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

Advisory Committees

  • Print
  • Share
  • E-mail

February 25, 2011: Vaccines and Related Biological Products Advisory Committee Meeting Transcript

February 25, 2011

DoubleTree Hotel Bethesda
8120 Wisconsin Avenue
Bethesda, MD 20814

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

Proceedings by:

CASET Associates, Ltd.
Fairfax, Virginia 22030
caset@caset.net

Table of Contents

Topic: Strain Selection for the Influenza Virus Vaccine for the 2011-2012 Season

Call to Order
Jose Romero

1

Administrative Remarks
Don Jehn

1

Introduction
Jerry Weir

7

U.S. Surveillance
Lisa Grohskopf

12

World Surveillance/Virus Characterization
Nancy Cox

18

DoD Vaccine Effectiveness Report
Kevin Russell

51

Vaccine Responses
Zhiping Ye

70

Availability of Vaccine Viruses and Reagents
William McCormick

77

Comments from Manufacturers
Samson Lee

83

Update on H5N1 and H9N2 Surveillance and Virus Characterization
Nancy Cox

99
Open Public Hearing113
Committee Discussion and Recommendation118

Update on Febrile Seizure Signal after Influenza Vaccine
David Martin

141


P R O C E E D I N G (8:30 a.m.)
Agenda Item: Call to Order
DR. ROMERO: Good morning everyone. We will begin. I will turn this over to Donald Jehn to make some introductory remarks.
Agenda Item: Administrative Remarks
DR. JEHN: Good morning, all. Welcome to -- today is actually the 125th Vaccines and Related Biological Products meeting. I am Don Jehn, the designated Federal official for today's meeting and would like to welcome all of you. I would also like to congratulate Dr. Romero on being accepted to be Chair of this Committee for this year. We have, also, a couple new members that have not attended a VRBPAC yet -- Dr. Gray and Dr. Schoolnik. We will be going around the table later to introduce everybody.
Today's session will consist of presentations that are open to the public, as described in the Federal Register Notice of January 20, 2011. