to one geographical location only? Many times we may see that a new isolate may appear in one location and then it might just simply disappear. So we may not have to worry about those viruses.

Now, if an answer to these first two questions is yes, then we need to know if the currently used vaccines work well against the new isolates. So serological studies are conducted to compare the inhibition of these isolates against serum obtained from those vaccinated with the current vaccine. And this is to answer if the current vaccine strain is well matched or whether it will work against the newly identified strains.

Finally, if we determine that the current vaccine strains do not match well against the new isolates, then we need to ask, is a suitable vaccine candidate available for including in the next seasons formulation.

So, if you make a recommendation to include a new strain in the vaccine and it does not grow well in eggs, it would not help much, since all the currently licensed vaccines in U.S. are made in chicken eggs.

On this slide we have listed the recommendations for 2006, 2006-2007 influenza season. Those are the vaccines that are currently being used.

Last season was one of the few in recent years when we had two strain changes, and those were the H3N2 and the B strain. And as is shown on the slides, you can see the ones in the red are the actual strains that were used by the manufacturer of influenza vaccines in the United States.

Just a few years ago, the number of influenza vaccine manufacturers with inactivated license vaccine had gone down from four to two, leaving only Fluzone and Fluvirin on the market. This happened when King Pharmaceuticals and Wyeth went out of the influenza vaccine business.

However, since then we have made a lot of progress. We have overcome the shortage situation of 2004 and we have two new inactivated vaccines available for use.

Besides Fluzone and Fluvirin, which have been on the market for quite some time, we have now GSK's Fluarix and ID Biomedical's FluLaval licensed. GSK's Fluarix vaccine was licensed in 2005, and last year we licensed FluLival.

Besides these four inactivated vaccines, we also have MedImmune's live attenuated vaccine, FluMist, which was licensed in 2003.

This slide shows the time line for vaccine production, and it also highlights the activities relative to influenza vaccine manufacturer, which go on for the whole year, in order for us to have vaccine for use during the September to January time frame.

For the vaccines to be available for use, one of the other production activities, as I said before, goes on for the entire year.

New isolates are made available to the manufacturers by CDC and CBER throughout the year, and the manufacturers keep working on them to make sure they are usable in case they are recommended for use in the vaccines.

Then from ordering the eggs to making the monovalent strains for use in the vaccines in the U.S., the activities go on from January and sometimes up to the late end of fall. And to that point, they will switch over to the strains for making monovalent for use in the southern hemisphere vaccine.

So the formulation of trivalent vaccines usually starts somewhere around June, and the activities related to filling, testing, and release of the vaccine could continue until the end of the year.

And finally, any, sometime around July when the animal strain change supplements are approved, the distribution activities begin, and they continue for the next few months.

Now, in order for all of this to happen, the support activities go on for the whole year. Around this time of the year, WHO, U.S. Public Health Services, and VRBPAC gives recommendation. Then towards fall, the Southern Hemisphere recommendations come out.

And the whole process of surveillance, identification of new relevant strain, and preparation of regents continues for the entire year.

So basically, even though we consider influenza vaccines as seasonal vaccines, they are seasonal only in the sense of use. In the sense of manufacturing, they are essentially a non-stop process with hardly a break in manufacturing.

So time taken for the trivalent vaccine lot to be available after a strain change, now since there can be a significant impact of the Committee recommendations on the manufacturing process, later in this session, the industry representative will go over this thing in more detail.

As you can see from the slide, it can take up to six to eight weeks for the manufacturers to optimize a strain for production after it has been made available to them. So any delay in identifying the strains to be used in the vaccine could delay the availability of vaccine in the fall.

Now, another great limiting step in this process is the availability of reagents needed for the manufacturers to standardize the vaccine and the assigned potency value.

Now, although the manufacturers may start production of some of the vaccines monovalent at risk, a timely VRBPAC recommendation is critical for trivalent vaccine to be available in fall, in time.

Basically, it could take up to 18 to 20 weeks for the first trivalent lots to become available after the change in a single strain and the recommendations. If there is an additional strain change, it could delay the process by another few weeks. And if all three strains have to be changed, the situation could become really challenging.

The WHO held its meeting for the 2007-2008 Northern Hemisphere formulation from the 11 to 14 of this month, where they reviewing the surveillance information and the information on antigenic and genetic characteristics of the viruses circulating around the globe. They also reviewed the serological data as how the circulating viruses, isolated from different parts of the world, react against the serum obtained from currently used vaccines.

Now, based n the discussions, WHO gave the recommendations for the vaccine composition for the 2007-2008 season for the Northern Hemisphere, and that's published on their website, which is listed here.

And I also have that listed on this slide, the recommendation basically reads, "It is recommended that the vaccines to be used in the 2008 Northern Hemisphere winter contain the following:

An A/Solomon Islands/3/2006 (H1N1)-like virus.

An A/Wisconsin/67/2005/(H3N2)-like virus.

And a B/Malaysia/2506/2004-like virus."

And out of these three, only the H1N1 A/Solomon Islands is a new recommendation. The other two have been from the last years vaccine.

Finally, the agenda for the Committee would be to review the surveillance data on epidemiology and antigenic characteristics, and the serological responses to the vaccine, and the availability of candidate strains and reagents, which would be presented by the next few speakers, and to discuss which strains should be recommended for the antigenic composition of the 2007-2008 influenza virus vaccine.

So, at this point I will stop and, unless there are any questions from me, I'll turn it over to the next speaker.

DR. KARRON: Our next speaker will be Dr. Anthony Fiore from the CDC.

DR. FIORE: Hi, I'm Anthony Fiore from the CDC, Influenza Division. I'm standing in for Joe Bresee, and I am going to provide an update on the influenza surveillance for this current flu season.

This depicts the various different surveillance systems which come into CDC, are compiled, and then go back out again to you, and the public, and healthcare practitioners, and so on.

We do conduct laboratory based surveillance with strains coming into CDC from a variety of different sources to be characterized.

We have Sentinel Provider surveillance, which consists of a variety of different providers who provide isolates and clinical information on patients that come in with influenza-like illness.

Population-based hospitalization surveillance, which is conducted in a number of different areas in the U.S., which were, those of you who went to David Shay's talk yesterday saw the map of where our population-based hospitalization surveillance is conducted.

Assessments by state and territorial epidemiologists about flu activity weekly from their states.

A pediatric mortality surveillance system, which was setup in 2003-2004 after the early and fairly severe season that we had that year.

And vital statistics and registrars who provide us with data for 122 cities mortality surveillance for pneumonia and influenza, and sorry, influenza, right.

And the next five slides are going to sort of give you a time lapse version of this years flu season. The season started off in November with a number of interesting small town, apparently school-based outbreaks in North Carolina, some influenza-B outbreaks. And over the course of the next few weeks, it kind of evolved to be a southeast-U.S. phenomenon.

And again, mostly concentrated in the southeast toward the end of last year, and then as we came into this year, spread throughout other states.

And you can see by the color-coding, regional activity as assessed by the state and territorial epidemiologists is depicted in blue, and widespread activity in red.

And then this past week you can see where we have a widespread or regional activity in most states.

This is the Sentinel Providers surveillance system. And what you see here is the last two years tracked by the solid lines, and this current flu season by the red line with the diamonds on it. And it predicts the percentage of visits to the Sentinel Providers for influenza-like illness by week. And as you can, there's a dotted baseline there which depicts what the average is thought to be, the national baseline, and then over the course of the year the people who come in for INI is shown by the lines.

And what you can see here is this season, which is depicted up to about week seven or so, we have a fairly typical percentage of visits for ILI. It tracks more or less the same as the last two years.

Here is information from our influenza hospitalization surveillance. This is laboratory-confirmed influenza in the new vaccine surveillance network, which is three sites around the U.S. And it tracks hospitalizations due to lab-confirmed influenza among zero to four year olds.

And this shows the last six seasons, the previous five seasons by the solid lines, and this season by the incomplete red line with the red circles on it. And what it shows you is that this season is very similar to four out of the last five seasons, that the blue line is the 2003-2004 season, which had early reports of illness among children, severe illness.

This is the Emerging Infections Program, this is, I believe, eight sites around the country. It again was shown yesterday by David Shay on the map. Again, this is lab-confirmed influenza vaccinations. This is just shown for the last three years, sorry for the last four years counting this one. And this shows not only 0 to 4 year olds, but also 5 to 17 year olds. The younger children are shown with solid lines, the older children with the dotted lines. And you can see this season's information, again with the red lines with the circles showing the 0 to 4 year olds, and then right there on the x-axis the dashed red line showing the 5 to 17 year olds. And this is a cumulative number of hospitalizations and that's why the number go up like that.

And then this is the pneumonia and influenza mortality surveillance system, often called the 122 city surveillance, showing the pneumonia and influenza mortality as reported by death certificates and registrars around the country. What is depicted by the black lines, the top black line is the epidemic threshold, and the bottom, the bottom black line is the seasonal baseline. And then the actual reports are tracked with the red lines. So you can see, we haven't actually spiked up over the epidemic threshold in this flu season.

Now, I mentioned that in 2003-2004 we started a pediatric death surveillance. And as of February 15 of this year, CDC has received 15 reports of influenza-associated pediatric deaths. Ten of these were among children five years of age or older. Three had underlying medical conditions. Five had no known underlying conditions. And two's previous health status are yet unknown. Nine of these children were unvaccinated, speaking strictly of the ten that were 5 years of age or older.

And then as compared to the last couple of flu seasons, in 2004-05, 44 deaths were reported. In 2005-2006, 48 deaths, and then going back to 2003-2004, 153 deaths.

This is the Collaborating Laboratories Strain Surveillance Summary for 2006-2007. And what you see depicted here is, let's see, yes, of the 10,458 viruses typed, 84 percent of them had been Influenza-A, 16 percent, B. Of the 28 percent of those that have been sub-typed, 87 of the As were H1, and 13 percent of the As were H3. And we've seen an increasing percentage in the last couple of weeks of H3s.

Switching over to anti-viral resistence data generated so far this year and as compared to last year, you'll recall last year there was the identification of high levels of adamantine-resistants among isolates, starting in 2005. And for comparison, in the last flu season, two out of eight, or 25 percent of the H1s are adamantine-resistant. And 192 of 209 H3s tested, or 92 percent, were adamantine-resistant.

There are fewer of these adamantine-resistants, or the proportion of the adamantine-resistants is lower this year, but there is still a considerable amount of it. And global surveillance so far, three percent of the 199 H1s tested have been resistant. Forty-four percent of the H3s tested have been resistant. Among U.S. isolates, the numbers are similar, one percent of H1s and 33 percent of the H3N2s, although we don't have very many of those tested so far.

As far as resistance to neuraminidase inhibitors, oseltamivir or zanaminivir, among the isolates tested so far since 2005, and that's 437 isolates, there have been none identified as being resistant.

And just to update you on issues that will come up in front of the ACIP over the next year, and also came up at the meeting last week, of course the new vaccine strains, which we're here today to discuss. The recommendation, the ACIP recommended, as you would expect, that adamantine not be used in treatment of influenza-A viruses.

We did make one change to the recommendations. And this had to do with harmonizing the American Academy of Pediatrics and the ACIP recommendations for young children in a specific subset. As you would recall, young children who are six months to less than nine years of age who get vaccinated for the first time are supposed to get two doses. What, the disharmony occurred had to do with children who only got inadvertently only got one dose in that first year. And what the ACIP recommendations have been changed to, and what the AAP recommendations already were was that those children who got one dose in their first year of being vaccinated in that age range, in the second year of being vaccinated should go ahead and get two doses.

Now, what was not changed was the age groups or the risk groups that were recommended for routine vaccination. And there is going to be continuing discussion of advancing the recommendations to include other age-groups or risk groups over the next several years. But for this coming flu season, the recommendations have not changes as far as that goes.

And that's all I have to say. I can either take questions now or do you want to wait after the session, either way.

DR. KARRON: We can take questions now if there are any.

Okay. Thank you, Dr. Fiore.

Our next speaker is Dr. Nancy Cox from the CDC, who will tell us about worldwide surveillance.

DR. COX: Thanks very much. It's a pleasure to be here. And I'll try to make my presentation as comprehensible as possible. For those of you sitting in the back of the room, I know you won't be able to see the slides, so there are plenty of seats in front if you would like to move up and have a better view.

I'm going to be talking about what we're seeing globally with respect to influenza activity and influenza viruses.

I will be talking about hemagglutination-inhibition data, with post-infection ferret sera.

I will be talking about the genetic analysis of the HA and the NA genes of the viruses.

And I will not be talking too much about the serology, the post-vaccine serology unless you have specific questions, based on the data in your packages, because I think that data will be covered by Dr. Zhiping Ye later on.

Influenza H1 activity has been relatively light, although influenza H1 viruses have predominated in the United States and a few other countries, and in a few other countries have caused moderate outbreaks, or moderate levels of activity. But really, H1N1 hasn't caused many problems, except in the United States.

If we look globally at the number of H1N1 viruses that were isolated within WHO's global influenza surveillance network, we can see that for 2006-2007, which is shown in the blue line, we really have relatively low numbers of H1 activities isolated worldwide.

What that means, of course, is that we have a limited amount of data. If we look at the viruses that are characterized in all of the WHO collaborating centers, our own included here, we had the largest number of H1 viruses characterized during this period from September to the current time. But there were also a significant number characterized in Australia.

So, what we see is that we have, the majority of viruses are still New Caledonia like, but there are a number of low reactors. And they were seen particularly at the WHO collaborating center in London, where they had a higher proportion of low reactors.

I apologize for the HI table, but I think it's important for us to go over carefully and to understand the kind of data that we're looking at on a weekly basis.

All of the WHO collaborating centers generate one or two HI tests per week on average, or at least we do. And we often have to retest viruses if they appear to be low and we want to make sure that they're actually low reactors.

So what we have here are the reference antigens, starting with New Caledonia, the vaccine strain, and we have a number of other reference antigens, which we have put into ferrets to develop post-reaction ferret serum.

So these across the top are the corresponding reference ferret antisera. And what we're really looking for is a lower level of reactivity than we see with the homologous virus interacting with the homologous serum.

So here you see a number of viruses, the Kentucky, the Virginia, and the St. Petersburg, which are very well inhibited, equally well-inhibited by anti-serum to the New Caledonia vaccine strain, as New Caledonia is itself.

Starting here with the Hawaii/15/2001 strain, we see quite a marked reduction in the ability of anti-serum to the New Caledonia virus to inhibit hemagglutination of this virus.

And if we look across here, we can see that it's true not only for the New Caledonia serum, but for the Kentucky serum and the Virginia serum, as well.

So Hawaii/15/2001 was one of the first viruses that we saw which had a specific amino acid change and a corresponding difference in activity with the New Caledonia serum.

When we put the Hawaii virus into ferrets, we got a homologous titer of 320 and viruses like it in having that same amino acid change, which I'll talk about in more detail later, are well-inhibited by anti-serum to this particular virus.

So we've, we've been able to really distinguish viruses that have a change from lysine to glutamic acid at amino acid 144. And that is antigenically important area of the globular head of the HA.

So we can really group these viruses into two groups, those that are well inhibited by the New Caledonia serum and those that are not, but that when we make antiserum to these viruses, they cover viruses with that 144 change.

So here we have a lot of test antigens that have been isolated, many of them during the last couple of months in December and January in the United States. And really the majority of them are well-inhibited by the New Caledonia serum. However, you can see this Texas virus, which is not, is well-inhibited by serum to these viruses, including the Solomon Islands reference strain, which I'll talk about later.

Likewise, if we look at viruses from Asia, we see that there are some viruses that are still well-inhibited by the New Caledonia serum, but we see a larger proportion of viruses from Asia, which fall into this other group with the change from lysine to glutamic acid at amino acid 144, and they're well inhibited by antisera to these reference strains, including the first one, the Hawaii/5, and the Solomon Islands/2006, which is the current reference strain.

I should note here that this is Solomon Islands IVR 145, which is a high growth re-assortant, which has been produced in Australia and circulated to the other collaborating centers in time for us to produce antiserum against this particular virus.

So, in summary, while we still have a number of viruses which are well-inhibited by the New Caledonia antiserum, we see a growing proportion of viruses, particularly in Asia, which do have the signature change and which are better inhibited by antiserum to viruses like the Solomon Islands virus.

If you look at just our own CDC data, you can see quite clearly that although the proportion of viruses that are, that have a reduced titer to New Caledonia is really a moderate proportion, and actually somewhat less than we saw during the period of April to September. We have, we see viruses predominantly from the United States during this October to February period, but if you look at where we're seeing the low-reacting viruses, it's in Asia and Central and South America. And so, specifically, the majority of the viruses that have been sent to us from Asia during this period from October to February, as well as the period from April to September 2006, from Asia, are low-reactors.

Now, I apologize. This is going to be very difficult for you to see in the back of the room, and I realize that your handouts are not color coded. The color coding actually helps a great deal because we've color coded the viruses by month of isolation so that we could really depict where we're seeing, if we're seeing a trend toward more viruses being in one group or another.

What you can see here is that the H1N1 viruses genetically divide into two distinct groups, or clades, Clade 1 here and Clade 2 at the top. These changes did not confer antigenic changes on the viruses, however. So we were not able to distinguish viruses from these two groups until we started seeing the change 144, the lysine, oh, this is the lysine to arginine, which is a different change, until we started seeing the lysine to glutamic acid change pop up.

And it's a little bit, well, it's really quite interesting actually, because we're seeing that this change is occurring in separate subgroups. So we see the change up here independently occurring, or so it appears to be independently occurring, and in this group here, where the Solomon Islands referenced strain and reassortant viruses.

And so no matter where the virus is on the tree, if it has that lysine to glutamic acid change at 144, it is poorly inhibited by the New Caledonia serum and well inhibited by antiserum to the Solomon Islands.

And so what, when you see a change that is occurring, apparently independently in different parts of the tree, what you tend to think is there may be selective pressure in the population to select that particular amino acid at that position.

If we look at the evolutionary tree of the neuraminidase genes, you can see that we don't have as many strains on this tree and we haven't really sequenced quite as many neuraminidase genes. We'll be sequencing more in the future because we'll be trying to do more high throughput sequencing, but you'll see that the Solomon Islands virus here is in the Clade 2 just, neuraminidase is in the Clade 2 just as the hemagglutinin was.

So there is correlation between where in the tree the hemagglutinin is and neuraminidase. And you'll see that we do have a number of viruses -- I should back up one and say that we do, we're still seeing quite a few viruses from the U.S., recent viruses from the U.S., in this group down here, which doesn't, and most of the viruses in Clade 1 do not have that 144 change.

So here is the old vaccine strain, which I should point out is the 99-strain. The Solomon Islands reference strain in Hong Kong/2652 is another of the reference strains that has been important in our understanding of what's going on with the H1N1 viruses.

So if we look, step back and summarize what we've seen globally, we can say that H1N1 viruses have been circulating at a low level, but sporadically in Canada, South America, and the Russian Federation.

H1N1 viruses, however, predominated in the United States. Many H1N1 viruses have remained antigenically like the New Caledonia vaccine strain, but a proportion of recent H1 viruses, particularly those from Asia, have been antigenically distinguishable from the vaccine strain. And, as I mentioned, these viruses were more closely related to early-Hawaii/15 strains and then these other reference strains, specific, and I want to note the Solomon Islands/3/2006. And the majority of those viruses do have this mutation that I spoke about.

Okay. We'll move on to the H3N2 viruses. H3N2 activity has been very moderate in many parts of the world. We started out with some H3N2 sporadic activity in a lot of the world, and then some increasing activity in Scandinavia. And then by January, we were seeing a number of European countries with slightly greater intensity of H3N2 activity, and Canada was having quite a bit of H3N2 activity.

If we look at the number of viruses that were H3N2 viruses that were isolated with the Global Influences Surveillance Network, again looking at the blue line here, we see that there really were not many viruses isolated globally compared to previous years, where we had a lot more H3 activity than we had this past season.

And this is reflected in this table here. If you look at the H3 table, you can see that between February and September of `06, there were over 1,000 viruses isolated, between October `06 and January `07 we've only had, or had only 319 viruses characterized by all of the four WHO collaborating centers. And that was as of the second week in February.

This is our most recent H3HI table. And again, I'll walk you through it fairly slowly and carefully. We have included only one HI Table in our package this year because we felt that it was the best summary of what's going on and includes the Ferret Sera from some of the most recent viruses. And these data were not even available from the WHO Meeting, so these are very fresh data.

So we'll look here, starting from the left with the California/07. This is an old vaccine strain. The current vaccine strain, Wisconsin/67/2005 and its corresponding high-growth reassortant, 161B, and its corresponding antiserum.

We have other viruses, including one from the U.S., Kentucky, one from South America, Santiago, a Florida, and then you'll see the last two antigens, Nepal/921 and Canada/1212, which I'll talk about later.

The Canada/1212 was actually used as a serology antigen, as was the Santiago virus. And you'll see those data later.

Now, what we've been seeing over time is that there's an increasing number of viruses that are poorly inhibited by antiserum to the wild type Wisconsin/67 strain or the actual vaccine strain, R161B. And you'll see that these viruses, however, react quite poorly with all of the ferret antisera that we've been able to generate.

Now, this table just shows a small number of the failed antisera that we've been able to generate, either using cell isolates or egg isolates. And I'll talk more about how many egg isolates we've actually had in hand later on, because that's increased significantly and I think it's important for the Committee to know.

But what we can see here in this reference panel is we don't get a great deal of differentiation. It's not like the situation that we saw for the H1N1 viruses where you could see that clearly there were viruses which were low on this side, and then a group of viruses that were high on this side, and the antisera that corresponded to those viruses didn't inhibit the other viruses very well.

So, we have here what is a much less clear-cut situation. We have really struggled with what these low-reactors mean. And there is a tendency right now to think that these, many of these low reactors are actually low avid viruses in the hemagglutination-inhibition test. And the way that you can eyeball this and determine whether they are low avid or not is to look at what the difference is in the titer between the virus and the reference strain that you're looking at, and it if a four-fold difference, just multiply everything by four; if it's a 16-fold difference, multiply all these values by 16, and see whether you get a pattern that's similar to that for the reference strain.

And what we've determined, and I'll show you a graphic representation, is that many of these appear to be low-avid viruses. Now, we don't really understand as much as we would like to about avidity. And we think that, well we know that the receptor binding properties of H3 viruses have been evolving over time, and that some changes occurred a few years ago that we believe have affected our ability to discriminate viruses using the hemagglutination inhibition test.

Nevertheless, whether we're looking at viruses from the U.S., Canada, Europe, or Asia, we do see viruses that are well inhibited by the Wisconsin antiserum. And if we look at these two most recent antisera, we see that while they cover some of the low reactors somewhat better than the Wisconsin antisera, or both that the antisera do, they really, you still are seeing a lot of viruses which have a 4-fold or greater reduced titer against these strains.

The Nepal/921 is an egg-isolate. And that was put into ferrets. And the Canada/1212, which is related genetically is a cell-based virus put into ferrets. And whether you have an egg isolate or a cell based isolate, you still see these low reactors.

So if we don't take avidity into account and we just strictly look at the number of viruses that are, that have a 4-fold or greater reduced titer to the antiserum to the vaccine strain, we see that we have about 59 percent of viruses which are low reactors. And we do have relatively small numbers compared to what we've had in the past, but these viruses are geographically spread, so we're seeing low reactors around the world, but they are behaving similarly in perhaps being low avid.

We have been trying very hard to improve ways to look at the HI data, the ferret data, and these methods will be extended to look at the human serologic data. Derek Smith from the University of Cambridge, has developed a field which he has called antigenic cartography. He's pioneered this using data from the WHO Global Influenza Surveillance Network. And basically what it does is mathematically calculate different distances, antigenic differences between viruses and then display this information graphically.

Well, if you're accustomed to looking at HI tables and you look at them everyday, your brain is actually doing a lot of it. It's amazing what, how much your brain can actually do on its own, but it's nice to have a visual display, especially when you're getting up in front of a group like this.

So Derek has, here is some of the old vaccine strains, the Sydney, the Wyoming, the New York, here is our current vaccine strain here in blue, and some virus strains that were used previously. And we're starting to see some outliers. You want to see some viruses really clustering around your vaccine strain. And if you look over time, which I haven't done here because we don't have enough time to review old data, but you would see that for the time that the vaccine, this vaccine was used, that there were, the majority of viruses were really clustering very closely around this. And each one of these squares represent a two-fold difference in antibody title.

And then if you look at our CDC data generated from viruses that were isolated in December, you can see a scatter somewhat away from the list of Wisconsin/67 virus. And then if you look at January again, you see the scatter.

Derek has written a program which will account for avidity, and so he's incorporated into his program an avidity correction, and so I'll just go through those same data. And what you see is that the viruses are actually, when the avidity corrections are done, the viruses are actually pulled closer to the Wisconsin/67 vaccine strain, although there are some viruses which are fairly far away, getting to be 4-fold, or 8-fold, or greater. But certainly the differences are less for the July data.

For the December data, we are still seeing viruses out here. This is the Canada virus, Canada virus here, and Nepal virus here. So these viruses are really not being pulled toward the Canada and Nepal.

And then for the few January viruses that we've had in hand to analyze, we're still seeing this scatter over here, but a few more viruses that are clustering around Canada and Nepal.

If we look at the genetic data, we've actually sequenced a fairly large number of HA genes. Even in the last couple of weeks, we've been able to sequence well over 50 HA genes from recent viruses. And we were very interested because we were starting to see two primary groups emerging. One here, which is represented by the Brisbane/9/2006 virus, which was isolated in June during the Southern Hemisphere season, and another which is represented up here by Nepal/921, which I pointed out on the HI table, and the Canada/1212 virus, which was isolated in December.

I should mention that a lot of the data on here has, the data have been provided, the sequence data have been provided by some of the national labs. The Canada lab provided us with sequence data. And we actually get data exchanged among the four collaborating centers and as many of the National Influenza Centers as possible. So we really have a lot more data than I'm showing you on here, but I'm just really trying to demonstrate what we're seeing.

Again, the most recent viruses are shown in purple. The December viruses in pink, November in orange. So you're looking for the more intense red colors here to see where the trends are. And what we're seeing is that the most recent viruses, and it is about 50/50, slightly more in this group than this group, but about 50/50 distribution of recent HA's into this group here, which has characteristic arginine to glycine change at 142, and then some additional changes. You can see that there are subgroups here with additional changes. We do have a lot of viruses, and they are from Asia, from the U.S., and Europe in this group as well.

Likewise, if we look at the NA genes, here is our Wisconsin vaccine strain. Here are the two groups that I was referring to before, the Brisbane Group and the Nepal Group. And you can see that the viruses, again, they're color-coded so that the most recent viruses are shown in purple and pink, and the NA's are segregating along with the HA's.

So, in summary, Influenza A H3N2 viruses have been difficult to analyze. However, activity caused by H3N2 viruses was low, generally speaking, around the world. However, there were outbreaks during the period September to the current time in Madagascar, Canada and a number of European countries.

Many of the viruses globally were antigenically closely related to Wisconsin and Hiroshima. The Hiroshima virus is the vaccine virus used in Japan. But an increasing proportion of viruses was antigenically distinguishable from the vaccine virus, viruses.

And increased heterogeneity was observed in the HA sequences from recent viruses and no emergent antigenic variant group was identified. And that is, I think, the key to focus on. If you'll remember, we really didn't see the same king of dichotomy in the ferret antisera that we saw with the H1N1s.

Okay. We'll move on to Influenza B Viruses. Influenza B has also circulated at low levels. There really hasn't been much activity, except for sporadic activity in a few school outbreaks and so on. But if we look overall, the Influenza B viruses have circulated at relatively low levels, which is shown very clearly here in the number of Influenza B viruses isolated in the WHO's Global Influenza Surveillance Network. And you see almost a baseline the number of Influenza B viruses.

So the total number of viruses that we've had to look at is even smaller. And we have them divided into two distinct lineages. And of course, this afternoon's discussion will concentrate on the fact that we have these two distinct lineages of Influenza B viruses.

So these are the four WHO Collaborating Centers, and sorry that this didn't get, that it must've got transposed. But you can see that in Australia they had 95 percent of viruses being Victoria-lineage. In the U.S., we had more B viruses than any other Collaborating Center to examine. And 76 percent of the viruses that we examined were Victoria-lineage viruses, thus matching the vaccine lineage.

But we're seeing some low reactors, nothing really different than what we've seen before, as I'll show you in the next table. Here, we have color-coded in this yellow mustard color the viruses that are on the B/Yamagata lineage. And to the right we have the viruses that are on the B Victoria lineage, including the Malaysia vaccine strain, which has been used in all countries by all manufacturers.

These are the most recent viruses we have. We have some from December, just I think only one from January. We'll expect to get some more later on. But we have viruses from the U.S. and from Asia, and we see that if we look at the homologous titer here of 1280, we do see a number of viruses that are reduced in titer. But that has been something that we have been seeing for a long time.

And what we know about the B Victoria viruses is that once they are isolated in eggs, they lose an important glycosolation site, which is right up at the tip of the hemagglutinin. And once they lose that glycosolation site and are put into ferrets, they induce antibody that is not as broadly cross reactive as the cell counterpart. And we have put many viruses into eggs and have found that even if you retain a glycosolation for one or two passages, if you pass it sequentially you eventually lose that glycosolation site. So this has been problematic for egg-based manufacturers.

So, if we look at what we are seeing at CDC, we're certainly seeing the majority of viruses in the Victoria lineage. That is, in the same lineage as the vaccine strain. A smaller number, which we're calling Florida-like, which would be a potential vaccine strain if we were to move to the Yamagata lineage, but of course, globally and within the strains that we've had from the USA, we've certainly seen a predominance of the Victoria like viruses.

If we look at the HA genes, now we're looking at the Yamagata lineage. And this is just for full information, not really as pertinent to our discussion today, but just to let you know that if we were to have, to move to the Yamagata lineage, we have some very good vaccine candidate strains, which produce antisera that do a very good job at inhibiting viruses on that lineage. So this is the old vaccine strain, so we do have contemporary egg run viruses that could be used, should that become necessary in the next few years.

For the Victoria lineage, you see that we really have much slower genetic change than we have seen for the H3N2 viruses. And this is typically what we see. The HA of the Influenza B virus does evolve more slowly.

So here we have the vaccine strain, Malaysia/2506/2004, and then the viruses that are more contemporary that we've been seeing. And you'll see some of the viruses that are egg isolates. As I've mentioned, they've all lost the glycosolation site. And we have a number of them. They are noted by the hatch mark, so you can see that we really do have quite a lot of egg isolate.

The neuraminidase genes are also not changing all that much. Here is the B Malaysia vaccine strain. Here are a number of the viruses, and again, you can see that we do have a number of egg isolates shown here.

So, in summary, Influenza B viruses have circulated in many countries; however, outbreaks or large outbreaks, apart from institutional outbreaks, have not been reported since September 2006 and January 2007.

Both lineages have continued to circulate, but the Victoria lineage viruses have predominated. And if you look at the WHO data overall, it was about 82 percent of viruses were of the B Victoria lineage.

The Yamagata lineage viruses were closely related to those strains that I mentioned that were egg isolates and could be used. And most B Victoria lineage viruses were antigenic closely related to B Malaysia, taking into account that we have an egg isolate that's gone into ferrets and we're looking at mostly cell-based, cell isolated viruses.

Now, this is just to show you that we've been working very hard to increase the number of egg-grown viruses. And this is shown by fiscal year. And our fiscal year and our influenza seasons sort of start at the same time officially on October 1. So we have really increased the number of egg isolates that would be available to manufacturers, should that particular egg isolate look like it's an appropriate vaccine strain.

So, for the last fiscal year, which ended September 30, 2006, we had 255 egg isolates distributed fairly well across the different types and subtypes. Now, I should mention that obtaining H3N2 isolates has been exceedingly difficult. And if we put 100 clinical specimens into eggs or into kidney cells and then pass on to eggs, we're likely to get out five if we're lucky. So it really does require a lot of work for the H3N2 subgroup of viruses.

I think I'll stop there and take questions. If anyone has questions about the human serology data that you see in the package, I can answer those questions now because I have the slides available.

DR. COUCH: Just one quick question, Nancy, for clarification to see if my assumptions are correct. When you say a low reactor, you spent a good amount of time on the H3 antigens with the ferret sera, a low reactor would be low reactor with say the Wisconsin antisera. But with your other ferret sera, no reactions that gave it a different identity? Is that --

DR. COX: That's correct.

DR. COUCH: -- what low reactor means?

DR. COX: So, okay, so when I say low reactor in my table, I'm really talking about the number of viruses that have a 4-fold or greater reduced titer to the antiserum to the vaccine strain. But when I was looking at the H3s, we were looking very specifically because we're seeing this pattern of low reactors across the board. So, in many cases, your low reactor is a low reactor across the board and you don't see better inhibition using any of the other ferret antiserum, or only moderately better inhibition.

DR. COUCH: I guess being specific, that would count for some of them, but when you say low reactors that reacted only with ferret sera, or is a battery of ferret sera?

DR. COX: We always use a battery of ferret.

DR. COUCH: Only the Wisconsin -- a battery?

DR. COX: We always use --

DR. COUCH: So it's a low reactor across the board?

DR. COX: We always use a battery, but in order to really be specific when we're talking of the WHO Meetings, we really are looking at the number of viruses or the proportion of viruses that are low reactors to ferret antiserum to the vaccine strain. So that's what is in those tables. But when we're looking more broadly, we're looking for patterns of reactivity. And so many, so for the H3s, the majority of the low reactors are low reactors across the board to the battery of antisera. And we never test viruses on their own. We always test with a whole battery of ferret sera.

I don't know if that answered your question.

Ruth?

DR. KARRON: Yes, two questions. One is so can you contrast these H3N2 viruses where you say there is no emergent antigenic variant group, and I assume that's based on the ferret antisera data, with say the situation we were in several years ago with H3N2 Fujian. There we saw an emergent new variant, is that correct?

DR. COX: We saw there an emergent new variant, which we could clearly see using ferret antisera. And we could see, we could see the two-way, 4-fold or greater difference. So antiserum to the Fujian strain didn't inhibit the old viruses as well, and antibody to the previous vaccine didn't inhibit the Fujian strain. So we had the two-way, 4-fold difference, which was very clear. And that also corresponded to changes that we saw in the genetic data.

DR. KARRON: Thank you. And then just a question either for you or for Dr. Fiore. Would you happen to know the school-based outbreaks of B in North Carolina, whether those were Yamagata or Victoria lineage. Is that something you would know?

DR. COX: The North Carolina, and Tony correct me if I'm incorrect, but the majority of those viruses were, they were Yamagata. And that was an early outbreak before the season really got rolling. And so we were feeling rather uncomfortable with the fact that they looked, that they were Yamagata lineage viruses and we had Victoria. But then it didn't hold true for the rest of the season.

DR. KARRON: Thank you. I think we'll go on now to Angela Owens from DoD who is going to talk to us about vaccine effectiveness and responses.

MS. OWENS: Hi. I will also fill in for Luke Daum as well, as we provide the sequence analysis overview.

Let me go back, sorry. We are actually part of an overall Department of Defense Global Influenza Surveillance Program which is a WHO collaborating laboratory and it has two parts.

One part is Sentinel site surveillance that is managed at the Air Force Institute for Operational health, and another part is a population based surveillance that takes base at eight training camps and is managed at the Naval Health Research Center in San Diego.

For our Sentinel sites we have, thank you, 43 medical treatment facilities that are located throughout the service. We also have 22 sites associated with four separate DoD Overseas Research Facilities. We provide naval wash kits, collection material, education material, shipping supplies, so they can send us specimens on a weekly basis. We contact them every week. We expect six to ten specimens every week from patients with a fever greater than or equal to 100.5 and/or a cough or sore throat. We also request an influenza surveillance questionnaire to be completed that describes vaccination history, travel history, any additional symptoms other than cough and sore throat.

We have so far collected about 1,200 specimens this season. Twenty-six percent have been influenza positive by culture and the majority have been Influenza A. We also test for adenovirus, enterovirus, RSV, pair influenza, and HSV, which is a background virus.

About 93 percent of our isolates have been sub-typed at this time. The majority have been Influenza A, H1, and the majority of B have been from the B Victoria lineage.

This is Luke's portion, so let me describe this. For Influenza B, the hemagglutinin genes of over 40 DoD Influenza B viruses were analyzed from the influenza strains obtained from summer of 2006 to present. Of these isolates, five strains were Yamagata like, with the remaining isolates being B Victoria like.

B Yamagata like viruses are antigenically and genetically distinct from the current vaccine strain, which is the B Malaysia vaccine strain. These five strains were collected from Peru, Maryland, and Illinois.

The remaining B viruses were B Victoria and shared 99 to 99.6 percent sequence identity to the current B Malaysia vaccine strain.

Depicted in this slide is a phylogenetic analysis of influenza strains represented by 38 Influenza B Victoria like strains and 5 B Yamagata like strains. As you can see, the B Malaysia indicated within the red box is find within the cluster of B Victoria like filled strains collected during the current season.

For Influenza A H3N2, the phylogenetic analysis of Influenza A strains collected during the 2006 through 07 season have been the minority this year, taking a backseat to the H1 subtype. At present, we've collected and sequenced about 70 strains, including a summary outbreak of H3N2 isolates in Nepal depicted as a distinct branch, which is at the top.

Shaded in the circle is Clade of Influenza A H3N2 Viruses that are forming a distinct branch in the current tree of H3 viruses. These viruses show a genetic variances in several nucleotides, inter-defined by the R142G mutation, which forms the distinct branch point, I apologize, let me go back. I think every time I touch this -- I apologize. I've been touching the bottom. About 50 percent of our isolates, including the July outbreak in Nepal are contained within this branch, phylogeny. Indicated in the red box is a current A Wisconsin strain and the older A California vaccine strain. And as you can see from the phylogeny, the A Wisconsin strain shows genetic variance from viruses belonging to this Clade.

And this is the A H1N1 phylogeny of the strains collected during the 2006 through 07 season. They've been the predominant circulating strain, and two distinct Clades have formed with Clade II viruses showing somewhat reduced titers according to the H1 analysis. While the two Clades are currently noted in this phylogeny, isolates obtained in the U.S. are Clade I, which are the current A New Caledonia vaccine strain. Clade II isolates were obtained from South America, which are Peru, Nepal, and one isolate from Saipan. Clade II viruses are defined by several key amino acid changes, some of which are located within the antibody combining isolates.

Again, we are a WHO collaborating lab. All of our information is uploaded to CDC's national surveillance system, and they also receive our isolates of interest.

For vaccine effectiveness, this is a descriptive preliminary review. We looked at patients seeking care from U.S. Military Treatment Facilities from October 1 to 12 February. This includes active duty and DoD beneficiaries.

Our outcome was lab confirmed influenza results by viral culture and isolation. We defined fully vaccinated patients as patients who received the vaccine greater than 14 days prior to the clinic visit date. Any of those patients receiving the vaccine prior to August were defined as unvaccinated unless they received another vaccine later.

Our vaccination data were gathered form the Military Immunization Database, and our influenza surveillance questionnaires. Those are completed at the time of the clinic visit.

A total of 796 specimens were included. 170 of these were influenza. 65 of the 170 hadn't identified vaccination status in the breakdown of FluMist versus the injection. 23 of these patients has a possible vaccine breakthrough, meaning they had the vaccine greater than 14 days prior to the clinic visit in which they obtained the isolate. All age groups were represented.

And the isolates obtained were Influenza A, H3, from California, Texas, and South Korea. And the South Korea information is also described in the sequence data that was mentioned before. H1 was identified in Alabama, Oklahoma, South Carolina, Texas, and Antarctica. And our Influenza B, one B was not sub-typed at this time. It was actually recently collected. And that was both identified from Alabama.

The Naval Health Research Center also provided a vaccine effectiveness review at the eight training camps, which six of them had influenza identified. All of the trainees are vaccinated upon arrival, so their vaccination to identify a fully covered vaccinated person is within the 14 days. They are not fully covered.

And they had 48 isolates. 38 of them of them were unvaccinated. And if you'd like more information regarding this review, please contact NHRC. And I have their information so afterwards I can get with you.

We did not, Geis, of course, is our funding and guidance agency. CHPPM and AMSA provided us with the vaccine information, and Pop. Health, CDC, and of course all of our Sentinel sites. We do expect a full review of vaccine effectiveness at the end of the season. This is a descriptive review. There is no random sample applied yet.

Here is our contact information. And I do apologize that Luke was not here.

Do you have any questions?

Yes?

DR. FARLEY: I was a little surprised at how often you did not know their vaccine status in the first part. Is that because they really weren't all enlisted military? It was families and other people? Because otherwise I would think your record keeping would be very tight.

MS. OWENS: Yes, for Sentinel site surveillance, our population, it's gathered from family clinics, pediatric clinics, hospitals, and ER's. The majority of our population is DoD beneficiaries, and unfortunately for the Navy and the Army it's hard to gather that vaccination status from DoD beneficiaries. That's where we come in play with the influenza surveillance questionnaires. That's why we actually look from the surveillance site, we request that they all fill out that. And unfortunately, a lot of these came from both Sentinel sites and non-Sentinel sites. We had about 22 non-Sentinel sites that submitted specimens this season.

But hopefully at the end of the season also we'll get more information.

DR. KARRON: Any other questions?

Thank you very much.

MS. OWENS: Thanks.

DR. KARRON: Next on the agenda is Dr. Ye who will talk about vaccine responses and availability of strains and reagents.

DR. YE: Thank you.

I think the whole purpose of doing the human serological study is to see whether the HI antibody response to the vaccine strain will confirm that the antibody response to the isolated, newly isolated, viruses reach as Nancy mentioned in her talk.

It seems the serum sample from human was pre-selected to choose the antibody, positive antibody response. So the whole purpose of this study is not to compare the antibody response from different vaccine string, but rather to see, to use this serum to study, to compare the antigenic difference of the newly isolated circulated viruses. Serum sample actually comes from different centers worldwide. As shown on this slide, the serum sample comes from, there are five serum samples that come from different centers. One is from Australia. The serum sample represent the human serum which immunized, was the vaccine that contained New Caledonia for H1N1 and New York/55/2004 for H3N2, particularly for the Austria serum samples.

Then the B strain is Malaysia and that is the same for the rest of the serum sample.

And the other serum samples from Europe, from Japan, and from U.S., the serum sample comes from the human serum, from the human population that immunized the vaccine strain containing New Caledonia for H1N1, and the Wisconsin like strain for H3N2. The actual vaccine strain contains Wisconsin itself or Hiroshima for H3N2 for the European, and also for the Japan. The European contained two serum samples, two serums, one is for Wisconsin itself, another one is for Hiroshima. But what I presented in this study is the same sample for Wisconsin itself. And also to have the same sample from a pediatric.

Okay. The antigen used for human serological study has been carefully chosen. And one is, of course, is the vaccine strain itself. Here is New Caledonia/20/99 for H1N1. And the representative or current vaccine, current strain used for serological study, choosing according to antigenic and genetic characteristics of the strain, but also the geographic characteristics of those viruses.

Here you can see that I have one strain from New England, from Hiroshima, and from Fukushima, and from Asia, also from European. And here I marked a group of the antigen with the asterisk over here as it represents the signature of antigenic differences in amino acid 144, substitution of amino acid from lysine to glutamic acid as mentioned in Nancy's talk. So we will see what the main response of those groups of viruses to the vaccine strain.

And here, this slide shows HI antibody response to H1N1. And this slide I choose a representative serum panels in one of the centers who did the serological study. By the way, there are four or five centers conducting the human serum study, and this is only one I choose as positive representative for H1N1. Here I chose from U.K study.

On the left column shows the serum panel, where the serum panel comes from. And here it shows the serum panel from adults and from European, which immunized for Wisconsin. They solved for H3N2, but here we're talking about H1N1. And then the panel coming from Australia. Since Australia human serum sample comes from human population immunize H1N1, so that's valid for this study. And here it shows the viral strain or antigens that has been used in human serology study. And here is what vaccine strain is solved and here is the representative circulating viruses.

And here I show the tradition of things that shows the percentage of post-immunization, HI or eco to 40, 1 to 40. And it also shows the percentage over 40, for the race.

But here I liked to focus on the GMT reduction. Unlike the serum panel studies in using ferret study, human serum we don't have the antibody of the serum against the individual as related to viruses. So what we focus on is to see the GMT reduction.

And I'd like to spend a few minutes to explain how we summarized the overall study from different centers.

Here you can see the post-vaccination GMT titer to the vaccine strain. Absolute number is not meaningful, but the comparison of the GMT vaccine strain to the isolated virus is what you want to focus on. Here you can see that the GMT titer to the vaccine strain is 1 to 90, 1:90, where the GMT from the newly isolated virus is a 36. What we want to see is whether this reduction is a 50 percent reduction. Here you can see that's over a 50 percent reduction. That's one layer of information I'd like to you focus on.

The second one is the different panels. This only shows the one panel of this particular strain of 50 percent of reduction. And in the next serum panel, it's the same thing. It's the GMT titer to vaccine strain and it compares with the newly isolated virus. And here, again, it's a 50 percent reduction. Now, we can see that there are two panels and two out of two panels have a 50 percent reduction. And remember that because of that, the way you translated to the summary data I will present later on.

And here is another two panels which I show, which are shown in this slide. So here is a panel from U.S. and Japan. Again, you see that this GMT to the vaccine strain itself is 273, where to this particular strain, England is 40. So, it's a, one time, one panel, a 50 percent reduction. By the same token, the same thing happened to the Japan group. The GMT is a 59, where the GMT to this particular strain is a 22. The reason I said even in the beginning of my talk, we're not to see response to the particular vaccine strain, rather we see the difference. Since the serum sample from Japan was not preselected to choose the high positive response of the post-serum sample, so they are, the GMT is relatively low compared to the other centers.

So now, have this one in your mind that I put the five different centers together to see the overall picture because individual lapse of data, you know, may not be well represented of overall data. And here is a summary of the GMT, 50 percent of GMT reduction by composed the serological study from different centers.

As I mentioned in a few slides back, four of four panels for this particular strain have a 50 percent reduction. So that indicates that there is a strain which is different from the vaccine strain. The same thing for A/Fukushima. Here is like 12, in 12 serum panel, some of them have a reduction. So now you can see, I can give you the sense of overall, of the strain and composed from different centers, different centers of studies.

And here I have to point out that the Solomon Islands/3/2006 in 19 serum panels, only two have the 50 percent of the reduction in this serological study. But however, overall, the representative circulating virus, the antigen that are used in serological study, shows that in 55 panels, 30 of them had 50 percent GMT reduction. That indicates those viruses antigenically is different from the vaccine strain. And the last column shows the average 50 percent of reduction. And also, that gives you the sense of the antigenic difference by using human serum.

And I apologize that the handout that you have was typed and corrected over here, this 51 percent. So that's an indicator that the 50 percent of GMT reduction isn't quite significant in this study. And it also shows that H1N1 newly isolated circulating viruses is antigenically different from vaccine strain by using human serum study.

And now we go on for the H3N2. And as I mentioned, the vaccine strain is Wisconsin-like. So the actual strain for the different vaccine components either is Wisconsin itself or Hiroshima. And again, this representative occurrence strain, which we choose according to the antigenic and the geographic differences. Here I like to emphasize again that the Canada and the Lyon strain which is asterisks, indicates there is a Canada or Nepal genetic group, as mentioned in Nancy's talk. So we want to see how this strain behaves in the human serology studies.

And again here I choose the one of the serum panels from the CDC to give you the sense of what's the antigenic difference by using human serum studies. I am not going to explain it again. Here, like folks on the GMT reduction of the newly isolated viruses, here in this column you can see that the GMT to the vaccine strains is a 101, where the GMT to the Brisbane is a 65. It's not quite a 50 percent reduction; however, the rest of the circulation viruses that we used for antigen in the human serum serology study included Canada antigen at more than 50 percent of reduction in a GMT. And something again, it's true to the other different panels, such as from U.K. and also from Japan, and from Japan and also from a U.S. serum panel.

And if we put the serum study from different centers together, now here we show the summary of the viruses. We use a 50 percent of GMT reduction in adults. And again here, we show the Brisbane in the 16 serum sample, serum panels, 7 of them had 50 percent of GMT reduction.

And another one is Sendai. And in 8 serum panels, 2 of them have the GMT, 50 percent of GMT reduction. However, the rest of the virus includes the Canada and the Nepal variance here. We can see that every one of them had a 50 percent GMT reduction.

And if we put the whole thing together, as shown in here in 73 serum panels, 55 of them had 50 percent of GMT reduction, as indicate that this virus really antigenically is different from, from the vaccine strain.

And again, as a true for the average, the average percent of a GMT reduction, 65 percent of reduction in the summary. So, that is for H3N2.

Now, we go on for the B-strain. Again, the B-strain, everyone uses the B/Malaysia, which is Victoria-like HA lineage. And the representative current strain was a two group. One was Victoria itself, and another one was the group that represent the viruses from the, represent the Yamagata lineage.

And here, again, I choose the serum panel from the CDC. And here you can see that the vaccine strain, here there are two vaccine strains, but I don't think anybody used Ohio for their vaccine strain, for licensing the vaccine strain. And here, as you can see that the Malaysia has a 69 GMT and Ohio is 126. And the rest of the strain you can see that the newly isolated viruses from Victoria like, you can see that there's been no reduction. Include the Yamagata strain, they are not 50 percent reduction. Probably that's due to adults that has been pre-immunized, immunized the previous year with Yamagata vaccines. So this is true for the rest of the serum panel including from Japan and U.S.

And here is a summary of the GMT reduction in adults. So here you can see that in 22 serum panels only four of them have 50 percent of GMT reduction. Seventeen is the average percentage of GMT reduction. So that indicates that the circulating viruses that we used for serological study are antigenically close to the vaccine strain.

And although the strain from Yamagata had lower GMT reduction, but compare with Victoria and the reduction is more than those with Victoria like strain.

So in summary of the serological study, with the sera collected after immunization with current vaccines show that:

With H1N1, the recent viruses was not well inhibited compared to the current vaccine strain.

And the same thing with the H3N2, the current viruses was not well inhibited compared to the current vaccine strain.

Where the B, for B, the recent B/Victoria lineage viruses generally well inhibited compared to the current vaccine strain.

And I'll stop here if you have any questions.

DR. COUCH: Again, just a question to be sure I understand and have the data straight. On your summary tables of GMT reductions and you say four out of four panels, that's four out of four sources of sera tested at that particular laboratory?

DR. YE: Yes.

DR. COUCH: And that's their results. So we're looking at maybe CDC or CBER data where you're looking also at the data from Japan, Australia, and Britain.

DR. YE: Right.

DR. COUCH: With that summary table?

DR. YE: I think do, as we discussed yesterday, the HISA, the variations from center-to-center from lab-to-lab, so in order to se the whole picture we have to use the serum from different centers. Here we have to use it from five different centers to get the whole picture.

DR. COUCH: But your serum went to those laboratories and they tested it. Is that correct?

DR. YE: Can you repeat your question please?

DR. COUCH: The FDA sera, the CBER sera, you have a panel of 24 sera. Those sera went to Australia, for example.

DR. YE: Right, okay.

DR. COUCH: And they tested. It went to Japan and they tested the same sera. All this is a change of sera.

DR. YE: I should've explained it in my talk. Yes, we exchanged the serum sample from center to center. Every single one of the center had the same serum panels from all the centers. So all the centers used the same serum samples. But the study may vary, such as using different red blood cells or the way they diluted for the same samples.

DR. KARRON: John?

DR. MODLIN: You didn't show us the data from the pediatric samples. I'm sure you've looked at that, but probably the numbers are small. But I wonder if you could just summarize the age range of which the kids with sera came from and maybe just give us a summary of what they showed.

DR. YE: I know you going to ask this question. Actually I have a back slide for that but I took it out.

DR. MODLIN: I'm getting to be that predictable.

DR. YE: I think although this same, okay, for this pediatric serum, since the majority of them are naive, relatively, so you will see the better picture. But in this particular study --

DR. MODLIN: Oh no, I'm sorry.

DR. YE: So to answer your question that the serological study from pediatric study more represent, for H3 for H1, is similar to adults. Where the B, you can see some difference between Yamagata and Victoria. Did I answer your question?

DR. MODLIN: Yes, in general. That's fine.

DR. KARRON: Actually, I have an H1N1 question. The A/England/555/2006, I didn't see that in the ferret sera panel or in the evolutionary diagram. But is that virus like Solomon Islands?

DR. YE: I'd like Nancy to answer this question. It should be there.

DR. COX: In your slides it should've been designated if it had that change. Let me look at that. I've got all the data here.

Could you mark it on your slide so that he could go back and look at your slides and see?

So the England did not have that change?

DR. YE: Right. England has no change in, I think it should be 144.

DR. KARRON: I guess I was just trying to understand was the England virus and H1N1 virus that is neither like the New Caledonia nor like the Solomon Islands. I guess that was just my question.

DR. COX: That wasn't a virus that we used in our serology, sorry. What was the strain designation again?

DR. KARRON: 55/2006.

DR. COUCH: What question are we trying to answer?

DR. COX: I'll get back to you. I'll find it and get back to you.

DR. KARRON: Thank you.

Other questions for Dr. Ye?

At this point, I think Dr. Vodeiko --

DR. YE: Yes.

DR. KARRON: -- is going to speak?

DR. YE: Yes, talk to Dr. Vodeiko from the FDA, our next talk.

DR. VODEIKO: Thank you again for this chance to present the next part of information. My name is Galina Vodeiko. And from the end of 2005, I am in charge of potency and reagent preparation in CBER with big help from Christian Anderson.

Let me introduce you information about availability from CBER on vital strains and for vaccine for reagents for potency.

I think I go to four.

We start from Influenza A, H1N1 viruses. The currently available strain is New Caledonia/20/99 and it's reassortant IVR-116, prepared in Australia. Many of you know the designation for many reassortant prepared in Australia started with IVR, and prepared in New York Medical School, it's usually X with a number.

With three kinds of candidates for new vaccine strain, one of them is A/Solomon/3/2006, and is available in high yield reassortant, IVR-145. The second strain is A/St. Petersburg/8/2006, and is available as a high yield reassortant, designated as X-163. And the third strain is A/Fukishima/141/2006, a high yield reassortant preparation is ongoing in Melbourne. And it expects to be available in late-February, so now.

Influenza A H3N2 currently available vaccine strains, we have two of them. A/Wisconsin/65/2005-like. One is a high yield reassortant prepared in New York Medical School, designated as X161B for A/Wisconsin/67/2005. And another, A/Hiroshima/52/2005 is Australian high-yield reassortant IVR-142.

We have one candidate for strain changes, it's A/Nepal/921/2006, and the preparation of high-yield reassortants is ongoing in New York Medical School and expected to be available in early-March.

For Influenza B, we have available current vaccine strains, B/Malaysia/2506/2004 and B/Ohio/1/2005. There is no superior candidate strains available by now.

Availability of potency reagents from CBER, H1N1, A/New Caledonia/20/99 strain, we have available reagent antigen and reagent serum. For H3N2, A/Wisconsin/67/2005, reagent antigen and reagent serum. The same reagents are available from other centers, in U.K. and Australia.

If new strain will be chosen, we expect reagents, potency reagents available by the May of this season.

For viruses B, in CBER we have reagents for both lineages. For Victoria lineage it's a current vaccine strain, B/Malaysia/2506/2004, reagent antigen and reagent antiserum. And for Yamagata lineage, we have seen from the previous year, reagents for vaccine for the previous year, B/Jiangsu/10/2006. The same reagents are available from other chosen centers.

If new strain will be chosen, as it is out of Advisory Committee work, the reagents are going to be available in May of this season.

That's it of what I wanted to say. Any questions?

(No response.)

No questions.

DR. KARRON: Thank you very much.

At this point, we'll take a break. And I think we're a bit ahead of schedule. So instead of reconvening at 10:35, we'll reconvene at 10:30, when we'll hear comments from the manufacturers.

(Whereupon, the above-entitled matter went off the record at 10:07 a.m. and went back on the record at 10:36 a.m.)

DR. KARRON: We're now going to hear from Mr. Albert Thomas, who will give us comments from the manufacturers.

MR. THOMAS: Good morning. My name is Albert Thomas. I am with Sanofi Pasteur.

I would first like to thank the Committee for the opportunity to present the comments from manufacturers at today's strain selection meeting, and would like to begin by discussing several of the critical factors that are involved with influenza vaccine supply, and how the strain selection process can impact each of those factors.

The first critical factor is the growth potential of each monovalent strain seed virus. There are many factors that can impact the total number of doses of influenza vaccine that can be produced, such as the overall capacity that is available to each manufacturer, as well as the average yield of all three monovalent strains, but most typically the number of doses of vaccine that can be produced is limited by poorest growing or least yielding monovalent string.

For example, a manufacturer may be successful in producing 40 million doses of the H1N1 monovalent strain, 40 million doses of the B strain, but if only 20 million doses of the H3N2 strain can be produced, there will only be 20 million doses of trivalent vaccine that are available.

To evaluate the potential impact of a low-yielding strain on vaccine supply, we only need to look to 2006, to last year. Due to the initial low-yield from the A Wisconsin/67/2005 strain, as part of the initial production, if it were not for the greatly improved yield of the improved reassortant, the X161B, the supply of trivalent influenza vaccine last year would've been significantly reduced. Even though a record number of doses of influenza vaccine were ultimately available last year, the late availability of the much better yielding A Wisconsin X161B reassortant did have a timing on the impact of vaccine supply.

The most critical overall factor is time. The timing for influenza vaccine manufacturing is limited at the beginning by the timing of the strain selection meeting, and is then limited at the end to distribute and administer the vaccine prior to the onset of the influenza season. Thus, the total time to develop production seeds, manufacture the monovalent components, formulate the trivalent vaccine, fill, package, release, and ultimately distribute is quite limited.

Also, please keep in mind that production seeds typically require at least four weeks from time of receipt for development and release prior to use in large scale manufacturing.

The potency of, the availability of potency test reagents is another factor that must be taken into account. The potency or hemagglutinin titer of each monovalent component lot must first be determined prior to formulation of the trivalent vaccine. And as we've heard, that's done via single radial immunodiffusion, which requires a strain specific reference antigen and antiserum. These two potency reagents must be manufactured and standardized for each new strain prior to initiation of trivalent formulation. The time to prepare and standardize the reference for reagents typically requires 8 to 12 weeks.

The final critical factor is the timing of the annual Biological License Supplement Approval. Since formulation of the trivalent vaccine typically changes from year to year, a supplement to each manufacturer's biological license must be submitted to and approved by the FDA each year. The license supplement includes changes to the vaccine product labeling, such as updating the trivalent formulation with the new strains. Large scale packaging of the influenza vaccine does not typically begin until this approval is received in the early-July time frame.

This next slide depicts a typical time line for trivalent influenza vaccine manufacturing. There is an arrow listed here signifying the time frame of today's meeting, the strain selection meeting.

The overall time line, as I mentioned previously, for influenza vaccine is based on the requirements to produce, and release, and distribute vaccine in time to support immunizations for the upcoming influenza season.

The desired time frame to begin distribution of vaccine is beginning in early-August with completion in early-November, as depicted by the yellow bar on the time line.

The past several seasons have been an exception to the typical timing in that distribution of vaccine has extended late into November, actually even into December in several years.

In late-December to early-January of the preceding year, time frame today, manufacturers typically begin production of the first monovalent strain at risk. This risk that the monovalent strain that is under production may not ultimately be selected in the upcoming formulation. Production at risk is necessary because the time to produce the monovalent component lots are limited. Again, as limited at the beginning at the time of strain selection and limited at the end by the need to be able to distribute and administer the vaccine prior to the onset of the influenza season.

Thus, at the time of the mid-February VRBPAC strain selection meeting, manufacturers are looking to begin production of the second monovalent string. Assuming the availability of an appropriate production seed, manufacturing of the second strain typically begins immediately following the strain selection meeting.

Due to the later scheduling of the VRBPAC strain selection meeting this year, several manufacturers may have already started the at-risk production of the second strain or risk over-production of the first strain that they had underway.

This time line depicted here is based on the fact that there would be one strain change, which is listed as strain 3. Prior to beginning the production of the third strain, the high growth reassortant would first need to be developed, and then manufacturing working seeds would be developed form that reassortant.

Please keep in mind that the development of a working seed for manufacturing typically requires four weeks from receipt of the reassortant at the manufacturer.

The final stage of production of the monovalent lots involves strain balancing, in which manufacturers are targeting the production of an equal number of dose equivalents of each monovalent string to support trivalent formulation. Balancing is required due to the difference in yield, per lot, of each of the monovalent strings.

There are about 30 weeks available from the beginning of the year until the early to mid-August time frame when the monovalent production would need to cease. The time that manufacturers produce at risk is typically six weeks, the timing from the beginning of the year until the typical timing of the mid-February selection meeting. So the at-risk production time is 20 percent of the overall time that is available to manufacturer the monovalent components.

If manufacturers were not able to utilize this at-risk time, this 20 percent, the overall vaccine manufacturing capacity would drop by 20 percent. So, for example, assuming an industry capable of producing 125 million doses, that would be a 25 million dose reduction from overall capacity if this at-risk time were not able to be utilized.

In parallel with the production of the monovalent component lots are the activities related to trivalent vaccine formulation, filling, packing, and release. The most critical element involved in this timing is the preparation and standardization of the potency test reagents for a new strain.

The preparation of the potency test reagents again typically requires between 8 and 12 weeks and begins once a seed is available for the new strain. Formulation of the trivalent vaccine begins following standardization of the potency test reagents, which is then filling of the vaccine into vials and syringes. Typically, a target for beginning trivalent formulations is early-June.

Following approval of the Biological License Supplement, packaging of the vaccine begins, and typically in early-July. And following final release of the vaccine, distribution of the vaccine would begin in early-August.

The greatest challenge that manufacturers have had is achieving this time line. Specifically in the past several years it's been related to when trivalent formulation can begin. The last year which manufacturers were able to begin trivalent formulation in this desired early-June time frame was 2003. Due to numbers of strain changes, availability of test reagents, or low yield, this has been delayed to the mid-June to actually early-July in some years, which has pushed out the time frame for vaccine distribution.

And please keep in mind this is a typical or desired time line. The time line again for each individual year will depend on the number of strain changes, the yield of each strain, as well as the timing for preparation of potency reagents.

The next slide provides an update of current manufacturing status. As previously mentioned, due to the limited time frame that is available for production of the monovalent components, manufacturers have chosen to begin at-risk production as soon as or at the beginning of this year.

Again, by manufacturing at-risk, prior to strain selection, manufacturers gain additional time to produce the monovalent component lots. In past years, manufacturers have chosen to produce the A H1N1 New Caledonia strain at risk. But with the greater potential of this strain changing in this year's formulation, manufacturers have had to select a different strain.

Based on the surveillance data available at the beginning of this year, the manufacturers of the inactivated influenza vaccine have chosen to produce the A Wisconsin strain at risk. MedImmune has recently begun production of the B Malaysia strain at risk.

And again, based upon the timing of this years strain selection meeting, several manufacturers may have also had to begin the production of the second strain at-risk, or risk overproduction of the first strain.

In conclusion, successful influenza manufacturing and vaccination program is based upon cooperation among all the parties involved. The consideration of both antigenic match, availability of C candidates, including high growth reassortants, as well as the potential growth of each candidate's strain is necessary to ensure influenza vaccine supply.

A tangible example of these results are the increased availability of egg-isolates and high growth reassortants, which manufacturers are able to evaluate for potential growth characteristics of strains that might be antigenically similar but do have significantly different growth characteristics in large scale manufacturing. And I believe Dr. Cox had presented some of that, some of the data listing the increased number of isolates that have been available in recent years.

Another very tangible example of this was during last year, when the initial yield of the A Wisconsin strain was perceived to be very low. And obviously that would have an impact on vaccine supply. New York Medical College very quickly developed an approved reassortant. That reassortant was reviewed and ultimately approved, and it was able to be phased into manufacturing. And a whole new set of reagents were produced in a record time to ensure the nation's supply of influenza vaccine in 2006.

So, in summary, it is necessary to consider the various factors, such as the appropriate selection of strain, based on antigenic and genetic match, as well as the availability of C-candidates and high growth reassortants in order to best ensure the supply of the influenza vaccines.

And once again, I would like to thank the Committee for the opportunity to present the comments from manufacturers at today's meeting.

DR. KARRON: Thank you, Mr. Thomas.

Questions?

DR. JACKSON: I wondered if you could comment on how production of thimerosal free or reduced product interacts with your time line that you showed us?

MR. THOMAS: Sure. This may be different for each specific manufacturer, but particularly the example of no preservative formulation for Sanofi Pasteur is essentially the same time line, however we are not adding the preservative. The biggest constraint for the no preservative formulations are that they are filled into unit dose vials and syringes. So it's primarily a filling constraint, both from the capacity point of view as well as the timing. For example, filling multi-dose, 10-dose vials, you could essentially fill the equivalent of 10 times the number of doses in a given time than you would for unit dose.

So the time frame is similar, however you are limited on how quickly you can fill and package the product because it's in a unit dose or single dose presentation.

DR. KARRON: Thank you, Mr. Thomas.

MR. THOMAS: Thank you.

DR. KARRON: Next, Dr. Pandey will present the strain selection options.

Excuse me, Dr. Pandey, I apologize.

There is, next on the agenda is an open public hearing.

Christine?

MS. WALSH: Thank you, Dr. Karron. 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.

I have not received any requests at this time.

Is there anyone in the room who would like to address the Committee?

I see no response.

Dr. Karron, I turn the meeting back over to you.

DR. KARRON: Okay, Dr. Pandey, now it's your turn.

DR. PANDEY: Thank you.

Now, I will be presenting the options for strain selection for 2007-2008 season influenza vaccine.

As the Committee has heard before, there are implications of strain selection, both in terms of vaccine efficacy and availability. If the recommendations match the strains that will likely circulate in the given season, then there will be a great benefit to the public health. However, if the recommended strain well for the manufacturers, we may not have enough vaccine available for use or there might be delays.

So, as you can see on this slide, despite two strain changes last year, the vaccine production went pretty well and the vaccines were available almost on schedule.

The supply of the vaccine, despite these two changes has, as in the previous presentation you heard, that we had a record 110 million doses available. And it shows that it definitely met or exceeded the demand.

So we must applaud the manufacturers for a job well done considering all the problems one can face when there are more than one strain changes.

And now coming to the options for the vaccine composition, for Influenza A H1N1, we can retain the current vaccine strain recommendation, which is in New Caledonia 2999 like virus.

Or the other option could be to replace the current strain with the Solomon Island/3/2006 like virus, as the WHO has recommended.

Or the option could be to replace the current vaccine strain with something else, another alternative H1 isolate.

Now, of all these three possible options there are pros and cons.

The advantage of keeping the current strain, obviously, is that manufacturers have worked with this strain for years. They have the reagents available. But then the disadvantages of keeping the current strain, obviously, we have heard that it is a poor match.

So then the option for A Solomon Islands, if we were to switch to that, the reagents are going to be available in May. The manufacturers have already gotten some experience with this vaccine, and based on what I have heard, that it is reasonably, it goes reasonably well.

And we don't have another option at this time, I guess, for if we were to change to a different strain.

Now, the option for Influenza A H3N2. Again, we have the similar options. You know, either we can retain the current strain, which is A Wisconsin/67/2005 like virus, replace with an alternative H3N2 isolate, or another option that manufacturers definitely don't like is to defer the decision to a later date, in case there is more data going to be available in helping make that decision.

So, if the recommendation again, as I said, is to retain the strain, we have the reagents available and we have the manufacturing experience.

But if we were to change the recommendation to another strain, the availability of reagents is an issue that won't be, as Galina mentioned before, it won't be available before May.

For Influenza B, either we can retain the current B/Malaysia/25/06/2004 like virus, which is of B Victoria 287 lineage, or our other option could be to replace it with an alternative virus from B/Yamagata/16/88 or B/Victoria/2/87 like lineages.

Now, if you were to retain the B/Malaysia like virus, which was in last years recommendation as well, and also has been recommended to be retained by WHO, we have the experience with this strain and the reagents are available.

But if we were to change, then currently there are no better strains available and the availability of reagents could also become an issue, and also how it would out for the manufacturers.

So, finally I come to my last slide, which is basically the question every year for the Advisory Committee is that what strains should be recommended for the antigenic composition of the 2008, 2007-2008 influenza virus vaccine based on the epidemiology and antigenic characteristics of the influenza virus strains circulating in the human population that the committee has heard, the serological responses to circulating influenza viruses of persons immunized with the current influenza virus vaccines, which was presented earlier, and also the availability of suitable vaccine candidate strains, which also the Committee has heard.

So I'll turn it over to Dr. Karron.

DR. KARRON: Thank you, Dr. Pandey. At this point I'll open it up for discussion.

But I actually want to ask one, can I ask you one specific question, which is, if you go back to your H3N2 slide and you talked about the possibilities being A/Wisconsin or something else, would you, based on the data you've heard, if there were a something else, would you think, for example, it would be a Nepal-like strain?

DR. PANDEY: Well, that definitely came up as a possible, you know, option. But I don't know at this time and I think that's for the Committee to discuss if that could be a good option.

DR. KARRON: Thank you.

So at this point I'd like to open the strain selection up to discussion, questions, and if there are additional questions for Dr. Pandey or the manufacturers.

Everything is very -- oh yes, I'm sorry.

Dr. Wharton?

DR. WHARTON: Given that we now have a number of, we've got the wonderful privilege in the United States now of having multiple influence of vaccine manufacturers producing for the U.S. market, but some of them do not exclusively produce for the U.S. market. A couple of them do have major production facilities outside this country. And presumably those production facilities are not only, will be having to deal with recommendations from other national authorities. What is the impact of that on realistically if this Committee were to make a recommendation different from the WHO recommendations on how we would get vaccine from those facilities that have to deal with potentially two different sets of recommendations?

DR. KARRON: So this is a question for the manufacturers?

MR. THOMAS: Maybe I can sort of answer and maybe some of the other manufacturers would also like to participate. But a selection of two, for example, H3N2 strains, different one say for the U.S. and possibly a different one for the WHO recommendation would be extremely difficult. Manufacturers who produce vaccine for several markets would have to produce four strains. So there are inherent inefficiencies in doing that. The overall number of doses of vaccines would be reduces, as well as the additional complications of preparing another set of reagents and having to test the different strains. So that would have a significant negative impact on vaccine supply.

And I'd also like to point out with the discussion here of the H3N2 that there is currently no other production seed currently available other than the A/Wisconsin/67. I know there was a potential there for evaluating another strain, but currently no other seed exists today.

DR. KARRON: Thank you. Yes?

DR. MCINNES: So, I think I heard that it would be very difficult but not impossible. And I'd like to probe a little bit more about what would be the feasibility of actually getting a high growth, or using a high growth reassortant for H3N2 that could be used for production. And I don't know who wants to comment on that.

DR. COX: As soon as we realized that the so called Nepal Canada group of viruses was increasing, we went back and looked and found that we had a Nepal egg isolate, which didn't grow particularly well during initial passaging. However, on subsequent passage it seemed to pick up a little bit. That virus has been distributed to a number of individuals, including to Doris Boucher at New York Medical College. And I know that she's been working on making a high growth reassortant for that particular virus. Of course we have no idea how it will grow or what its antigenic properties will be because, of course, after going through a number of passages to select for high growth in eggs, we often do see changes. So I think there are a number of unknowns. But maybe Doris would be willing to make some comments about when the high growth reassortant might possibly be available to distribute to manufacturers.

DR. BOUCHER: I can only say it's under development. So far everything is proceeding according to plan, but we don't have it as of now. We are trying, we would like to ship it off to the CDC for them to begin analysis next Wednesday, a week from today. But we don't know what will happen with the testing. And as Nancy said, we don't know how it's going to grow.

DR. KARRON: Nancy, I wonder if I could ask you to comment on some of the ferret antisera data, Nepal, and the Wisconsin strains, and the differences that you see. And some of these are the tables that we looked at, in trying to discriminate between the Nepal strain and the Wisconsin strain.

DR. COX: Yes, I think that the H3 table in the CDC package is actually fairly instructive. We were hoping to see a different pattern. That's on page 19.

What you see here is the Nepal antiserum on the right, which has been made to an egg grown virus. So that was the virus that was sent to Doris to make a high growth reassortant. As well as antiserum made to the Canada/1212, which is a cell isolate. So we tried to control for the fact that, you know, sometimes you see a little bit different results for cell and egg-grown isolates that are genetically similar, or at least in the same general and genetic group.

And what we see, if you look down the rows, if you compare the Wisconsin/67 wild type titers, as you look down the table, and then the Nepal, the titers against the Nepal and Canada/1212, you will see that you have the same viruses being low reactors for all three antisera.

So you may get a two-fold higher titer for the Nepal and Canada antiserum, but you still get the same low reactors. And so that has been the real issue.

And as I mentioned in my talk, there have been some studies done that indicate that the receptor binding pocket of the H3 viruses has evolved. And of course the shape of the receptor binding pocket can have an impact on what you see in an HI test. So antibody could still be binding to the hemagglutinin, but not inhibiting the, if the pocket is lighter, not inhibiting the ability of the virus to agglutinate to red blood cells.

So I think that what we are going to be doing in the future with the H3 viruses is to look at virus neutralization tests for a small number of viruses. It's a very labor intensive test, but I think it might really help us to discriminate exactly what is going on here.

DR. KARRON: Right, so I guess also to summarize what you're saying, we don't really see a difference in, or much of a difference, more than a two-fold difference when we look at the HI test, Wisconsin and either the Nepal or the Canada strain?

DR. COX: That's correct. So we were hoping to see that the Nepal and Canada antisera would cover the viruses that are in the same genetic group better than the Wisconsin virus does. But really there doesn't seem to be that much difference.

DR. KARRON: Right. And there's certainly, going back to some of the human data, there's certainly a difference when you look at the human sera. But I would think that the problem is really that we don't have the reverse experiment in humans. We don't know what would happen if you immunize a person with A/Nepal or A/Canada strain. We don't, we of course don't have that information because that's not what we do. We can only look at responses once you vaccine.

Yes?

DR. FARLEY: As a follow-up to that, is there a way of, because I guess if I'm reading this correctly, the Canada and Nepal cross-protect for each other that they seem to have, am I reading that correctly if you go down further? Now, I guess the question is, is there a way of knowing among these low-reactors what proportion the burden of disease that is taken up by Canada and Nepal isolates as opposed to these various others that are listed here, with mostly U.S. designations?

DR. COX: We, because we really can't discriminate between viruses that are in the Nepal group and in the Brisbane group using the antisera that we have, we've done a lot of sequencing. And so, what we're seeing, I think I mentioned, so for the 91 H3N2 viruses that have been sequenced, and those are viruses collected since October of 2006, we found that 48 percent were in the Brisbane group and 45 percent in the Nepal group.

So we're seeing an almost equal distribution of viruses in those two groups. And it doesn't seem, it, I would say that the Asian viruses are predominating in the Nepal group. But that, for example, there was a fairly, the National Influenza Center in Seoul, Korea had sent us quite a number of H3N2 viruses from an outbreak that occurred in November and December. And those viruses were distributed in the two groups. So even if you look at a particular country you can see that viruses fall into both the Brisbane and the Nepal Canada groups.

DR. KARRON: Dr. Couch?

DR. COUCH: I would, if you permit, I think it's probably worth pursuing H3 because that's the one that's bothering us. We started there and Dr. McInnes indicated that we've already heard that follow-up too. If you look at the ferret sera and the H3 strains, you know, there's no big differences there anywhere, as Dr. Cox has pointed out. If you pick all these various isolates now, you can say well, you know, it looks like there's maybe a little shift toward the right side over there of which Nepal can just be one example.

So I make my little chart each time on this one, so antigenic change I ended up saying well, probably zero here. No big antigenic changes we can hang our head on, you see, as part of the decision making. Epidemiology always comes with a question mark and wait to see what Dr. Couch tells us. And we had no major problems with H3 anywhere in the world so we don't have the benefit of viruses that dominate in an outbreak that would help us decide that one is about to move there.

And then the final one I always look at is the human sera. And the human sera results for H3 is quite bothersome because some of those strains that are out on that side there's not very much in the way of cross-reactivity. And Dr. Ye pointed out, you know, that the reduction that you deal with GMT's for the H3s, with these various laboratories, is comparable to the H1, which is a little bit of a discussion item, but less so maybe than this one, you see.

And the final statement to make at H3, around this table and every time we do this, you know, that is the most important decision we make because that still is clearly the most common epidemic virus with the most, the most serious impact against humans, with attack rates and in hospitalizations and disease.

And so I came around here bothered about the Wisconsin decision that we were told had been made by WHO, but we all accept the fact, the position that was pointed out to us by the industry representative. If we talk about doing something different, we are really tampering with vaccine supply and perhaps significantly. So I feel like we're in a bind here this year on H3. And some other people may want to comment on that as well. That one is my tough one.

DR. KARRON: I'll say that I am particularly bothered by this sort of discordance, if you will, between the ferret sera and the human sera. And I was wondering do you want to comment on that, Bob? Does that bother you too?

DR. COUCH: The differences between the two? Yes, that bothers me. I mean I think a ferret, you know, if I want to try and biologically do something with this, I say well let's look at, and you've expressed here, let's start seeing more pediatrics here. Because if we say those ferrets are helping us differentiate these, then the most comparable individual for humans, which is our primary interest not the ferrets, is going to be how those children do. And the children, what we have this year, their Wisconsin antibody didn't like the strain either. So the ferrets maybe didn't agree with the children very well here. Now, so what data do you like for your decisions? I like to see them all fall into place, but maybe that represents what I do and my strongest interests. I want to be sure those antibody responses to that antigen cover the ones that may be coming out in the future in humans. I'm a little more concerned there than I am whether the ferrets manage to pick up a difference or not. But we'd like to see both of them.

DR. KARRON: John?

DR. MODLIN: Right. Ruth, I don't think I got an answer to my question earlier, and that is the age range for the pediatric sera. I mean it's a big difference whether they are two-year-olds or whether they are nine-year-olds.

And it may be that is the entire range, but Dr. Ye, do you know what the age range for those panels?

DR. YE: Yes, the age range is 6 months to 36 months. So they are quite young.

DR. MODLIN: Indeed, as Bob said, they could be very useful, particularly if we had larger numbers. So I think the recommendation might be, you know, in future years if we could look at a larger number of pediatric panels that would help, at least with this particular conundrum.

DR. YE: Yes, I think this year, normally we send this to CDC ourselves, so probably we should send it to different centers to give, you know, more confirmative data from it. We're limited to the limits of this sera sample, so sometimes it's harder to share with other centers.

DR. KARRON: Pamela?

MS. MCINNES: We started a little bit of this discussion at the break, but I need to, I'd like to look at page 23 in the CDC. And I need to just have somebody explain this to me. So the A/Wisconsin was our vaccine strain, right? And the A/, wasn't A/Hiroshima an alternative?

DR. COX: Yes.

DR. MCINNES: So I'm confused about the data.

DR. COX: Oh, this is a different Hiroshima.

Thank you.

DR. COUCH: I'm with you. I looked at it on the chart. It's right next to Wisconsin, Hiroshima/33, so I assumed that was the same one we were talking about last year. But I had the same problem with that that Pamela is talking about.

DR. COX: It's a 2006 strain. And in the old one was a 2005, so there's something wrong with it.

DR. COUCH: Something wrong with it?

DR. COX: Yes.

DR. COUCH: I see, okay.

DR. COX: Yes.

DR. COUCH: If you look at table 21, I think Sasha has got it. Hiroshima may not be 33, but down at the bottom lists the, should've been 2005. It's a different Hiroshima.

DR. COX: It's an old, the one in the vaccine is an older strain.

DR. MCINNES: So I can be just a little less worried.

DR. COUCH: I hate to tell you how long I spent worrying about that particular one.

DR. COX: I apologize.

DR. KARRON: Is there more discussion? Are we ready to select our strains, as ready as we're going to be?

(No response.)

Okay, I'm going --

DR. MODLIN: Ruth, I'm sorry. Just maybe a little bit more discussion about the H1 strain and then the recommendation for a WHO. I mean it sounds like to me that the experience here in the U.S. is a little bit discrepant from what the rest of the world has been this past year and so it's easy to see why the WHO made their recommendation. I guess we haven't had any discussion about what this really means for us for next year. My assumption would be that it would be presumptuous to think that we were going to experience the same H1 activity as the rest of the world is over a period of time, at least that one could predict, they would be more likely to predict that. But we haven't had that discussion and I would just be interested to what other people think of that.

DR. KARRON: Nancy?

DR. COX: I think what we have to take into consideration is we've had a predominantly H1N1 year this year. So that, generally speaking, brings up the antibody levels in the population. And so, we hadn't had so much H1N1 activity before, so just in looking at what might come next, I think it's more likely to be something different next time.

DR. COUCH: Do you want to talk H1? I've got my table on H1. The ferret data here said we've got different viruses. And so I was waiting for the epidemiologic data on that one, and it's a U.S. epidemic. It's, you know, scattered around a little bit in Asia, but presumably it's different and the antibody results are erratic. Instead of a yes and a no that New Caledonia was the highest growth, and I might of gambled on keeping that one. But WHO voted to go for another one, you see, and I said well, we had a significant year and Influenza A viruses drift. And we've got pressure, as Nancy said, you know, that we have to have some drift coming up here. So H1 has got to change, if not the coming year the following year. And New Caledonia has been there a long time.

So I rationalized my way into a week support for changing H1.

DR. KARRON: While you're at it Bob, do you want to comment on the B? There's not too much probably to say there?

DR. COUCH: B's are actually easier. Maybe if somebody wants to redo it after we have our discussion this afternoon that might be different, but no, no discussion on B, no problem.

DR. KARRON: Pamela?

DR. MCINNES: So let me return to the H3N2 dilemma and also I am very troubled by the sort of lack of concordance, the comfort that we have that the ferrets are compared with the human serum. You know, we can say well maybe the HI test is not the best way for us to do this, but in effect, the reality is we do have, I think, some troubling data here. And the question on the table would be, I'm aware of what WHO Committee recommended, I'm going to throw out the idea of two H3N2s.

DR. KARRON: Nancy?

DR. COX: I think that it would be wise for us to have comments from the manufacturers about that.

MR. THOMAS: I guess some of the questions would be, initially, what would it replace, first of all? And again, we're into the same is this a, how different of a product is this from the licensing aspect? Does this require clinical trial, that whole aspect. So there are a lot of questions on defining what the product actually is and ultimately deciding is it the H1 strain that is removed and there are two H3s. I'm assuming it's still a trivalent formulation, based upon the question. So I think there are a lot of questions on definition of the product, licensing aspects.

Now, specifically for the manufacturing point of view, the concern that I mentioned before about manufacturers producing the A/Wisconsin at-risk, obviously you would alleviate that because a lot of that product is already produced, but we're still introducing another H3 strain that today a production seed does not exist. There is no yield date available for that seed, nor do we have a definitive time line of when that see would be available, which could impact overall vaccine supply as well as the timing of when reagents could be prepared and when vaccine would ultimately be available for distribution.

DR. KARRON: I was just going to say sort of back in response, I'm, what's troubling me the most is I don't, is trying to understand low responders and what those viruses are, and what it means, and how we interpret the tests. I think despite all the manufacturing caveats, if I had seen that there were really good responses to the ferret, the ferrets had very good responses and they were clearly these were very different viruses I would've said, you know, maybe we should be postponing our decision until March until we have, until we know if we can get an isolate that represents the H3N2 strains. My concern in looking at this data is I don't know that a new H3N2 strain would do better. And if you gave that to children that that would induce a better HI response, or whether there is something about these low responder viruses that's different and that we have the, you know, we need a different test to really understand this.

So that's what's, that's what I think is troubling me the most.

DR. COUCH: And I would hope that when you do those HI tests that you've got controls in there to indicate that you're not dealing with a low responder antigen. And when you run these HI, these HI batteries, see we don't have all these ferrets here in my lab.

DR. COX: We have controls.

DR. COUCH: But you're suggesting a low responder. I would think that ought not be a question of the --

DR. KARRON: But if you give a ferret A/Nepal and look for their HI titers to A/Nepal, they're relatively low. Correct? I mean there's something about this virus that induces low levels of antibody as we currently measure them in HI tests.

DR. COX: Actually, Nepal does find, there are some viruses that are, will give you a homologous titer of 80, and that really indicates to us that that's a low avid virus. But these viruses, both Nepal and the Canada/1212, when put into ferrets, elicit titers of 640 or so. So I think that we have sort of the, we really have a contradiction in the data. And I honestly have wrestled with this and have lost sleep over this data, these data.

DR. COUCH: I thought you were talking about the humans here not the ferrets here. Because the ferrets here, I've been hearing problems with them.

One more and I quit. And you can end up, the antigens don't all agree. I mean that was part of, somebody talked about the HI yesterday. But we've got three of them for H3, Santiago, Canada, and Hiroshima and they're uniform. That's, see, if you look at the H1 data there's an erratic one in there every now and then. I can bypass erratic results, but all three of them?

DR. MODLIN: Could I ask a very basic question, and I'm embarrassed I don't know the answer to this. But for a particular antigen, if a human already has a high titer of antibody and a relatively high titer and receives an inactivated antigen, how much of a boost do we expect, or does that high titer actually inhibit a boost like it does with other inactivated antigens? And if that's the case, then it seems to me that if we have a discrepancy between our animal sera and our human sera that perhaps the human data may be a little less reliable in terms of making these types of decisions. Am I way off base, Bob or Nancy?

DR. COUCH: One of the things you'd like to know to fully understand the data, which is maybe what you're driving at, is I would like the battery to each time say is that the same group of individuals? Were they vaccinated last year? And the industry might not prefer that, but I'd like to know which of those vaccines they received as well. Because if you look at some of the data, I don't mean to be picking on my friends, I worry about some of those antigens in the Japanese vaccines, if you just look at the comparisons of the battery of sera there. But we don't know that, you see.

And your point would be that if they're already high from Wisconsin and you re-vaccinate them and they're high, then you're going to have less likelihood of finding a cross-reactivity to one of those other strains.

DR. MODLIN: Exactly.

DR. COUCH: I guess I think that's probably correct but I don't know that for sure.

DR. KARRON: Zhiping?

DR. YE: Dr. Couch already answered the question.

I just want to comment on that. Because of the serum in humans, especially for adults, previous years may expose them to the same antigens or different antigens, so their responses is kind of order than the children's one or the ferrets studies.

DR. COUCH: You could follow that up with saying well if that's the case and that antigen does change, we want that new antibody and we better give him that antigen to get it.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: A question for the manufacturers. Would, if we try to put 60 micrograms of hemagglutinin in a vaccine, wouldn't that automatically equate to a 25 percent reduction in the amount of vaccine available?

DR. HETHERINGTON: Yes, I think that is the point that there is a maximum capacity for total antigen that gets produced in the U.S. And you can cut it up anyway you want to, but a trivalent vaccine you get "x" does and quadravalent vaccine you're going to get 25 percent less. And also the increased risk of delay in production because you have yet another antigen, another seed stock you need to get up and go, so timing is at risk as well. So you're really taking two hits on going to a quadravalent vaccine.

DR. KARRON: But as maybe a follow-up question, you can answer both of these, Mr. Thomas, is what percent of the vaccine produced this year did we actually use total, of the vaccine that was made available by manufacturers because we have increased our capacity significantly.

MR. THOMAS: So the first question regarding the increased formulation, the answer exactly would be an equivalent to a fourth strain, in terms of monovalent requirements, so you'd have that 25 percent decrease.

And I'm assuming the question for a vaccine that wasn't administered, was produced last year in this current season, the biggest impact there was the timing of the vaccine. The fundamental feeling is that the timing was available, the vaccine was available in a time frame that everyone desires, the September/October time frame into early November, then there wouldn't be, there would be much less vaccine unused.

A great deal of distribution of vaccine this past year, due to issue of yield with the A/Wisconsin, which then created a delay in the reagent preparation shifted a great deal of vaccine supply into the late-October, November, into December time frame. So based on what we see from immunization programs, obviously the sooner we can get the vaccine as available, the success of the immunization program will increase greatly.

DR. KARRON: But actually, just to follow-up on, I understand that timing can be critical, but do you actually know what percent of the vaccine manufactured this year was actually administered?

MR. THOMAS: I don't have any data on that.

DR. KARRON: Do you know that?

DR. COUCH: I think most of us understand and appreciate that problem that would relate to an individual decision like say a decision this year to put two H3s in there, you see, and you cut the supply by 25 percent, perhaps more, depending. If we're talking about changing and having new concepts, and we're going to talk about one this afternoon, you're talking about it evolving slowly so that the industry can adjust to that. And they've adjusted to providing, you know, I remember the time in which you were lucky if 20 million doses were made and used. Now we're talking about over 100 million, you see, so that with time and the desire to make it and sell it, the industry can adjust, but not just like that. I think that's what we're hearing in the answers.

DR. KARRON: Lisa?

DR. JACKSON: It seems like there could be costs and purchasing implications that might not be insignificant as well.

DR. COUCH: Could you say that again?

DR. JACKSON: I'm sorry, my usual clarity. It seems like -- that's even worse -- there could be costs and therefore purchasing implications if four-valent vaccine were say more expensive than a three. And you know, as you were saying, the ability to adapt quickly to that kind of change, you know, may cause additional problems with distribution and purchase.

DR. KARRON: Pamela?

DR. MCINNES: I mean I think we're cognizant of all of these factors. The risk, I mean to just say because of all these issues we're just pragmatically going to go along with something bothers me. I mean I think the data are worrying. And this is, whether we like it or not, this is a collective effort. This is not someone is just the recipient and just marches along. It's not in anybody's interest to have a vaccine that isn't, you know, the best decision we could've made with the data that are on the table. So I am going to think about all those other factors and the balancing of them, but I first am going to wrestle with what I think is the best, what I think is the best decision based on this data.

DR. KARRON: Steve?

DR. SELF: Ultimately, it seems to me the tradeoff is one of coverage versus efficacy. So adding a component may make the vaccine a little more efficacious, it will probably reduce coverage due to timing and supply. And those tradeoffs are very hard to make even when you've got modeling results in front of you. We have none of that and, you know, I listen to this and I honestly have no sense at all about how much an improvement in efficacy we could obtain, what the impact on coverage would be, and how at the end that would balance out in the population impact. I mean this is, this is a very interesting discussion, but I find it informed by very little.

DR. COUCH: We have made, and I'm straightforwardly honest with you, as you might say the wrong decision in past years, and the outbreak that succeeded with the vaccine that did not have a good match, on occasion, has really been severe. But that was an innocent error. We couldn't help it, you see, that was the only information we had at the time, you see, when the decision was made. So that concern is there. I guess my only point was you were suggesting well you lose a little bit. No, we've got the risk of losing a lot. That's our concern.

DR. KARRON: Monica and then Nancy?

DR. FARLEY: I wonder if those who, Dr. Ye and Dr. Cox perhaps, is there anything that we would gain, or you won't know for sure, but can you comment on whether a delay of the decision on H3 has much chance, if any, of clarification for us over the coming number of weeks. Is there anything that we can do with the current strains that we have, in terms of additional testing that we think might help us sort out the low responders, or will additional strains coming in, is there much chance that the volume of additional input would be there that, you know, to help inform us, given the fact that we know it's a big tradeoff, a negative tradeoff, from the manufacturing perspective?

DR. YE: I think if we wanted to have more data probably the best we can do is to conduct human serology study using the Nepal strain. But I don't think that will be reality, because we'd have to send it to the different centers to do a similar study. This precedent we are doing this to give more data.

DR. COX: I can't, can't really add very much. I was just conferring with Dr. Klimov and he thinks that we probably have several dozen H3N2s that haven't yet been analyzed that are just coming in. We know they are coming in and haven't yet been analyzed.

We are able to generate sequence data very, very quickly, which will tell us which of the two groups the viruses are falling in genetically. It takes a bit longer to generate the HI data because we have to grow the virus and do the HI test. But there would be limited additional information. There would also be limited additional information on how the high growth reassortant that Dr. Booker is producing grows. But again, in a three week period of time, the additional data would be limited.

And then there would be difficulty but a possibility of conducting microneutralization tests in the interim. So that would be, that would be difficult to do within a three week period but it could be done.

DR. KARRON: Bob?

DR. COUCH: Well, just to extend on that one because I'll be entirely straight with the Committee, I'm waffling. I don't want to say no. I'm waffling between abstain and defer. But if we defer, and that was going to be one of the questions, you asked about more strains, but if we defer and the FDA just now is looking for the reassortant, once you have the reassortant you have to know that it works well and then you've got to make the antiserum. And then you've got to distribute that. We're probably talking about really almost May or June before you can even, the industry can even begin to work with a new H3.

DR. WEIR: I think that's correct from what Dr. Ye said that it would be probably unlikely that we could generate more serology data very fast.

But just to clarify one other thing from Dr. Cox, would you not also after the high growth reassortant is made, would you not need to generate ferret antisera to that before you test it to the isolates to see how well it would cover, to really give some useful data about whether that would be a candidate or not?

DR. COX: We always test the high growth reassortants to be sure that they have similar antigenic properties to the wild type strain. However, it doesn't preclude our distributing it to see how it grows for the manufacturers.

DR. WEIR: But I thought the high growth reassortant for the Nepal was not available yet.

DR. COX: That's right. It's not available.

DR. WEIR: So what I'm saying is after it is available, then you would have to generate ferret antisera to that before you saw how well it would really cross-react?

DR. COX: That's correct. But that wouldn't preclude its being distributed.

DR. WEIR: Okay. But it would still take time to generate that additional data?

DR. COX: Two weeks. Two weeks. Two weeks to make the serum and then the test could be done almost immediately after that. And one other thing that I need to say that needs to be emphasized I think every year, we are really limited by what the epidemic does, you know, what certain viruses circulate, where they circulate, and how many of them circulate. And we're also limited, to some extent, to the timing of when they get sent to us. But the season was a very mild season generally and that is true worldwide. And it really didn't takeoff terribly early, except for some of the school outbreaks that we had in the United States.

So this is one of the situations that we often face where we would like to have a lot more data, but the majority of, the concerning data that you've seen here today was generated in the last three weeks or so. So it's, we're really racing with the virus, and it is a moving target, and it's a very difficult business.

DR. KARRON: First Steven, then Bob.

DR. SELF: Yes, so that's a perfect segway to a question about the epidemiology. So I see that within the H3, the low reacting viruses are sort of on the rise. But I don't see what the best current data is for the balance between H1 and H3. I see last year based on the plot was predominantly H1. Am I, is there current data sort of on the balance of H1 and H3 infections?

DR. COX: Are you talking about in the United States?

DR. SELF: In the U.S.

DR. COX: Tony, I think that because there are so many unsubtyped viruses that have been recently identified in the United States, it's difficult to say. But as Tony mentioned, there appears to be an increasing proportion of Influenza As that are H3s. But they were predominantly H1s this year, whereas last year that was not true.

DR. COUCH: But it is --

DR. SELF: You're right. It's on page 4. Yes, so we're arguing about, for the H3 problem, what this year may be a pretty small fraction of the total cases. Is that correct?

DR. COX: Yes. So basically this year we've had predominantly H1s. So we wouldn't expect to have predominantly H1 next year, although influenza is not predictable. And I always have to say that over, and over, and over again.

So when H1 circulates in the United States again, we might expect to see a different virus because the New Caledonia viruses have been around for so long.

With respect to the H3s, we've had relatively less disease caused by H3s, but H3 activity appears to be picking up somewhat relative to H1 activity.

Did that make any sense?

DR. SELF: Yes, it did. I'm still trying to get a handle on just the magnitude of this subset of H3 viruses, what the likely magnitude of that problem for next year.

DR. COX: That is totally unpredictable.

DR. SELF: Okay.

DR. COUCH: There is such a thing as the Harold-wave, which as been popularized by a group from Houston, suggesting that late phrase like that, that that was the proceeder for the epidemic the following year. And there are at least three or four clean examples of that, where that's been the case.

I'm willing to take us off dead center, if you want, unless there is more open discussion.

DR. KARRON: I do just want to ask a question and go back to the H3N2. So my sense, however, is when it comes to making a decision about that the only, the two options really are to retain the current strain or really to defer, because at this point we do not have a Nepal strain. I mean we don't have a high growth reassortant. So we couldn't, as a Committee, make that recommendation. We could say that we would defer our decision. I just wanted to put that out.

And with that, I think it is actually, unless there is anyone else who wants to make any comment, question, I think it's time to actually talk about the individual strains.

And I am actually, first we'll start with H1N1. The three possibilities, as outlined by Dr. Pandey, are to retain the current vaccine strain, which is A/New Caledonia, to switch to A/Solomon Islands, or to replace the current vaccine strain with an alternative strain.

Dr. McInnes, I'm going to start with you?

DR. MCINNES: I was looking also at the decision from WHO, and I sort of do take a little bit the same view as Dr. Couch in this about concurrence with it or having difference or non-concurrence with that.

And looking at the H1 data, I would support changing that strain, the vaccine strain to the A/Solomon Islands, the H1N1-like virus for this upcoming season.

DR. KARRON: Thank you. Dr. Hachey?

DR. HACHEY: I would also support replacing the current vaccine strain to the A/Solomon Islands-like virus.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I would also support changing the current to the A/Solomon Islands.

DR. KARRON: Ms. Province?

MS. PROVINCE: I too support changing the current strain to A/Solomon-like.

DR. KARRON: Dr. Jackson?

DR. JACKSON: Yes, I agree with the change as previously stated.

DR. KARRON: Dr. Word?

DR. WORD: I would agree with the changes as previously stated.

DR. KARRON: Dr. Hetherington, do you want to comment?

DR. HETHERINGTON: I agree with the comments so far.

DR. KARRON: Dr. Wharton?

DR. WHARTON: I concur with my colleagues in changing to the A/Solomon Islands.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: I concur with updating the H1N1 strain to A/Solomon Islands.

DR. KARRON: Dr. Self?

DR. SELF: I agree.

DR. KARRON: Dr. Farley?

DR. FARLEY: I agree.

DR. KARRON: Dr. Couch?

DR. COUCH: I already said I had a weak agreement, but I agree.

DR. KARRON: Okay. Dr. Modlin?

DR. MODLIN: Yes.

DR. KARRON: Okay. And I also agree with changing to the A/Solomon Islands.

Please Christine?

MS. WALSH: Just to summarize that vote, it was unanimous, 13 votes in favor of replacing the current vaccine strain with the A/Solomon Islands.

DR. KARRON: Can we actually go to the next slide?

So there are actually three options listed up here for H3N2, but I think we have concurrence among the panel that really there are only two options that we can realistically consider.

One is to retain the current strain, which is A/Wisconsin.

And the second is really to defer a decision to a later date, pending the potential availability of a Nepal-like high growth reassortant.

So this time, Dr. Modlin, we're going to start with you?

DR. MODLIN: Well, obviously we're in a box. I'm very much concerned about the fact that these new strains have appeared so recently. And obviously the question is does that predict increased activity for these new strains next year. And I think virtually everybody has acknowledged that we don't know what the predictive, the likelihood is here.

I would point out that this represents obviously a major problem for the manufacturers and this would be a second new strain if we were to defer a decision, with the possibility that there would be a second new strain. That would be the reason why we would be deferring a decision in the first place. And that creates real issues with respect to concern about supply and cost, as Dr. Jackson pointed out.

I'm also, I recognize that we have this discrepancy between this data from ferrets and data from humans that bothers me a little bit. And I suppose if I had to make a choice between the two, I would probably come down on, based on the discussion we've had and also recognizing the fact that I'm not a respiratory virologist or an expert in this area and I'm new to this sort of decision making, but it seems to make sense that maybe putting a little bit more weight on the ferret data as opposed to the human data, recognizing the pitfalls there.

So yes, I come down with a recommendation to retain the current strain based on all this, weighing all the information that we have. It seems to be, to me the better or the lesser of two evils I guess, would be a better way to state it.

DR. KARRON: Dr. Couch?

DR. COUCH: I agree. I've sort of already said my piece on this one I guess. But I had a weak support of H1. I have a very weak support, but I would vote with going with A/Wisconsin. And for two reasons, primarily one is you heard me say that I think vaccine and some antibody is better than no antibody. And that even if we miss, we'd have some benefit there and we'd have plenty of doses of vaccine. Plus, the fact that I think if I could afford the luxury, my vote would've been to defer. But I don't think defer is likely to gain us anything in this decision.

So I guess what I say is I vote yes to go ahead, but I would like to add a qualifier to that and ask CDC and whoever else is appropriate to continue to monitor this one, and maybe these new strains you're seeing, very closely. And I don't propose this as an option, but to just at least point out that in the past when this has happened, and these new strains have appeared, we have made supplemental vaccines, the last one being A/Taiwan when we missed on the H1 decision and then we added an A/Taiwan supplemental vaccine that was given to us as a special supplemental vaccine. So I would not propose that now, but I would hold that out as an option in case we miss on this one and we still have some time.

So I do have concern about the H3 decision, but I'll vote with going with WHO recommendation. And somebody tell them we're not happy with what they did.

DR. KARRON: Thank you. Dr. Cox, would you like to offer an opinion?

DR. COX: I'd rather abstain. Thank you.

DR. KARRON: Okay. Dr. Farley?

DR. FARLEY: Well, I'm reluctant in my answer as well. I'm particularly concerned that deferring this year would be more problematic than it always is to defer. And that is that the manufacturers have almost uniformly chosen to do their at-risk production of this particular antigen, and so that not only be potentially be asking for a two component change, but we would have lost the two months of production that have already gone into it. So given all of that, but in light still of the concerns, I agree with whatever we can learn about these, this emerging issue, both from a testing standpoint of how best to look at these things when there are questions, are there additional tests that we can do and is there way that they can be done in a timely fashion given all the constraints of how it happens, how the epidemic unfolds, which we can't control. But continuing to study them so that we will understand where it is going is very much, I think, is something I would concur with.

And, in addition, then my vote would be in favor of keeping the Wisconsin component.

DR. KARRON: Dr. Self?

DR. SELF: I vote to retain the current strain.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: Well, I was always taught by my mentor, Gordon Meiklejohn, to pay more attention to human data than to ferret data. And in this case the ferret data looked reassuring to retain A/Wisconsin but the human data did not, and I am puzzled by this.

If there were a likely, a good likelihood that we could come up with additional useful information not at a later date but by some date certain, call it mid-March, call it the end of March, that would favor an updating to A/Nepal. That likelihood does not seem to me to be great. And yet at the same time, I think while I'm sympathetic with Nancy's reluctance to make any predictions for next year, looking at the mortality data for the last several years makes me concerned that next year is going to be a pretty significant H3N2 year. And the question is what virus will predominate.

Given, however, that the likelihood of additional information is not good, I would vote to retain A/Wisconsin/67/2005. If something dramatic happened in the month ahead, I hope we could reconvene on sort of an emergent basis, but I don't think the odds favor that at all.

DR. KARRON: Thank you. Dr. Wharton?

DR. WHARTON: I would concur with retaining the A/Wisconsin, but have to say that I really feel like between the at-risk production the manufacturers have already done and the WHO recommendation, which presumably will be affecting the U.S. suppliers who are located in Europe, we simply have no choice. I think those two things together would provide such a hit to supply that whatever benefits might accrue from a better match were we to wait, and all these other things that might happen do happen, that we simply would be in a very unacceptable situation regarding the influenza supply.

That is a really place to be. And I don't know what kind of signals the manufacturers look for when they make these decisions about at-risk production. I am sure good efforts are made to have those be the most informed decisions possible, and I do not know if there is any signal that could have been detected regarding these potential issues of the H3 strain at the time those decisions were made, but I hope there can be some consideration of making sure that the at-risk decisions are the best ones possible, because at this point I feel like we don't have any choice.

DR. KARRON: Thank you. Dr. Hetherington, would you care to offer an opinion?

DR. HETHERINGTON: Well, just briefly. The Committee obviously is faced with a very difficult decision, but I think it's all about coverage and delivering the vaccine in a timely manner to get what positive benefits we know will exist out of this, as opposed to putting more, excuse me, putting more at risk because of the timing and trying to gain an additional benefit that really is not quantifiable with the data we've got, unfortunately. And hopefully that situation will improve, but you're stuck with the pragmatic.

DR. KARRON: Dr. Word?

DR. WORD: I guess as I sit here I listen to many of my colleagues who have more experience in influenza and, you know, we keep hearing about issues with manufacturing, you know, they've started things up. Then I keep saying why are we here. Because if you are presented with the information and you're here to make a decision, I know that the WHO has made theirs, but then if because of, you know, various constraints from other areas, we're not going to be able to make the best decision that we think is best for this particular country, then I'm saying I'm not sure why I sat here and listened to all this. I mean so many people here felt uncomfortable with moving forward, yet they're saying I can't get this information quickly enough, and I guess with that I'm not as comfortable moving on with the A/Wisconsin. Even though I'm struggling and I'm still trying to figure out the best way to phrase this because I know you're saying you may not get additional information in a timely fashion, we should get some vaccine out to people, but then I'm going to say we're going to revisit this every single year, aren't we? I mean wouldn't that be the same discussion every year if something happens?

So, I'm going to extremely reluctantly agree too. I don't know, I'm torn. I want to say no.

DR. KARRON: Are you saying retain, defer, or abstain?

DR. WORD: I don't want to abstain. I have a thought. I would defer in good faith.

DR. KARRON: Okay. Dr. Jackson?

DR. JACKSON: Well, I agree. I mean the concerns voiced are very concerning. And if we have a significant mismatch here that's obviously something we want to avoid. But I agree with Dr. Wharton that it seems we're in a box and we don't really have much of a choice. And, you know, delays in vaccine supply really impact vaccination programs, of course, but in particular, vaccination of children which is an area of increasing emphasis. And what we find where I am is if we don't have vaccine by, at the latest, early November, we really don't get children, interest wanes, and they certainly don't get two doses. So we just really are dealing with a situation which we have really limited good options. So I would vote to retain.

DR. KARRON: Ms. Province?

MS. PROVINCE: I echo Dr. Word's sentiments. It seems that, not every year, but every year we face these same kinds of questions since I've been on the Committee. We're driven, understandably, by limitations of the manufacturer, but I don't want to be driven, I don't want my decision to be completely driven by limitations of the manufacturer. Although I know what the realities are, I understand those, but I think we need to look at the processes that we are going through, examine those, and figure out sort of from year-to-year how can we get out of this box that we seem to be in more and more frequently, and maybe make better decisions and have data available at a time where we can act on it and still accommodate manufacturing schedules.

So reluctantly, I too vote to retain the current strain, but with those caveats.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: I think Dr. Word's comments, I would perhaps argue that this is somewhat unusual to have the difference between the human the ferret data. And the timing of the isolates coming in late and having a late epidemic in the U.S. is part of it, and contributes to a complication that we couldn't really predict.

And I have to say that being on this Committee is fun because I hadn't though of children being more like ferrets than humans, although I'm thinking about it immunologically I understand that. But I think that getting children immunized is important, and if they're more like ferrets then I'm reassured by that.

So I vote to retain, but I echo Dr. Couch's comments that I think it's important to monitor and to keep the option of a monovalent supplement as an option if indeed we find there's a serious mismatch.

DR. KARRON: Dr. Hachey?

Dr. HACHEY: I'm going to agree that the problem is we just don't have a good fit this year, as far as the current vaccine. But I really don't see a clearly superior strain that we have an option to pick. And any delay is associated with clearly some substantial risk in regards to production, supply, and delays. More data would be nice, but that doesn't look like it's going to happen, at least data that is substantial enough to have a high likeliness of altering the decision.

So I vote to retain the current strain.

DR. KARRON: Dr. McInnes?

DR. MCINNES: I'm not comfortable with retaining the current strain. I think there are some additional data that could come to the table. I think we would, we have a potential within the next month, 4 weeks, 3 and a half to 4 weeks to understand about this potential, this reassortant, how it's going to perform. I think CDC has indicated that they do have some additional viruses to look at and I want to acknowledge the extraordinary amount of work that they do, and that they have put on the table, and that they continue to be willing to do. And I would vote to defer.

DR. KARRON: Thank you.

I am going to vote to retain the current strain. With all of the concerns, both voiced by I think many of the people around the table, I'd actually, two points that I'd like to make. One is I'd really like to echo what Dr. Wharton said. I think that if, for example, the B strain had been made at-risk instead of the H3N2 strain, this would've been a very different discussion. And I realize the manufacturers are working with the best data they have, but I don't know how those decisions were made and I would urge them to review them carefully each year, as I imagine they do.

The second thing is I would like to have some kind of mechanism for dissemination of the additional data that will become available in the next month or so from the CDC and from other centers to members of this Committee. Not necessarily because it will have an impact on any decision-making, but because I think all of us are concerned about this decision and we would like to be able this data as they become available.

And Christine needs to summarize the vote. And then did you have a comment John? But I'll let her summarize.

MS. WALSH: To summarize the vote on the options for H3N2, there were 11 votes to retain the current strain, A/Wisconsin, and 2 votes to defer the decision to a later date.

DR. KARRON: John, did you have a comment?

DR. MODLIN: Well, just one comment. I guess we're now in a post-hoc position of second guessing the decision to go with A/Wisconsin for the first strain, as opposed to say the B-strain. And I guess the question is when that decision needed to be made did we have anymore information at that time that we would be seeing this shift in the H3 strain compared to a similar change in the B-strain? I haven't seen that we have and so I just would question. Obviously you roll the dice and was there anymore information that would've been informative when the dice were rolled a couple of months ago compared to what we have now. And I haven't see that we have it. So that was the only thing that I would raise.

DR. COUCH: I guess we're saying the same thing, but when the industry has told us repeatedly that they have to commit before we sit around a table and make a decision, I guess what I'd say is if your commitment is H3N2 be very thoughtful that is the one of greatest concern to us. And if you cannot make that one, we'd prefer it.

DR. KARRON: Norman?

DR. BAYLOR: I just wanted to make a comment. I guess Bob, the commitment has already been made. They have committed. They have started, many of them already. We do have the march, you know, as we follow-up VRBPAC, which we usually have sometime in March, and that is the time that we could review the data from CDC. We can arrange that meeting anytime in March if that data would suggest that we need to have a further discussion. So that's open, although, again, the commitment, I think the manufacturers have already started, and they can correct me if I'm wrong, how far they've gotten. But if the data are so impressive that we need to make a change, we can have that discussion at the March meeting.

DR. KARRON: Did you want to make a comment, Mr. Thomas?

MR. THOMAS: Yes, I could provide a little insight into the timing or the decisions making for at-risk production. Just bear in mind the decision to produce the A/Wisconsin strain at-risk was decided over seven weeks ago. And that was based upon the best available surveillance data at the time. And if you recall, based on some information already presented here, at that time both B/Yamagata and B/Victoria strains were co-circulating. There was still data on both. So the B-strain was uncertain. There was indications that the New Caledonia strain, which is what had been selected in years past, would be changing, based on the availability of the Solomon Islands. So there was a new egg isolate. The B-strains were co-circulating. And at that point we all realized that the H3N2 strain is the one that has the most potential changing year-to-year, but again it was the, essentially our only available production candidate. And at that time there were no additional egg isolates available, nor did the surveillance say that there was going to be a grouping of an antigenic drift that was currently identified.

But the manufacturers would completely support another method here of looking at how we would begin that decision for producing at-risk.

DR. KARRON: Thank you. Dr. Choeling?

DR. CHOELING: Kathleen Choeling for MedImmune. So I think maybe I should explain just for transparency why we made the decision to go ahead with B, having the same information. And I think maybe the timing was a little bit different when we made our decision. It could've been. But the other thing that the Committee probably may or may not know is that we make our own reassortants. So the CDC supplies us new isolates in a very timely manner. So when we got the A/Nepal H3N2 strain, we were aware that there was a possibility that that strain could change.

So I think it's, there are a number of different reasons for that difference in our deciding to go ahead with the B/Malaysia at-risk.

DR. KARRON: Thank you. And speaking of the B, that's our one decision left to make.

Christine is going to put up that slide.

And Dr. McInnes, we're going to start with you?

DR. MCINNES: I would vote to retain the current B/Malaysia/2506/2004 like virus, B/Victoria 287 lineage.

DR. KARRON: Dr. Hachey?

DR. HACHEY: Vote to retain.

DR. KARRON: Dr. Stapleton?

DR. STAPLETON: Vote to retain.

DR. KARRON: Ms. Province?

MS. PROVINCE: I also vote to retain.

DR. KARRON: Dr. Jackson?

DR. JACKSON: I also vote to retain.

DR. KARRON: Dr. Word?

DR. WORD: I vote to retain.

DR. KARRON: Dr. Hetherington, do you have any opinion?

DR. HETHERINGTON: No other comments.

DR. KARRON: Okay, thank you. Dr. Wharton?

DR. WHARTON: Vote to retain.

DR. KARRON: Dr. Eickhoff?

DR. EICKHOFF: Vote to retain.

DR. KARRON: Dr. Self?

DR. SELF: Retain.

DR. KARRON: Dr. Farley?

DR. FARLEY: Vote to retain.

DR. KARRON: Any, no, okay. Dr. Couch?

DR. COUCH: Retain.

DR. KARRON: Dr. Modlin?

DR. MODLIN: Concur.

DR. KARRON: And I also vote to retain.

Yes, Christine, please summarize.

MS. WALSH: For the option, options on Influenza B, there were 13 votes, unanimous decision to replace -- I'm sorry, retain, retain current, the B/Malaysia virus.

DR. KARRON: Thank you.

This concludes our morning session.

It is about 12:15 and I would like to propose that we reconvene at 1:15 instead of 1:30, unless does this pose any particular hardship for anyone?

Yes?

DR. WEIR: I thought we were getting an H5 update?

DR. KARRON: Oh, I apologize.

Nancy, an H5 update, of course.

DR. COX: Gosh, I thought I was going to be let off the hook.

This is not the right presentation.

Okay. This should be easier. This is just for information only, but I thought it would be very useful to update the Committee on what's been going on with the H5N1 viruses that are circulating.

I'll just give you a bit of history and recapitulate what's been going on since 1997. Currently there are two discreet lineages of H5HAs that have descended from the A/Goose/Guangdong virus. And the A/Goose/Guangdong virus is really then ancestral virus of all of the H5 viruses that we have. That is that it's the nearest to the ancestor of the `97 strains that caused the 18 human cases with 6 deaths.

In `97, I should remind you that there was evidence for direct avian to human transmission with limited, very limited, and rare human-to-human transmission documented. That has remained true since then.

Then we didn't hear very much about H5N1, although we thought that it was probably continuing to circulate in South China. And then in late 2003, there was a sort of an explosion of reports of activity in Southeast Asia. And actually, that was at the end of 2003. Earlier in 2003, there had been two human cases with one death in Hong Kong that was from a family that had traveled to Fujian Province to celebrate the Chinese New Year.

And then retrospectively, it was determined that there was a death in Beijing that was thought to be SARS at the time but it was one of those SARS-negative patients who then subsequently was tested for H5N1 and found to be H5N1 positive.

So since the end of 2003 until today, we have cases in 12 countries. The latest country to be added is Laos, and I didn't get it on this slide. I'll have it on the next slide. But Nigeria and Laos are the two latest countries, there it is, Laos and Nigeria are the two latest countries to report human cases of, and Laos reported the first human case yesterday, or WHO reported it.

So we have 275 cases, 167 deaths. It's a 60 percent case fatality ratio, and you can see where the cases are occurring. The current hotspot is really Indonesia. And then there's a lot of activity going on in birds in Africa as well. And we've heard of recurrence of H5N1 in birds in a number of a different countries, and we heard quite a bit about what was occurring in the U.K. while we were in Geneva a couple of weeks ago. And we also have heard about bird outbreaks in Bangladesh and a number of other countries that hadn't previously reported outbreaks in birds.

So, with respect to what we're doing globally, there are some basic principles and practices that the WHO undertakes when there is a newly emerging strain.

And developing of H5N1 vaccines is one component of WHO's overall strategy for pandemic preparedness. And there are four WHO Collaborating Centers, as you know, with an additional four H5 reference laboratories. And we share the H5N1 antigenic and genetic data very frequently. It's actually put into a share compartment, which allows us to really compare what is going on. And then WHO convenes periodic teleconferences of these reference labs to discuss the data and apportion tasks require for vaccine candidate reference virus production.

And we really have to have integration of antigenic, genetic, and epidemiologic data from both the human and veterinary health sectors in order to make the best decisions about which viruses to select as potential vaccine strains.

And consequently, the candidate reference viruses are really chosen on the basis of all of these considerations.

So, as I've mentioned the HA sequences divide into two distinct phylogenetic clades. Clade 1 viruses have circulated in Cambodia, Thailand, and Vietnam and caused human infections during 2000 and through 2005. And then subsequently caused two cases in Thailand in 2006.

In contrast, clade 2 viruses were circulating in birds in China and Indonesia during 2003 and 2004, and then spread very dramatically westward after the very well known outbreak of H5N1 at Qinghai Lake in Western China. And it is postulated that migratory birds did assist in the spread of the virus to the Middle East, Europe, and Africa.

Clade 2 viruses have caused the majority of human infections since late 2005. And there are multiple subgroups, genetic subgroups, in the so called clade 2. And they can be distinguished both genetically and antigenically, and some of them have very discreet geographical distribution.

So the majority of the H5N1 virus detected in Africa, Asia, and Europe in birds, which have been associated with sporadic human infections are in clade 2.

Clade 2.1 viruses circulated in poultry and caused human infections in Indonesia. And as I mentioned, Indonesia is somewhat of a hotspot.

Clade 2.2 viruses have caused outbreaks in birds in Africa, Asia, and Europe. And these are the Qinghai Lake viruses. And these were most recently associated with human infections in Egypt, Nigeria, and we're not sure yet about the virus from Laos.

Viruses in clade 2.3 cause poultry outbreaks and human cases in China.

And then there are viruses outside this classification, the 2.1, 2.2, and 2.3, which have been isolated from domestic poultry in Asia. And there are two emerging clades, which are represented by A/Goose/Guiyang/06 and A/Chicken/Shanxi/2006. And the virus A/Chicken/Shanxi/2006 is particular interest because it was mentioned at a meeting in Beijing in early December that birds that were well vaccinated with the current inactivated vaccine using the A/Goose/Guiyang strain were having breakthroughs caused by this particular virus.

So this is the way we, you can look at my next, unfortunately will not be quite oriented like this, but these are the clade 1 viruses. And we had the vaccine candidates, Vietnam/11/94 and 12/03 about which you heard a lot about yesterday that were developed using reverse genetics to take the multi-basic cleavage site out of the HA, and then they were put on a puree backbone, safety tested extensively, and then subsequently used to manufacture the vaccines that you've heard about yesterday.

Clade 2, subclade 1, so 2.1, consists of these Indonesian viruses which have been fairly homogeneous antigenically with the exception of the viruses that were isolated from the Karo cluster. This was the large family cluster that occurred in Northern Sumatra. And those seem not to be the predominant viruses circulating in Indonesia.

And then we have Clade 2.2. I mentioned these were viruses that actually descended from the bar headed goose Qinghai Lake virus. We have a number of vaccine candidates that have been made by reverse genetics that are in this group.

And then clade 2.3 here, which is circulating primarily in China, and the majority of the Chinese human cases fall into this group. And we have to vaccine candidates here, the Anhui/1/05, which was isolated from a human infection, and then the Japanese white-eye/Hong Kong/06 was obviously isolated from a bird.

When we look at these virus antigenically, there is also good differentiation. So we have the clade 1 viruses, which inhibit each other well but don't inhibit the clade 2 viruses, so the antisera don't to the Vietnam/11/94 virus really don't do very well in inhibiting viruses in clade 2.

These are viruses in clade 2.1. The antiserum to these viruses don't inhibit clade 1 viruses very well and tend not to inhibit viruses in clades 2.2 and 2.3, although they do better with 2.3 viruses.

Viruses in clade 2.2, likewise, in antisera inhibit to these viruses and inhibit each other pretty well but there is differentiation, good differentiation.

And the same is true for viruses in clade 2.3.

So, we can see very clearly here with color coding, clade 1, clade 2.1 in green, 2.2 in yellow, and 2.3 in blue.

If we look at the profiles of these viruses in terms of their resistence and susceptibility to anti-virals, with respect to resistence to amantadine and rymantadine, the M2 channel blockers, we see that clade 1 viruses are resistant and there's a particular amino acid change in the M2 protein that confers resistence.

Clade 2.1 viruses are a mixed bag. About 80 percent are resistant and there are two different resistence changes that are seen among those viruses.

Clade 2.2 viruses have been sensitive.

And Clade 2.3 viruses also have been sensitive.

If we look at susceptibility to the neuraminidase inhibitors, we see that clade 1 viruses are generally sensitive but there have been several resistant mutants isolated from treated patients.

Clade 2.1 are sensitive.

2.2, again, generally sensitive but moderately resistant viruses were detected from patients from Egypt that were treated with Oseltamivir.

Clade 2.3 viruses are sensitive.

I won't bother going through that particular HI table because I think you've seen enough HI tables.

I just wanted to point out the two new subgroups here that I had mentioned before, Shanxi virus, and you can see the horizontal distance is the distance that really counts on these trees. And the Shanxi virus is out here from the backbone, the consensus. And then also we have another group that appears to be emerging. There are a number of 2006 viruses in this group, and these are isolated mainly from Guyang. So these viruses are being sought from our Chinese colleagues to be used, possibly used to make candidate vaccine viruses. And we really haven't characterized these in terms of their antigenicity or their susceptibility to the anti-viral drugs.

So, in conclusion, I want to make it very clear, and I think everyone in this room certainly knows that H5N1 viruses remain a pandemic threat but have not yet developed the ability to be transmitted efficiently from person-to-person.

We've seen some changes in the viruses. In particular, occasional viruses around the receptor binding pocket, but those viruses have not persisted.

We're not able to predict which, if any, H5N1 antigenic or genetic variants might acquire the ability to be transmitted efficiently. We see distinct geographical distribution of the H5N1 genetic and antigenic variants, and therefore we really find ourselves as a group, global WHO group, unable to make specific recommendations for use of one particular group or subgroup of viruses because it's not possible to predict which of the viruses in the distinct antigenic or genetic groups might acquire the ability to become officially transmissible.

Instead, we are taking the approach that we should provide potential vaccine viruses and that we should encourage the regulatory authorities to produce the reagents that would be needed to make vaccines, both for experimental purposes and for stock piling purposes.

So, as you know, the Vietnam strains have been available for some time. The Indonesia/5/05 clade 2.1 is available from CDC. Antigen and serum should be available soon from CBER, and Dr. Ye may be able to comment on that.

The reverse genetics, modified Turkey/Turkey/1/2005 clade 2.2 is available from the NIBSC in London. And they also have the antigen and chief serum available.

The Qinghai Lake, clade 2.2 is available from St. Jude. Reagents are not yet in production. And that's true also for the Whooper Swan/Mongolia.

The A/Anhui/1/05, clade 2.3, virus is available from the CDC. This virus was made during the visit of post-stock from the National Influenza Center in China to CDC and was made together with our staff. Reagents are not yet in production.

I'd like to acknowledge all of the many, many collaborators around the world, without whom I could not have made this presentation.

Thanks very much.

DR. KARRON: Thank you, Nancy.

Questions or comments at all?

Now, it's lunch time. We will reconvene at 1:30.

(Whereupon, the above-entitled matter went off the record at 12:30 p.m. and went back on the record at 1:35 p.m.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

(1:38 p.m.)

DR. KARRON: If everyone would please take their seats. We're going to begin the afternoon session.

We're going to go ahead and start with an introduction by Jerry Weir with the FDA.

DR. WEIR: Is it on? Thank you.

I'm Jerry Weir, the Director of the Division of Viral Products and I am going to give a very brief introduction to the last of our four sessions for this Vaccines and Related Biological Products Advisory Committee.

The topic of the fourth session will be a discussion of the circulating Influenza B strains that you have already heard about today and yesterday somewhat.

As you know, Influenza B viruses are not divided into subtypes, and currently vaccines contain a single B component. There are, however, two distinct antigenic and genetic lineages of Influenza B, which co-circulate. They are designated by their reference strains that correspond to them, and they're referred to as the B/Yamagata lineage and the B/Victoria lineage.

Influenza viruses from each of these lineage periodically become dominant as they circulate. And one of the main reasons that we're here this afternoon is because today and in previous VRBPAC discussions, the question has come up as to whether there should be some consideration given to some alternative vaccination strategy in order to address this issue of having two distinct lineages circulating.

Just to remind everyone, this slide shows the composition, the B-component composition in the vaccine for the last few years. And if you look to the far right you can see in parenthesis the lineage that each strain comes from. And you'll see that what's happened is we've actually alternated every couple of years with a strain from the Yamagata lineage and the Victoria. And of course, we just recommended that we keep the Victoria lineage for one more year.

So, the reason we're here today, the agenda is as follows.

Besides this brief introduction that I'm giving, our colleague, Dr. Robert Couch, who is on the Committee, has agreed to provide a background to this issue.

And then following Dr. Couch, Dr. Gagneten from CBER will discuss briefly the regulatory implications of any of the alternative strategies that are brought up and that will be presented by Dr. Couch, as well as yourself.

And then we have the manufacturers scheduled to give comments, to get their view points.

And then we have an open public hearing scheduled after that so that anyone else in the audience that is interested can put in their two cents.

And then what we would hope to have is a discussion among the Committee that will discuss this issue.

So what the goals of the afternoon are, are as follows:

We want to review the available data regarding the Yamagata and Victoria lineages of Influenza B. Discuss the genetic and antigenic relatedness, epidemiology, cross-protective responses to vaccines derived from each strain, as well as the morbidity and mortality associated with Influenza B.

And then have, the speakers will assess the various options to provide vaccine coverage for strains of both lineages.

And then we'll examine the regulatory and manufacturing considerations for these options.

And what we would like to see the Committee focus their discussion on after everyone has had their input is the following three items, which we can I guess put back up there at the end:

Please discuss the medical significance and concerns presented by the two circulating lineages of Influenza B, and whether such concerns can be addressed by means of an alternative vaccination strategy.

Second, please discuss the feasibility of the various options for expanded vaccine coverage of the two lineages of Influenza B. And you'll see as the speakers go through their data that these could, some of these options will include types of quadrivalent vaccines, supplemental vaccines, as well as some other options.

And then finally we would like you to discuss the possible future steps for both manufacturers and the public health agencies, keeping in mind of course the context of the global nature of influenza vaccine recommendations and production.

And so that's all I have for an introduction and I guess we can, unless someone has a question, we can move on to our background.

DR. COUCH: All right. Thank you, I'm Robert Couch, Baylor College of Medicine. I was introduced earlier yesterday but not again since then.

I have three introductory comments first. One is I'm your third choice speaker. When Jerry asked me to do this, I said look this needs to be either Nancy Cox or Roland Levendowski because year after year they've presented this data on the lineages and immune responses. And he came back to me and said well Nancy has got a big assignment for the morning, which I think everybody would agree she did have. And he thought Roland had an assignment for H5 yesterday, so I'm not sure what happened to that one. I said, all right, I think I'm probably a reasonable third choice. So you got the third choice for this one.

Now, the second is I said, he said how much time. I said give me 30 minutes. He gave us 40. So what I did was to add a little bit more, you might say, on the biology of Influenza B, epidemiology, and so forth before we go into the lineage data, which is what we'll have to wrestle with, with regard to discussions.

And one other comment I want to make is my purpose is to provide the background and the options. I won't take a stand. Maybe we'll come to that later on, but the options are for the discussions of the group afterwards. So that's what's coming to you.

And the final comment is in that regard, this has been going on for some time. And I've brought it up on occasions in the past, but I'm more comfortable saying, not calling this a problem, but calling this a concern. And it has been a concern, and it still is, and it may continue to be a concern that I think needs to be addressed. And that's what Jerry said as part of the introduction here.

All right, with kind of an introduction then, let's go into the subject. Some of this, this is repetitious obviously, but the classification of the human influenza viruses is, and we've already heard A, B, we don't talk about C. It doesn't exhibit that kind of variation that gives us concern. These are our medical problems, A and B, and as you know, subtypes. And within those subtypes, and for B we drift, and that's what the task we have every time we meet here is trying to identify that drift, prepare for it, make the vaccine decisions so it has control.

Now, our current concepts are that these are bird viruses. And they crossed that species barrier and established themselves in man and that's what we're getting as a problem out here. Whereas, our current concept is that this is a human virus that had our animal origin; we still don't know what it is. So Influenza B are human viruses that we have to live with but share a lot of characteristics with the Influenza A.

And I've chosen to take that tact and a little bit of the background information in contrasting Influenza A. First, just some comments about the virogy.

These viruses have a similar structure. They have the same component parts. They have the same replication sequence. There are some differences, but they're not major ones. But the Influenza B and A have unique nucleoprotein antigens, and that defines them as Influenza A or B, the nucleoprotein antigen, and they are different.

Now, that follows a lot of other differences that are associated with that. They both exhibit drift, as we said, but Influenza B do not exhibit shift and therefore no pandemic and no subtypes.

Antigenic drift is at a lower rate than that that has been described for Influenza A. And this is from one of Nancy Cox's manuscripts, the evolutionary rate for the HA1 gene and the protein, H1N1, H3N2, and B, and over a lot of years, from 1943-94 here, for the Influenza B, nucleotide changes per site, per year. You see the As are about the same and the Influenza B is a little less than half of that.

Amino acid changes per site, per year, up around five and down to a little over one and a half. And that clearly accounts for some of the differences in the epidemiology between the A viruses and the B viruses as we know them.

Now, a few comments about infection and disease, and in that contrast. There are some differences in substrate preference. Those that you have worked with these know that B likes tissue culture less than A. A does pretty well with eggs, those vary with time. And the SA performance, whole virus for HI for Influenza B, split products, more sensitivity, A doesn't have that problem. So there are some differences here, but they're all in the laboratory set. But those relate to differences in the infection in the laboratory.

As we know it to have a similar transmission, a similar pathogenesis, a similar infection pattern, and a similar illness pattern, and the same basis for immunity. And it's that we use for the basis for our decisions on an annual basis for the vaccines.

There are some differences in complications but there's a lot of overlap. Both viruses can produce pure viral pneumonia, and both produce, induce secondary bacterial pneumonia. Both produce otitis media or lead to otitis media in children, sinusitis in older individuals. Acute myositis is thought to be more common in Influenza B than in Influenza A, more in children than in adults, but described for both.

Acute encephalitis or encephalopathy is thought to be more common for Influenza A. The Japanese problem may be a good example, and it's very acute in association with the illness.

Reye's Syndrome, originally thought maybe to be a complication only of Influenza B, it turns out it's a complication, just a little bit more common, it appears, for Influenza B. And both produce various neurologic disorders that we'll call one of the myelitis disorders and there are quite a number of them.

Now, an anecdote on the side. And the reason I like this one, Influenza B and Parkland Hospital 7677. This has Jim Luby's mark on it. Some may know Jim. He's a great physician. He was a CDC EIS Officer and I have a lot of respect for what his beings, for what was done here and what was said. And this was their experience in that hospital: 15 confirmed cases, three secondary bacterial pneumonia, two with a severe chest disease disorder, and I'm not giving you details on these, high fever and rhabdomyolysis, two cases. Three of Reye's Syndrome, I guess it preceded the aspirin knowledge. Two with neurological syndromes, again, not giving you the details. One Steven Johnson Syndrom. One thyroid dysfunction. One pregnancy with toxemia and two deaths. All Influenza B.

And this is the conclusion out of that, "Quantitatively rare but qualitatively severe, complications and sequelae outside the respiratory system may be the most significant contributors to Influenza B, morbidity and mortality." And I don't think we have a good quantitative understanding for how true that may be, but clearly that is true, just the uncertainty is the overall significance.

Now, Influenza B is associated with mortality. Here is the data from Thompson, which most of us do. H1N1, H3N2, and B, look at H3N2, you see P&I Mortality estimate, 6,600, 28,000, all causes over 40,000. Influenza A here, H1N1. B is in between. You see 1,100, 5,200, and 8,300, less than H3N2 and greater overall than H1N1.

And this is one of the, this is a period of the epidemiology descriptions in a CDC publication of a ten year period. And it shows the circles, which they sometimes use. The size of the circle was the magnitude of the epidemic. And the viruses are the pieces of the pie. The slash marks are Influenza B. You see, here is about three quarters of them, Influenza B, a small number, and Influenza B epidemic, a sizeable one. Mostly Influenza B, but a small one. Influenza B almost half of a big epidemic. Influenza B, three quarters of a major epidemic. So that was actually four out of the ten years there, which is a little more common than most of us think of it, but a significant cause of epidemics for influenza.

These are age distributions. I know, and I think most people that Influenza B is more common in children. This is the kind of data that is the source of that generality. This is data out of Houston. Two successive epidemics, A/Victoria and the next year we had a B/Hong Kong epidemic. 1,100 isolates in this one and 670 isolates in this one. See, 0 to 6 months, 7 to 12 months, 1 to 4, 5 to 9, so on down the line, 45 to 64 and greater than 65.

If you look at Influenza A, you see that this is what gets you right away. It's all running about the same. Seven, six, seven, two, and this seven if four, versus the distribution, 100 percent is the total obviously. This one, and here is nine and five, which you see thirteen percent, eight percent, and then when this 30 percent grabs you. So if you look at five to fifteen, half of the isolates in the Influenza A, in the Influenza B outbreak are in that age group, a much higher frequency, lower and higher, and a greater distribution for the Influenza A epidemic.

That's the kind of data that leads to the association that it's more of a problem, not more of a problem but a major problem among children.

This is also, I experience in looking at, these are low-income clinics, charity clinics and private hospitals and private physicians offices. Just looking at the children, look at the low income. You see I put the distributions to the total here, 12 percent, 3 percent, 30 percent, 15 percent. 38 and 46, so it goes up here, and it's high down here. I dropped this one just because you can argue in the low income groups that we're getting an age group where it's hard to control the children, even get them to the doctor. But if you just go up to age 9 and take this group, you see, three quarters of those isolates in this group, a quarter of them under age five, whereas half of them are in the low income age group under age five.

That's been not an unusual characteristic, and it's also true to the certain extent to Influenza A. But they spread more and the rates are higher among low income groups. We can speculate as to the reason for that.

Hospitalizations for influenza virus infections, 1969-1995, and this is Simonsen's data. Only 1 H1N1 epidemic, but hospitalizations, you see, look at H3N2, 12 epidemics, estimated 85 to 220,000 in excess. Influenza B isn't the highest, didn't even get close to the lowest. So a much lower hospitalization rate.

On the other hand, if you get put in the hospital, this is the data out of DC Children's Hospital over a number of years, 131 Influenza As H3N2 isolates, 54 isolates of the children admitted to the hospital. Syndromes, upper croup, bronchiolitis, and pneumonia. And you see they overlap. They're both here. So you get croups a little higher than Influenza A, and other people have pointed that out. But if the child gets sick enough to put in the hospital, it's the same disease that you can see in an Influenza A infection.

And then go back to another age group so we don't forget the elderly. An outbreak in Influenza B nursing home, 1979. This represented an antigen change. I didn't look it up, but I presume it's probably the B/Singapore/79 change, which was one that was recognized. Nursing home, 359 cases, 129 yield a 36 percent illness rate, 5 hospitalized and one death. 93 percent of the individuals had been vaccinated, but there had been a significant change in Influenza B that caused that outbreak and then related to this occurrence in the nursing home.

So I would summarize the features of the Influenza B epidemiology as we know them today here. The major cause of an annual epidemic, about every two to four years. Infections occur in all age groups. Illness is most prominent among older children and young adults. Illness in infants and young children appears to be more common among the low income groups. Infections are prominent in the elderly in some epidemics, with excess mortality, but not in all epidemics. Overall impact is less than H3N2, but greater than H1N1. I think that's a fair statement that we would all agree with. But overall, Influenza B is a significant cause of absenteeism, clinic visits, hospitalizations, and deaths. And that's a reason Influenza B is in the vaccine and one of our considerations that we addressed this morning.

Now, if move on to the lineage subject, this is a, Nancy showed you one this morning, this is one from last year, but just to remind year that there are pretty clear cut change and differences in ferret sera. And you see here is a B/Yamagata lineage and a B/Victoria lineage. And we're coming up here, I'm not, I chose to just stay away from all these strain names. So what you're going to see is the lineage, B/Yamagata, B/Victoria, not the particular strain that represented that lineage in any given data. So I think it'll be a little bit easier for you to follow that way.

And this is a description of the sequence as it's generally known now. Now, before the 1980's, Influenza B was considered to be a single dominant strain in the epidemics, each year and in epidemic years.

In the mid-80s and early 90s, B/Victoria/87 like strains were dominant. B/Yamagata was present, however. It was recognized, some of it in retrospect recognition.

In the early 90s to 2000, B/Yamagata like strains were dominant, except for Asia, where they had both of them, B/Yamagata and B/Victoria.

2000 to the present, both lineages worldwide. That's the major reason for our addressing the topic today.

And another example, this was presented to this Committee in early January, at the January meeting in 2000, and I'm sure it was probably was Cumiac Nairomi. But they had had a major problem with B in two previous succeeding years, and this was the example of what we're seeing in Asia.

Here is the Yamagata derivative. That was their vaccine. Here was their outbreak two successive years with the B/Victoria. And they were, basically as he described it, smashed with Influenza B infection and disease.

And his summary and recommendations I thought were interesting. For the B components, should we consider B/Harbin, that's a B/Yamagata strain, B/Beijing, a B/Vic strain, or a third strain which efficiently covers the above two strains. You see if we had that, we wouldn't be having that discussion at all and our decisions would have been easy every time the subject comes up. So the question was raised but we don't have the solution to it, unfortunately.

And I like this recommendation from 1999-2000 recommendations from WHO. For Influenza B, either B/Yamagata or Victoria depending on your local conditions. They just putt and left it up to the individual countries to decide what they were going to do. They've not done that every year, but they sure did it that year.

And this is the circumstance before we started getting a little more concerned of this problem. Here is a geographic distribution, October `97 to 2000. This is one of Nancy's maps. And here we are in the Far East for Influenza B, you see, and the rest of the world blank. And then in late-2000 going over into 2001, then it started changing. Here is Asia. Here is Italy. It began to show up in parts of Europe and in North America. So then we began to see the Victoria lineage appearing. It was here around the late-80s and 90s, but reappearing in the Western Hemisphere.

And this is my table of the distribution of the lineage starting in `98-`99. Now, this comes from that stack of things on my filing cabinet behind my desk. And there are differences in the numbers here, and the reasons are, for a number like this, for example, I had the report from the full season. From a number like this, I only had a report that was given out at this meeting, for example, you see, in February. So they're smaller numbers. But when I had both, while the percentages are different, the patterns were the same. So I think we're all right with pattern thinking rather than necessarily the harder percentages.

And I chose to star those in which B was equal to or greater than 20 percent of the isolates. Well, that's a little bit arbitrary but in my mind when you get over 20 percent now, that's a significant contribution to the epidemic. So you can see the ones that are starred here.

So we go up here to 1998-1999, 100 percent Yamagata, and the vaccine that year was Yamagata. The next year, 100 percent. The following year, 100 percent. And then a B/epidemic, a B/contribution of significance and Yamagata vaccine. And then you can see then, then the B/Vic, we're talking about U.S. now, then the B/Vic appeared. And it was not significant that year, so we switched to a B/Vic vaccine, but we had a B/Yamagata dominant. And then we go with a B/Yamagata dominant the subsequent year and went back to a B/Yamagata vaccine, and now we have a significant contribution from Influenza B, and most of the isolates matched the vaccine.

B/Yamagata the following year, but now we have a Victoria predominance. And so this year it doesn't reach that 20. It's running around 17 percent, but two-thirds of Victoria, so at least yes, no. It's Victoria vaccine, but I didn't get it starred, that's it. But the Malaysia is matching at this particular point. So when it started circulating, we got two right and we missed on two.

And this slide does a good example. Nancy gave this one to us last year again. And I like to use this to say that this is a good example, I think, of the fact that these two lineages are jocking for dominance, and it comes and it goes in different parts of the years and has been coming and going at different parts of the world.

Now, what's going to happen? Is this temporary and will go away or not? Well, we don't know the precise answer to that, but this manuscript relates to that question.

Multiple Genotypes of Influenza B Virus Circulated between 1979 and 2003. This is an article out of the Memphis Group. And they completely sequenced 31 viruses. The lineages, according to them, were established somewhere around the mid-80s. It had been suggested by the CDC folks earlier that it was around 1983. There is other data centers, probably was close, sometime in the 70s, at least 1975. It may have been even earlier than that in which the lineages started separating and both started occurring. But by 1980, or shortly thereafter, it was pretty clear.

And they used as a parent a Russian strain showing the two lineages clearly are moving away from that. And calculated again evolution rates. This is, this is, this is a little bit higher than the one that had been seen by Dr. Cox, a little bit low, but see, at least suggested B/Yamagata, it may be a little bit more commonly changing than the B/Victoria and the same for the amino acids.

And you could get the impression, when you look at this data, that B/Yamagata looks like it's a little more likely, a little bit more of a dominant lineage, and this would be compatible with that but certainly not in enough data for you to make any kind of conclusions or any kind of planning on that basis.

And they agreed with other individuals looking at the same subject that the Influenza B had been undergoing a great deal of reassortment, mixing up all kinds of genes. And out of those 31 sera types, 31 viruses, they had 15 genotypes. I mean basically one out of every two was a different genotype.

Unrestricted mixing of lineage genes, they could find no paired relationship basis for them and so their conclusion had to be no survival advantage to either one of these lineages. So we don't come out of that with a feeling that one is about to move and replace the other, or any kind of our feeling or thinking we sometimes use for Influenza A.

And here are the genotypes, the 15 genotypes. Here is the Russian. Here is Victoria/87. There was an 85 virus that was identical. Here is Yamagata. That has a shared nuclear, a non-structured gene here. And there was a virus in Memphis that has all eight that appear clean. That's somewhat later. But you don't need to go, you can look at the stripes just are all mixed up there. It didn't appear to be any preference for any of these sorted out with any of the others, so they just concluded we can't, we have no idea what kind of virus is going to be showing up with the future. And that included the hemagglutinin genes, which of course is our focus.

Now, the problem with this is only 31 viruses. And this, I think, is important to be followed, in my view, as to what, is this prediction and anything that would say it's not right so that we would say these co-lineages are likely to continue.

Now, we just had a press release that NIAID had now completed the sequencing of 2000 influenza viruses, of gene sequences. And we sent 50 Influenza B strains over a ten year period to them for that purpose and I'm sure a lot of other people did as well. Somebody needs to examine that data bank now, and if that confirms the conclusions from here, then you would have to conclude these co-circulation is indefinite, in the indefinite future. And that's an important bit of information that would relate to any decisions that we'd want to make. So that would be one of my recommendations on data if we get to that point.

So here's a summary slide of the B Lineage surveillance data. Two distinct antigenic lineages of Influenza B have circulated at least since the mid-80s, probably before that.

Both lineages have circulated in Asia since emerging.

The B/Yamagata lineage predominated in North America during the 90s. The B/Victoria lineage was essentially absent but was in Asia.

Co-circulation of two lineages has been present in North America since 2001 with varying dominance.

And the available data currently suggests that co-circulation is likely to continue. And as long as we have co-circulation, we are no better off and we can't predict which it will be I think. It will be a guess.

Now, let's look at some of the antibody data that relates to the decision-making to this body. Serum HAI antibody after an inactivated influenza vaccine, and this is data presented at this Committee. Rather than present all these different strains, I did make selections. And I will say, I will claim they are only representative because you look at them, the titers and frequencies differently when you use them with different strains, but the patterns are the same.

B/Yamagata 1988-`89, adults and elderly, you see, the number, somehow this group seems to like 24. I'm wondering why not 25, but any rate they like 24.

B/Yamagata like, you see, percent rise, GMT, and percent equal to or greater than 40. So if you just look at this one, you see, 75, 178, 97 percent equal to or greater than 40. Elderly, somewhat lower, 97, 67. Somewhat to be expected.

Now, go over to the Yamagata, we follow the same thing. If you just look at the last column, it displays the problem, 100 percent, 75 percent to the elderly, 8 percent, 75 percent for the elderly.

Now, if you look at the opposite lineage not in the vaccine, you see, let's just, let's me, let's just take the equal to or greater than 40. It illustrates the data. 97 percent equal to or greater than 40, down to 59 percent. 67 percent down to 25 percent for the opposite lineage. Same here, 100 percent, 33 percent, 75 percent down to 17 percent. 88 percent, 38, 75 down to 50. See, a significant drop in the coverage for the opposite lineage.

Now, here's a B/Victoria vaccine, in which we've got 88 percent here, 75 percent, slightly lower in the elderly. You look in the reverse direction, now it's 88. I dropped to 64. 75 and I dropped to 63. So the patterns worked in both directions. And the reason for showing this one separate, the 2004 and 2005 vaccines is this one now has children stated, children sera.

Adults again, and we see the same pattern, you see, a B/Yamagata vaccine, 94 percent, 57 percent in the elderly. You go to the opposite lineage, it drops 94 to 46, 57 to 40. The same down here if you do the adults, 96 percent drops to 38 percent with the opposite lineage, both the Yamagata vaccines.

Now, let's look at the children. 264 sera, I wondered myself how they all of a sudden had that many in one year, but 264 sera. 79 percent rise, pretty good for the 5 to 8 year olds. 48, GMT, 64 percent, not bad at all, you see, we think, most of us think of these as less than a normal adult and elderly less, children are less. And that fits with that generality.

But if you look over at the opposite lineage, it's a much more of a drop, 8 percent. If you look at those, half, 6 months to 2 years of age, see we're looking at only 9 percent got equal to or greater than 40, and 0, nothing basically measurable to the opposite lineage in that particular age group.

And these children I don't know the ages of them, but you see the same general pattern here is 40 percent equal to or greater than 40 to the vaccine antigen and nothing measurable, presumably in any, among the children to the opposite lineage at all. So it really looks terrible for very young children.

Now, I'm sorry that the rest, the other half is not available for this data, which I put in and might not have if you hadn't sent us these articles. I had forgotten about Jan England and the two articles out of Seattle. But one of the points to make is that that previous one, see this one right here, 10, 9 percent, 0 percent. That is the same vaccine you're looking at here, which is now 62 percent GMT and 88 percent equal to or greater than 32. It's a different group of children, maybe a different vaccine, a different laboratory; in fact, we all ought to contract with that laboratory to test our serology maybe. But again, some of the variability that you can get in things.

Not quite as good with the B/Vic strains, but not bad for very young children. And these were very young children. They were all between say 6 and 23 months. So it's not always bad, but we don't have the opposite lineage data from that particular one.

And this last one is the summary of the work data that I had from the WHO reports, `98, Al missed a year too here. But four Yamagatas, a Vic, and a Yamagata. And if you look at equal to or greater than 40 and equal to or greater than 40 from Vic, which is the way to express your data, here is the drop across the board in the elderly, which you expect to see. And B/Vic, you see, now we're 79 percent to 39 percent. It drops to 30 percent or so in the percent of individuals what some people call protective titers. And children only in this last one, 13 percent equal to or greater than 40 to the vaccine lineage, and 40 to the opposite lineage.

So a little inconsistency, but by and large, very young children get very negligible benefit for the opposite lineage of the Influenza B that's in the vaccine.

And this is another. We heard from, Dr. Ye showed us that I've sort of gotten used to seeing some of this data now. What percent lower in those GMTs, you see, and most of us think in terms of roughly a two-fold different in GMT is getting at the area of significance. And they quoted in percentages, you see, so these would each be 50 percent or greater reduction for that opposite lineage for that particular individual. So those are the kinds of figures that we have to deal with.

Now, the summary of Influenza B antibody responses. Antibody responses among healthy adults to the vaccine like strains are generally good.

The antibody responses among the elderly to the vaccine like strains are reduced compared to young adults, but that is data we're custom to seeing.

Antibody responses among children to the vaccine like strains are reduced compared to adults, particularly in young children, again data that we're used to seeing, but maybe particularly, but particularly among infants.

Antibody responses to the lineage not in the vaccine are significantly reduced in all age groups. That's my judgment of significance, not a statistician's judgment. But I would consider it clearly clinically significant compared with the vaccine strain responses, and they appear to be minimal responses in very young children.

Now, one of the questions I ask is if we use the live vaccine in children, FluMist, does it do better at crossing that lineage. And I can only give you a little bit of information. I know Kathy Choeling is in the audience and I spoke to her a little bit earlier, and I think she can elaborate on these. But I got Arnold Monto to send me his data, this recent publication in the New England Journal of Medicine. And these are young adults, or 18 to 46, with a mean of 74, got live or inactivated, 278, 273, equal to or greater than four-fold increase with the Yamagata lineage, which was the vaccine, or the opposite lineage, equal to or greater than 1 to 32. And you can see the kind of data we just got through looking at for the inactivated vaccine, 85 percent, 30 percent for the opposite lineage, 98, 73.

Now, if you look at the live vaccine to only 12 percent had a response to the vaccine lineage, but four percent against B/Vic. And these ratios are basically the same.

If you look at these, this one is down about 25 percent. This one about 20 percent. But those are in the same pattern so that doesn't support any particular advantage for the live vaccine.

And this is the other source of data that I have, and that's the recent publication by Bob Belshe in the group on basically -- but the huge multi-center, multi-country study. And now, was the virus well-matched, no placebo in this study, you've either got live or inactivated. Here is the attack rate. Virus-positive illness was inactivated, live, and at 27.3 percent reduction. Not statistically significant, but reduced when the vaccine, when the infection virus matched the vaccine virus. Now, their definition of non-well matched is the opposite lineage or a B/Yamagata, which is significantly different. It's not defined in the manuscript what significantly different consisted of. But it probably doesn't really matter because when you look at that whatever poor cross-benefit you got from inactivated, which would be our concern, live is no better. So this is, there is no data here to support the live vaccine being better at crossing that lineage than was true for the inactivated vaccine.

So, with that as a background of information now, let's move on to consider the options. And the options were handed out to the Committee, and I added one to that list of options, which seems to me is fairly obvious, as I told you earlier, really we should've stuck it there probably to begin with.

And that is, don't change anything. So that is the first one up here. Continue the annual trivalent vaccine and do the best we can, which is what we got through doing this morning.

The advantage of that are a system is in place for each step to delivery of the vaccine. A single strain selection is made, reagent preparations, and manufacturing, and the delivery of that vaccine is, and the manufacturing and everything would be unchanged. See, once you have something in place it is tempting to leave it alone.

The disadvantages of that would be co-circulation would leave a proportion of persons unprotected. We've already talked about that. Children and elderly perhaps at the most risk for being unprotected. A major mismatch then would result from selection error, not from a novel virus emergence, which is the way we usually think, that we miss when a novel virus emerges like we had the discussion a little bit of H3 this morning. Those have been the -- in this case we might just make a bad selection. You hate to think that way. And then you end up missing and have significant disease that cause dyslexia.

Option two, and I added this one to the list, alternate annually the lineage strain in the vaccine. And the way I said that to Jerry, that's why I had a little less concern this past year or this current year than I had the previous year because we changed lineages.

And the advantages would be it would ensure protection to both of the lineages, at least to some degree, a little quantitative uncertainty there, among those who are vaccinated yearly, because you're switching yearly, your prime, you're boosting. It requires only one antigen selection, where we are right now. Reagent preparation, manufacturing would continue the same system we have at present. We'd be switching Bs every year, but we'd be close to the same system we're working with at present.

The disadvantages are we have the potential from reduced protection from a mismatch. We'd have to concede that. We have that now when we miss, when we don't guess correctly. And we'd have the potential for the reduced protection for those skipping a year. A number of individuals don't get their vaccine every year. You see, if you've got the two lineages, this one and one in the middle, you might be in trouble for that lineage. Not a half frequency, I don't think, but it clearly occurs.

Options now, here's a third option. Quadravalent vaccine with 15 micrograms of each lineage.

The advantage to that would be the expected responses to both lineages should ensue. And it should provide the expected protection for both lineages.

The disadvantages would be the possible increased reactogenicity. The total dose would be 60 micrograms. I think that would be negligible as my personal opinion, and certainly not a reason it, it could be evaluated, but not a reason for serious concern.

It requires a selection of two strains now and the reagent preparations yearly. So that compounds the problem at the decision level where we were this morning and at the reagent preparation at CDC. And the manufacturer, if he doesn't tell me that going to create a problem, I'll be terribly surprised. The lack of data on responses to a quadravalent vaccine, we would always consider that something of a deficiency though we would think it should be good. And my, this one is personal, I personally have a little problem with Influenza B then becoming the dominant vaccine component with 30 micrograms when the dominant vaccine component consideration should be H3N2. It's a little bit of the same idea that we heard this morning.

And I thought there was a third example, but whatever I had in my office I couldn't find it anywhere.

Vaccination with two strains, 1956-57, both B/Lee and B/Great Lakes were added to the vaccine. 63 and 64, both B/Great Lakes and B/Maryland. 73-74, B/Mass/71 and an additional monovalent B/Hong Kong/72 was made. I had no idea when it showed up, how well it was used, or that thing, but was added when this change antigenically occurred, which caused a major B-epidemic the subsequent year. So that supplemental vaccine was added.

The fourth option, a quadravalent vaccine with seven and a half micrograms of each lineage.

The advantages were if the response to both lineages should ensue. The overall B dosage would be unchanged, 15 micrograms. The reduction in immune responses and protection, versus 15, should be minimal.

The disadvantages are that it would reduce the usual dosage of the single most-likely antigen, which is a fix, to a great extent in our thinking. Some reduction in immune responses would occur and some responses in protection is possible. It would require the yearly selection of two strains and reagent preparation. Would this give the manufacturer problems or less so than 15. And lack of data, again, on the quadravalent vaccine.

Now, this is data to support that perhaps being okay. This is John Treanor's publication. And groups getting the full dose and half dose, and this is Influenza B, 2002, this must've been five years ago, the actual state, one of those shortage years when the half dose study was done.

The vaccine group, prior vaccinated group and no prior vaccinated increased four-fold you see. A significant increase, just 10 percent significant for a four-fold increase. And the same here, the 84 versus 73 percent. No differences in the GMT. No differences in the percent equal to or greater than 1 to 40. And if you look at this reverse accumulation of individuals, here I even forgot which one was which. But here is the full dose and the half dose among those who had previous vaccine. We do see decreasing numbers of 1 to 40, 1 to 80, 160 and so forth, and of those with no previous vaccine. So in that study, in healthy adults, there's a negligible difference between some slight reduction if you only gave a half dose if it's minor. The children, the data you'd like to have for more confidence are, however, is young children, and we don't have that data.

Now, the fifth option, both a trivalent and a quadravalent vaccine, so to reduce the magnitude of what you're proposing be made would be the reasons for this.

The advantage would be that there's a greater need for children and probably the elderly in which the quadravalent would provide that. The expected responses to both lineages should follow the quadravalent. The quadravalent should provide expected protection from both lineages.

The disadvantages to this would be that those given the trivalent may have reduced protection against the opposite lineage. And the available data suggests that all ages would benefit from that quadravalent, including those healthy adults. There were possible increased reactogenicity. You've already heard me for the quadravalent, I think, would be highly unlikely. It requires yearly selection of two strains, manufacturer problems, lack of data on the quadravalent, and again, an awful strong emphasis on Influenza B rather than H3N2, which I personally have a little problem with.

And the final option is production and delivery of a supplemental B, 15 micrograms it would be for the other lineage.

The advantage would be again the expected response you should get to both lineages should provide the expected and desired protection.

The disadvantages would be required, again, two strains to be done. The manufacturing problems for the supplemental vaccine, delivery problems, it makes Influenza B again the dominant antigen, rather than H3N2. You keep hearing that from me. But this would complicate delivery too, particularly for the unprime. Now we're talking about three doses, maybe even four injections depending on what you need from that opposite lineage. And it brings up the question of having to now really seriously consider adding spring vaccination, which has been proposed as a way to shorten what we have to do right with two doses in the fall. So there are some disadvantages associated with that one too.

Well, that was my assignment and the preparation I presume for the discussions.

DR. KARRON: Thank you very much. Next will Sara Gagneten from the FDA.

DR. GAGNETEN: Hello. I'm Sara Gagneten. I'm a Scientific Reviewer in the Office of Vaccines. And I didn't request for an extension of my presentation. I'm going to very briefly give you an overview of some of the regulatory considerations for the alternative vaccine options that Dr. Couch talked about just now.

The biology of Influenza B, Dr. Couch just went through it briefly just now, went through it very, in detail, and just wanted to mention in this slide the problems of coverage.

Starting in 2002-2003, as Dr. Couch mentioned, the vaccine, it was recommended that the vaccine contained strains from the B/Victoria lineage.

That year the majority of Influenza B viruses isolated in the U.S. were from that lineage, so the strain was retained the following year in 2003-2004. But that year, as you see, the majority of the viruses isolated in the U.S. were from the B/Yamagata lineage.

So for the following year, 2004-2005, there was a switch to the B/Yamagata lineage. And then that year things worked well and the virus was retained in 2005-2006.

But then the following year, the majority of the viruses isolated were from the B/Victoria lineage. And so this, as you can see, this slide illustrates problems of coverage that we've been talking about every two years, well, I mean, it's been happening every two years but that's coincidence.

So some of the, okay, I'm sorry. So I will talk to you briefly about the alternative vaccine formulations to expand coverage for circulating strains. And I will discuss the regulatory passageways for licensure of alternative vaccine formulations, manufacturers using this license process.

Just as a refresher I wanted to mention that each year after selection of the strains, manufacturers submit a BLA supplement to their licenses and that supplement does not contain chemical data.

Also, as a refresher, their license vaccine contain influenza viruses from two type A strains, one type B. They contain 15 micrograms from each strain for a total of 45 micrograms of hemagglutinin per adult dose.

So some of the alternative vaccination strategies were just mentioned. I'm just mentioning these three that would require some kind of licensing action.

So the option, one option would be to include, to develop a quadravalent vaccine with two type-B strains and two type-A strains at 15 micrograms HA for each one of the strains and a total of 60 micrograms HA per dose.

An alternative would be to develop a quadravalent vaccine with half the amount of HA, or 7.5 micrograms HA for each of the two B-strains and 15 micrograms hemagglutinin for each of the A-strains, for a total of 45 micrograms hemagglutinin of the monovalent Influenza B vaccine that would be administered with seasonal trivalent vaccine.

Now, for the regulatory considerations for alternative vaccine formulations. When you factor using a license process, I wanted to mention that at the clinical level the quadravalent vaccines would require clinical immune response and safety data. And the monovalent Influenza B vaccine administered with seasonal trivalent vaccine would require clinical immune response data for administration to address issues of possible immune interference.

At the manufacturing level, the applications would require data for each virus strain as it is done annually. And in addition, we would need data for steps that differ from the license process, the manufacturing steps, such as formulation that would differ.

Lastly, administratively, the type of application that would be required, there could be a clinical supplement to an existing BLA or a new BLA is under discussion in CBER. And options related to trade, a change in trade name, surveillance, considerations impact the type of application that would be required.

Lastly, we would require revision of the labeling.

I will conclude, and this is just to mention a few of the advantages and disadvantages. Dr. Couch went through it in detail, but generally, these options would represent an improved coverage against circulating influenza strains. They would also contribute toward preparedness of possible introduction of previously circulating strains, such as H2N2.

And the important disadvantage is, as you've heard from this morning, formulations that contain four influenza strains may cause manufacturing constraints that may affect the timing availability of vaccines.

So with this, the topics for discussion will come after the next topic.

DR. KARRON: Thank you. I think we'll move on now to comments from manufacturers.

Dr. Colgate.

DR. COLGATE: Good afternoon. My name is Tony Colgate. I'm from Novartis Vaccines based in Liverpool. And I was nominated by the former working group to give this presentation on behalf of the industry. And although you see the Novartis logo on the slides, it's not totally a Novartis presentation. There was input from all of the U.S. manufacturers, and indeed you'll see that the presentation is actually based on the presentation that was given by Al Thomas from Sanofi this morning. So basically I'm building on his presentation.

I really just want to set the scene initially. I personally find the influenza vaccine the most stimulating vaccine to manufacture. The main reason for that is it's a seasonal product. It's invariably a new product every year and, therefore, a new challenge. We have a liberty production period, so you get your product to market on time or you don't sell it. And it's changed by the next year, so it's lost.

And at present, all the influenza vaccines are derived from virus inflated in eggs. So basically that's all I'm going to talk about today. And the majority of FDA approved influenza vaccines are inactivated. And that's for all except for the MedImmune cold adapted live virus vaccine.

And many of the issues that I'm going to talk about here actually don't apply to that vaccine, but I'm not going to address those, but I think Kathy is here if you want to talk about them.

You saw this this morning. Basically, I put it up and you're going to see it twice more later on to emphasize that we have this fixed period really from somewhere at the beginning of the year, manufacturers start manufacturing at-risk, to the strain decision time, to August, in which we have to produce an antigen. And this again is assuming that in fact we have two, two seeds, or two strains that are known and only one working seed was to be produced.

A number of things are outside our control. One is the virus reference strain, which we have to, we have to get from WHO-approved laboratories, reassortant production, and also reagents. So basically it's not totally under our control. And really getting the vaccine out is a collaborative effort between industry, and WHO, and the WHO-labs. And in general, it works very well.

So we have to produce our three lots of antigen in this period, produce reagents, and then we have to fill, and formulate, and distribute.

Now, I put on the top here just to kind of remind you it's a seasonal product, new product every year, and limited production. So there's pressure on all the time.

Now, the growth potential of the seed virus, as Al said this morning, the quantity of monovalent influenza vaccine that can be produced is limited by the least product of the monovalent strains selected. So basically if you put an extra strain in there, we've got another constraint.

And each working see requires at least four weeks from receipt of the seed to develop prior to using in large scale manufacturing. Now, every monovalent that we produce has a minimum quality assurance requirement. We have to do virus inactivation validation qualification on each strain, process validation qualification, assay validation qualification, and we also, one of the manufacturers has to produce a purified antigen for the single radial immunodiffusion reagent.

Now, the potency test reagents are most important to us, as you've heard before. We can't formulate trivalent vaccine until we have some way of standardizing the vaccine. So they're required to determine the potency of the monovalent.

And, again, as I've said before, this is not under our control entirely. We are relying upon control agencies, CBER, in this case, to produce, standardize and supply reagents for all new strains. So basically not only will there be more pressure on industry, there will also be more pressure on CBER.

And as we've said before, production begins at-risk prior to the mid-February decision. If we don't do this, we are endanger of not producing sufficient doses. And as we already know, one strain is usually produced at-risk. I don't think I need to dwell on that. We covered that well this morning.

So I've got a look at a couple of scenarios and I don't think I've covered all the scenarios that Bob Couch suggested, but I think I've got some of them. Basically, the brief I had was to look at a vaccine containing 45 micrograms, which is basically what the trivalent vaccine contains now, and the vaccine containing 60 micrograms.

Now, for 45 micrograms, there are two options. One is 15 micrograms of the A-strain and seven and a half of each of the B-strains. I've heard that discussed many times in the past, but I was interested to see actual clinical results. I didn't know there was any clinical data on that.

And the other is, could be to put in approximately 11 micrograms of HA of each strain, but between 15 and 20 years ago the vaccine actually used to contain 10 micrograms of hemagglutinin, but that was changed in preference for 15 micrograms. And I'm not sure that anybody will want to go back on that.

And both of these formulations I thought may have clinical challenges with lower HA content per strain.

The other alternative is a vaccine containing 60 micrograms in total, 15 micrograms of each strain.

And another alternative we have is to reduce a monovalent B-strain in addition to the trivalent vaccine.

If we look at the first scenario, the good thing is that we could potentially produce the same number of doses as the trivalent manufacturer. But again, as I've said before, it's subject to the growth characteristics of the fourth strain, or the strain which is the lowest yielding. But what we would have to produce is initial B-strain seed and that would require manufacturing and testing. And as I said before, there would be the additional testing for the fourth strain regarding virus inactivation, process validation, assay validation, and production of the purified antigen for the reagents. And this of course would, as I said before, give additional work to control agencies.

Possible difficulties that certainly our policy control people perceived when I was discussing this with them, they were worried about the accuracy of the assay with two B-strains. They have not looked at that before. They were worried that there might be symbiosis during the two B-strains and that it would be difficult to accurately measure them. But this presumably would be down to specificity of reagents. And that's something that we would need to consider if we're going ahead with this.

So basically this is the same picture as before, except we divided the third, or divided the third part into a three and a four. So basically we're using the B-slot to produce two halves of the B-strain containing seven and half micrograms instead of 15.

The second scenario with a 60-microgram HA total vaccine is a different situation. And assuming that there were no changes to manufacturing capacity and timing of strain notification, we could only produce 75 percent of the doses compared with trivalent vaccines. So this has a significant impact.

And as you'll see when I put the chart up again for the third time, you'll see it may require productions of two strains at-risk because of the shorter periods. We may have finished, run out of steam on the first at-risk production before the strain decision is made. And also balancing of the four strains would be more difficult at the end.

There would be the additional work, as for the first scenario, which is the production of the new strain, additional testing, validation, qualification, more work for CBER and possible difficulties for two B-strains to be assay.

And here we are, as you can see, with a smaller slot here, we may be in a situation where we are looking for the second strain to manufacturer before the first one. There may be ways around that, but I guess that would require some pre-notification of A/B strain before the February decision.

Now, again, assuming no changes to manufacturing, and capacity, and timing of strain notification, if you try to produce an independent B-strain, this has even more adverse effect if you're trying to vaccinate everybody with a trivalent and a monovalent because it requires two vials of, one of the trivalent vaccine, one of the B- for every vaccination. And that would reduce filling capacity by 50 percent if a second strain is not identified very early.

If only subjects were to receive the monovalent B-vaccine, i.e. children, then the impact on monovalent strain manufacturing depends on the size of the population selected. You need to identify timing for vaccination of the selected populations, still a potential impact on filling capacity and therefore, a potential impact on the number of those who supply.

All I'm considering here are really basically the mechanics of doing this operation. I haven't considered any clinical requirements, regulation, or legal pathways, which I think have been covered previously.

And it may require a timing of strain recommendations for all the four strains. An earlier recommendation may be required if we're going to get the required doses and number of doses. But you suggested that maybe we could produce the fourth B-strain out of season. That would require some kind of decision in advance of which strain that should be.

In addition, as we heard yesterday, most manufacturers are now actually producing H5 antigen during the closed season. And in Europe, vaccine manufacturers are actually producing for the Southern Hemisphere in the down season. So that suggestion is a little bit limited, but it's worth discussing.

Just to mention, we, obviously, cell culture is the flu product of the future. And multiple manufacturers are working on cell culture influenza vaccine, but at the moment none is approved in the U.S. or I don't think anywhere else either. But it's getting close, I think, in Europe.

Cell culture has production attributes that may facilitate manufacturing of a tetravalent vaccine, but that has not yet been established.

So in summary, all of the scenarios that I've discussed increase the workload and complexity of a season of product. That potentially changes other year and is subject to exactly manufacturing time constrains.

The second scenario, the 15 micrograms of each strain, the 60 micrograms of HA total would reduce existing production approximately by 25 percent, assuming no changes are made to manufacturing capacity and the September release, because the release date is dependent on the date the virus seed is supplied, growth rates, and yields, as already described. But this could be overcome with a corresponding increase in production capacity, but that means planning and time.

Production with addition monovalent vaccine B-strain is also likely to impact on vaccine supply.

So, in conclusion, influenza vaccine manufacturers is complex, increased complexity with a four vaccine strain, and the balance between the supply and timing to deliver with additional strains.

Having said that, if desired by health authorities on the basis of public health need, the vaccine industry is prepared to try to resolve the many issues together with health authorities.

Thank you.

DR. KARRON: Thank you, Dr. Colgate.

Next on the agenda is the open public hearing.

Christine?

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.

I have not received any requests at this time.

Is there anyone in the room who would like to address the Committee?

(No response.)

Dr. Karron, would you read the open public hearing statement please.

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 at 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 top 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 speaking.

DR. CHOELING: My name is Kathleen Choeling and I'm an employee of MedImmune. What I wanted to do is just follow-up on, to provide a little bit of additional information on the immune response, the cross-lineage reactivity following vaccination of children with FluMist because there have been a lot of questions, I know, following some of our findings that show we have broad cross-reactivity against drift strains within a lineage.

There were some question whether that would extend across the lineage. So what we did is look at a few different pieces of pertinent information. And to summarize it before I tell you, it basically agrees with everything that Dr. Couch summarized so nicely earlier. But I wanted to just reinforce that.

The first thing we look at was the response in ferrets. And when you vaccine ferrets with FluMist containing one Influenza B lineage, there is no immune response developed in the ferrets to the opposite lineage that's not in the vaccine. If those ferrets are then challenged with either lineage of B-virus, there's complete protection against the lineage that's contained in FluMIst, but not any protection against challenge with a strain that's in the opposite Influenza B lineage. So that's what you would expect.

Then we also looked at the immune response in young children, 6 to 36 months of age, who are vaccinated either with FluMist or with an activated vaccine. And they got two doses of vaccine. And we looked at their serum antibody responses following vaccination, after their one dose or two doses, and if you look at the HAI response in these children, you can see a very nice vigorous antibody response, as measured by HAI to the B-lineage contained in the vaccine.

And also if you look at drift strains within that lineage, you see a good antibody response that is highly actually than what you see with an activated vaccine in that age group.

If you then test those same sera using a microneutralization assay, you can then see a vigorous immune response to the B-lineage contained in the vaccine, a good response to the drift strains, but absolutely no microneutralizing antibody detectable to the vaccine, to the non-vaccine lineage.

So I think those data all would agree with what Dr. Couch told us earlier.

And then finally we, Dr. Couch presented the head-to-head study that we just completed, the Belshe publication, and showed that you couldn't draw any conclusions based on the wide confidence intervals between the two vaccine strains. And as Dr. Couch mentioned, the way that study was analyzed was that the not well mapped strains consisted of a bucket of strains containing Yamagata, drift strains, and also the non-vaccine lineage of the Victoria strains.

So there is no way you can sort that out looking at those data, the immune response to the vaccine lineage or the non-vaccine lineage.

I looked also back at some previous years in which vaccine efficacy studies had been done with FluMist, when fortuitously the circulating strain was of the opposite lineage as contained in the vaccine. And again, all those studies were analyzed in the same way where the non-matched strains consisted not only of the opposite lineage, but also of drift strains within the vaccine lineage. So, it was very difficult to make any conclusion based on these data that we have any reason to think that you would achieve cross-lineage protection from FluMist anymore than you would from an inactivated vaccine in this age group.

Thank you.

DR. KARRON: Thank you. At this point if there are no other people who would like to make a comment during the open public hearing, we'll move to Committee discussion. And what I'd like to do is put up the slides that Dr. Weir had up at the beginning with some of the discussion points.

Okay. While he's actually putting those slides up, I have a question of my own. I think maybe they're for Dr. Cox or Dr. Couch, just something of interest to me, which is this has to do with drift among Influenza B strains. And I was wondering within the B/Victoria lineage or within the B/Yamagata lineage, how much drift do we see over the years as compared with A-strains?

DR. COX: Oh, okay. Antigenic drift is slower just like the genetic changes are slower in B than in A. And we don't really know the reasons why, but it's, I would say roughly half to a third of the rate, genetically, and antigenically, probably about the same.

DR. KARRON: So just I guess to sort of understand, when we have switched back and forth over the years between Yamagata lineage and Victoria lineage, with each switch are the viruses very different? You know, a Victoria that we chose lineage virus that say we chose this year as opposed to two, or three, or four years ago? I'm just trying to get a sense of that.

DR. COX: It probably, you know, I would have to go back and really look at it analytically, but if there's a significant time interval then there would be a difference. If not, then you know, if it's only, for example, we've seen in the Yamagata lineage we've seen B/Florida, which was one of our reference strains-like viruses for a number of years. And we don't see a change really from those B/Florida-like strains in the Yamagata lineage.

And like -- right, but when we hadn't had circulation of the B/Victoria strains for a period of time, then we had a big change between the previous B/Victoria strain that had been in the vaccine prior to that, resurgence of the Victoria.

DR. KARRON: Thank you. Dr. Farley?

DR. FARLEY: Well, I had a follow-up to that and a couple of other questions. But in looking at the, it was slide number 3 on our handout from Dr. Gagneten. It almost looked like, which was kind of showing the percentage of Yamagata versus Victoria in one year, and what the vaccine was, and then what happened the next year. We, for the last four or five years, we've been in the pattern of going two years with one and it's almost, if you looked at that, if you had just alternated years rather than done two years in a row for each one of them, it seems like we would've been closer to the mark. And I wondered, I mean if we went back and kind of re-analyzed it, would there have been something near the end of the previous season, the season that was a good match, the last things that were isolated, would that predict the reversal that you'd see the following year.

And I just wondered if one of the other options to discuss was just that automatically we assumed we were going to alternate years and can we come up with a system where it would give us an early option, in terms of the strain, so that could be the first thing they work on the next year, whether the off season or at the very beginning of the following season.

So that's one of my questions. And I guess the other is much more hypothetical. But it seems like with this kind of stable 2 lineages that it's a perfect candidate for molecularly constructed antigen, where you would put the key parts of both of those together. And it seems to me that, again, the idea of trying to modernize the process that this, the Bs, would be the first good candidates for trying to work on that option, that approach, and get it incorporated into our thought process. I mean I know that will take years to go through the regulatory issues, but it seems like this one would be a good one to put in the two stable antigens.

DR. EICKHOFF: Well, actually my comment sort of takes off from what Monica just said because of the several options that Bob Couch spelled out. The one that I found most intriguing was simply to alternate between the two lineages year-to-year, pretty much irregardless of what one expected to be predominant that year. And that was the one option that Mr. Colgate had no comment on, did not consider it, and so I wonder if Mr. Colgate might comment about that option in particular.

DR. COLGATE: That would cause us absolutely no problems at all.

DR. KARRON: Actually as a follow-up question to that, would there be a potential advantage in terms of, an actual advantage in terms of vaccine production? That is to say if you could make the Influenza B strain every year at risk because you would know it's either going to be Yamagata --

DR. COUCH: Every year you would know one strain you could start with in B-strains.

DR. COLGATE: If you could tell us that, we'd be very, very happy, yes.

DR. KARRON: But just because then you would never be in a situation of potentially not having, you know, if there were going to be say, as we talked about this morning, the H3N2 change --

DR. WEIR: I was going to say --

DR. KARRON: Anyway, yes?

DR. WEIR: -- and I think Nancy is too, you still have to pick the right strain.

DR. KARRON: Right.

DR. WEIR: Okay.

DR. KARRON: Right. And Nancy, yes?

DR. COX: I think that it's also important when you're thinking about, when you're thinking about the big picture you have to know not only what proportion of Yamagata and Victoria lineage strains are circulating in the U.S., but what proportion they made up of the entire influenza activity that was ongoing. And that's, you know, in some years it can be a big problem if you have a mismatch. In another year, if you have relatively little B-activity and very few outbreaks, if you have a mismatch, it really doesn't have the clinical impact.

And so I think that looking at it simply, you know, in this dichotomous way over-emphasizes the problem that we sometimes face when we have a mismatch. Sometimes it really isn't that important clinically because we have very little B-activity.

The other thing is that if the U.S. decides to do this, would decide to alternate, I just want to emphasize we would have to choose the right strain. There wouldn't be an automatic okay, go ahead with the old, whatever it was before, Victoria lineage or Yamagata strain. But also, the U.S. could be out of sync with the WHO recommendations very easily. So that would be a potential disadvantage for the manufacturers that Tony didn't bring up.

So, and what sometimes happens is that globally you'll see one picture where a certain lineage will predominate, whereas, in a particular country or particular region of the world you'll see another picture. And that was very true because Victoria viruses, I'm sure you covered this while I was out of the room, but Victoria viruses continued to circulate in China, specifically in Southern China, co-circulate with the Yamagata lineage viruses. Well, they caused very little activity, if any, elsewhere, anywhere else in the world.

DR. KARRON: Right. And I assume that it was issue like that that led to the recommendations that Bob talked about, which is use whatever is appropriate in your region.

DR. COX: Exactly.

DR. KARRON: So in those sort of years you would not be out of sync because I assume that everyone was manufacturing differently and according to the needs of the region.

I wonder actually if this kind of issue has every been discussed at WHO or ever come there as an issue, this issue of a problem with B.

DR. COX: It has not been presented in this formal way, where pros and cons of the different alternatives were really spelled out and discussed. So this is a really good way to look at the problem.

I had one question for FDA. If a tetravalent vaccine, including two B-strains and two A-strains were to be licensed, would you have to go through the same regulatory process if you were to have three A-strains, that is two H3s and 1 H1, and a B-strain. Or would the fact that you have 60 micrograms or whatever number of micrograms would be decided be, and you looked at that quantity of antigen and found it to be safe, and effective, and not to be interference. Would that be sufficient data for you to be able to generalize and say you could have two H3s, instead of two Bs if, for example, the B/Victoria lineage did circulate only in China, as it did in the past for a period of time?

DR. WEIR: Okay, I think I got it. So you're saying if we license the quadravalent with, in this case, two As and two Bs, then could we switch and As and Bs in any sort of combination in the future without --

DR. COX: Yes, yes. So that say we were facing a year like this year where the As --

DR. WEIR: So, in other words, two H3s the next year if that became?

DR. FARLEY: And just to add a little complexity, more complexity, could H5 be one of the antigens?

DR. WEIR: Okay. So we haven't thought about this that much. I think that, I'll just speak off the cuff and then let Norman correct me. I think that the simple example, if you really had four and you licensed it and it was safe and effective, probably strain changes would probably be pretty easy to manage. Now, I haven't thought about the H5 possibility.

DR. BAYLOR: And I think that adds to the complexity because an H, unless you take it even further, an H5 we're gaining experience on H5. So in the near future, H5 may not cause as many problems, but you could take it out to an H7, an H9, and there where we may not have had a lot of experience, I think it would be much more complicated.

DR. WEIR: Yes, I guess the other scenario is a little simpler because we would at least be assuming the 15 micrograms is effective as well, whereas if we don't know that with the H5. But as Sara pointed out, I mean we are talking about clinical data to support the, you know, not only the safety but the efficacy, you know, to make sure no interference from one strain to another, the addition.

DR. KARRON: Dr. Self and then we'll take a comment from the audience.

DR. SELF: I'd like to go back to this point about the antigenic variability within and between lineages. I mean my read of this data and some of the comments was that most of that variability is between lineages. And the slow rate of evolution would sort of support that idea. If that's true and variation within lineage is fairly slow over time, then the selection earlier of a strain within a lineage would be much easier. And that would have some important implications about manufacturing.

However, Nancy, you indicated and I was getting, you know, vigorous head bobbing across the way that selection of strain within lineage would still be a very important. Where, which is it? Would selection of strain within lineage be able to be done reliably sooner if lineage was set?

DR. COX: That's a very, very difficult question to answer. If we were looking at switching to the Yamagata lineage, we have strains that we've had basically, one strain we've had in our back pocket for a number of years. So given the fact that we are very actively pursuing egg isolates and characterizing them more vigorously than we have in the past, I would say that on balance, because the B-viruses do tend to evolve more slowly, it would be easier but it wouldn't be guaranteed that we could come up with that earlier selection. But it would certainly be a lot easier than trying to do that for the H3s.

DR. SELF: So then the strategy of alternating years, which I find sort of theoretically fascinating, but I'm not sure is the best solution, or maybe a quadravalent but would split those. I was impressed at how similar the data looked for the seven and a half would make that total B-component able to, manufacturers to start that process at considerably less risk earlier. Is that fair?

DR. COX: I do, I think that's fair with the caveat that those data were derived from immunization of young, healthy adults.

DR. SELF: So one of the down sides, one of the cons listed by Dr. Couch in the option with the quadravalent half dose was the lack of data --

DR. COUCH: The half dose data was healthy adults. And I'm told, this is hearsay, but there is a large study done by the military of basically the same thing, but it was all much larger numbers, all healthy adults and found the same thing.

And the data we want are the two on each end, preferably the children. But we don't have that data.

DR. SELF: But lack of data in the cons distinguishes from all the other cons in that you can remedy that. Most of the other cons you can't do that. You just have to live with that.

DR. COUCH: I want to ask just for my information and maybe help a couple of questions of Tony. If you took the scenario of I don't care seven and a half or fifteen, you can comment on the two, but two of them, see, and let's take the seven and a half I like best, you only have to make half as much but you have to make two of them, what does that do to the time frame? You know, do you have to close down and a long time to start up again and use the rest of your eggs and so forth?

DR. COLGATE: Not really because we're changing between the H1, the H3, and the B anyway, especially at the end when we're trying to balance the strains. So it's really a matter of having a reference strain early and being able to get the reagents so that we can formulate. So it is really, again, things which are, to some extent, out of our hands. As I was trying to explain before, we have to work in cooperation with everyone else. So if we have the strains early enough and the reagents are there, then it's just more hassle basically is putting four strains together in a season.

DR. COUCH: And then if you went for a full 15, could you then slide that back instead of starting in December, why, we give you one of the, I mean in January we give you one of the antigens and you start in December?

DR. COLGATE: That's a possibility. But as I said before, some, the problem is that some manufacturers in Europe are producing for the Southern Hemisphere as well. So their production, and also H5 production goes in that time. I guess the only answer really to 15 micrograms of each is to increase capacity. And I mean that can be done with sufficient notice, basically, and investment. I mean all these things can be done if it's done in a controlled, planned way.

DR. COUCH: No question about hassle, but some of them are a little easier and doable than some of the other options that we talked about.

And my last question, I'm not sure whether it's for you or for FDA is I don't know, and maybe somebody else does, where do we stand with regard to development of cell culture vaccines in the pipeline and that sort of thing? Can comments be made on that because everybody is waiting for that other option to come into the considerations for flu vaccines.

DR. COLGATE: I don't know.

MR. TSAI: I'm Ted Tsai. I'm an employee of Novartis Vaccines. Novartis has an MBCK cell culture vaccine that for which an application has been submitted to the EU and for which we have some plans for the U.S., including a manufacturing plant that's been, for which construction is already underway in Holly Springs, North Carolina. So there is a cell culture vaccine based upon MBCK cell production that is emerging very soon.

DR. COUCH: You may not be able to say it, but you've got to have some sort of time line. Is that three years from now, five years from now, or can't you say?

MR. TSAI: We can't predict what the EU regulatory authority will say.

DR. COUCH: Well, assuming they're cooperative.

MR. TSAI: Well, they have the application. And as I said, we have plans to submit an application to the U.S. And there are other manufacturers with cell culture vaccines in the work as well for the U.S.

DR. KARRON: There is someone who has been waiting very patiently in the audience.

MS. CAVANAUGH: Nancy Cavanaugh, MedImmune. I just had a clarifying comment, I guess, about the quadravalent vaccine, and in particular the regulatory and clinical pathways that were described by Dr. Gagneten, and whether those would be similar for the live attenuated vaccines. Those were specifically described for the inactivated vaccines.

DR. GAGNETEN: Those, I'm sorry. Those ones would hold for licensed products. So it would hold for MedImmune also.

MS. CAVANAUGH: Thank you. And actually before you sit down, could I just ask in terms of as we're talking about this half dose, seven and a half microgram, obviously you don't measure your doses in micrograms, but does that have relevance when you're considering FluMist or is that then just a quadravalent vaccine and it almost doesn't matter whether it's ten, 7.5, or it's half that, which is not much. I mean how would you interpret that?

DR. KARRON: Right now we're considering both options. You know, the same dose for four vaccine strains or yes, half of each.

MS. CAVANAUGH: Okay. Thank you.

DR. KARRON: Dr. McInnes?

DR. MCINNES: I'd like to push the envelope on the amount of antigen that can be manufactured as opposed to reducing the dose that we're delivering. And I say that because I don't think we have ever optimized the amount of antigen we've delivered, to optimize for immunogenicity and deficacy. And even though we did see very similar responses in the study that was for healthy young adults. So I'm a little reluctant to just assume that, you know, moving towards a half dose, half strength concentration on each of the Bs is the solution.

So in trying to push again about the amount of antigen that could be manufactured to produce, at a minimum of 60 microgram per unit delivery, however that be divided up, I'd like to ask about the life of the manufacturing facility. I've only ever visited during the day. Does it work round the clock? Is it possible to push the amount of time that can be in a 24-hour period? I mean do you work all night?

DR. COLGATE: Basically, it really depends on your manufacturing. But the limiting factor is the number of eggs that you can incubate basically. So it's the number of eggs that you can handle in a working day. And that working day, I guess, could be 24-hours.

But really, the simplest way would be just to increase the size of your facility to buy 25 percent, I guess, and operate that way, rather than try to hot-bed everything. I mean if you try, if you stress the facility too much by running it 24-hours a day, it's going to crash. And there have been examples of that.

DR. MCINNES: Talking about the same period working 24-hours a day. No, I'm serious.

DR. COLGATE: No, the facility as well. I mean you basically, you have to allow time for cleaning and preparation and make sure everything is done in an orderly way. If you try to compress too much in, eventually you fall over. It's really better to do it in an orderly way and just increase your capacity for processing eggs to the amount that is needed to go out at the end of the day.

DR. MCINNES: Tony, I'm sorry, so I'm pushing you here. So I'm hearing that in fact we work a traditional daylight time. There may be some cleaning and campaigning going, but you're not loading eggs in, in what could constitute the other half of the clock, right? There's not a night stock that works the same kind of work as the day stock?

DR. COLGATE: We have an evening staff who actually do the cleaning in preparation for the rest of the day.

DR. MCINNES: Okay.

DR. COLGATE: And certainly in the downstream processing, we are actually working a 24-hour shift.

DR. COUCH: Am I correct that basically what you're saying, Tony, is we're going to increase dosage, you know, and I'm interested in antigens besides B, you have to have increased facility capabilities? Is that what you're saying?

DR. COLGATE: That's it basically. Do it properly or you get no vaccine in the U.S. one year.

DR. KARRON: Dr. Jackson?

DR. JACKSON: Just regarding the half dose. I mean I guess, I think we do know some things. We know in the elderly the response to the 15 microgram is diminished and that there is clearly a very strong dose response, you know, in Dr. Kyle's studies and others. If we give more antigen, we get a better response. And we think more is better in that regard. And then when we go down to the other end of the age spectrum, particularly infants, you know, Dr. Englund's studies and other work that's been done indicate that the response, in particular, to B after a single dose of vaccine is very poor. And even after two doses, in some cases, doesn't seem to be great.

We also know that in a good year only about half of children who are supposed to get two doses actually get two doses. So I'd be very concerned about cutting the amount of antigen in half in those two groups for fear of what might happen assuming of what we know what a correlation there might be between antigen, antibody level and true protection.

DR. KARRON: Dr. Farley?

DR. FARLEY: This is a question for the manufacturers. I'm wondering, from a practical standpoint, given this discussion, and if we were to choose to go down the route of quadravalent vaccine with say this 60 microgram total, so not making the compromise on the antigen content, what would be the time table of when this could even possibly happen? I mean are we talking about two years down the road, or more, or less?

DR. COLGATE: That's a difficult one really. I mean it normally takes two to three years to get a new plant up and running. Increasing capacity by 25 percent would really depend on the individual circumstances of the company. If they have space just to increase the size of the facility, then I guess it could, may be done earlier.

But I think also the regulatory hurdles are probably going to be the constraining issues basically. And I think we, I think we need some kind of clear directive about that. That is, is it required and there is a regulatory pathway and what kind of clinical requirements are also there. I think it would need to be spelled out to us very clearly exactly what is required and the mechanism for doing it. And if that's done, I'm sure, as we have done in the past, we would respond.

DR. KARRON: Dr. Wharton?

MS. BAXTER: Marguerite Baxter with Novartis Vaccines. I just wanted to add to Tony's comment to sensitize the Committee. The other factor that would have to be considered is it would actually be necessary to enact tax legislation to include a quadravalent vaccine in the vaccine injury compensation program. Because the way the law is written now, it only covers trivalent influenza vaccine. So that would also need to be a step in the process that would need to be factored in.

DR. WHARTON: Yes, it seems like given that this is likely to be a somewhat long range process that is being really the follow-up from the last couple of annual VRBPAC meetings on influenza strain selection with some in-depth discussion this afternoon would be for FDA to, you know, to be able to define for us, or to be able to define what the regulatory pathway is that such products would have to go through. And I'm sure there are some clinical studies that would need to be done so that we all would be sure we understood what a quadravalent vaccine would mean.

And so given that we are on a journey here, I think. I think we're not planning on getting there this afternoon that those are some steps forward that would move this process forward, just as this discussion has as well.

DR. STAPLETON: Yes, I think just as several people have eluded to, it would be interesting and should be feasible to retrospectively look at drift within different B-lineages and come up with some confidence intervals with how likely you are to have a major mismatch, based on previous years, which might actually improve the at-risk manufacturing process.

DR. COUCH: I just wanted to add that I tried dose-response data to Influenza B in infants and very young children. I couldn't find anything. None.

DR. KARRON: One other question I have and I don't know if this would be useful or not, but we're also, we're in a changing era with regard to influenza vaccination. We're vaccinating more children than we ever did before and we're increasing the age range in which we vaccinate children. And I guess I was also wondering as part of this journey, if you will, whether it's useful to do any kind of modeling to look at the various options, taking into account the B-lineage strains that have circulated, the rates of vaccination, if you alternated strains in a vaccine what would it do? If you had a quadravalent vaccine, let's say of 60 micrograms, what would it do? We may not know enough about half doses to really be able to model that, but whether it would be useful to do some of that as part of the thinking process.

Are there other comments, thoughts form members of the Committee, or FDA, or from the audience?

Yes, Dr. Eickhoff?

DR. EICKHOFF: As part of this process, could we formally ask CBER and/or CDC representatives to take the issue of alternating strains year-to-year to WHO for their consideration next year?

DR. KARRON: I don't know --

DR. COUCH: Well, since you're at the hand, this is almost to the side, sorry. But I almost did it up there but I forgot. I wanted to, I think we've done it before, let's thank CDC, and Nancy, and Dr. Klimov, and Dr. Ye for that presentation this morning because that's a huge amount of work that they bring to these decisions for us. And one of my reactions was that if they'd just give us less data, they'd probably have much less discussion, and the decisions we'd make would be much simpler, but we wouldn't encourage that.

DR. KARRON: Absolutely. Of course, I think we should thank all of the people who've worked very hard on our behalf.

I do though want to follow-up on Dr. Eickhoff's question and ask whether these deliberations could be brought back to the WHO?

DR. COX: Yes, I think it would be very important to bring these deliberations back to WHO. We do often spend extra time; we even started our meeting on Sunday afternoon this past year so we could spend half a day deliberating about H5 vaccines and going over that data. So we do find time for special topics and I think it would be very useful to invite, and we can invite outside experts. And it would be very useful to have the same kind of deliberations and really get feedback, because on that one occasion or two occasions, as Dr. Couch pointed out, we did have to say within the WHO recommendations, either B/Vic or B/Yamagata lineage virus, whichever is most appropriate, because the distributions were very different.

I think that it would be extremely useful to begin thinking about clinical trials. And exactly what it would take to put together a clinical trial that involved 15 micrograms of each B, and seven and a half of each B, that would really help answer some of the questions that have come up today and give us a lot more substance to deal with as we move forward with some of the difficult decisions.

DR. KARRON: Norman?

DR. BAYLOR: I just wanted to say, I mean, what we'll do as far as the FDA, I mean we look forward to working with the manufacturers and probably pursuing this discussion a little further. We have meetings with the Influenza Manufacturing Group Pharma, and this is something I think we can bring up as an agenda item and discuss the feasibility of this. And what we can do internally is we can do some, create some scenarios on what kind of clinical trials we would need or develop to answer some of these questions, you know, using, you know, with all the options, looking of the options of alternating, or a quadravalent, what would it take. We can actually outline what we think would be a likely clinical trial to design.

I think also, Ruth, your comment about the modeling I think would be important because I really, I think we need to know among the options what do we really gain. I mean it would really be helpful to say if we go this route we gain this much. And that way we'll have a better idea of which one of the options to pick, or do we pick any of them. I mean, as Bob had indicated in his first option, do you stay the course. And I think it's important to evaluate all of those.

DR. COUCH: My last slide which I didn't use said data is needed.

DR. KARRON: Any other comments?

In that case, I'd like to thank everybody for attending this VRBPAC meeting and we're adjourned.

(Whereupon, the above-entitled matter went off the record at 3:30 p.m.)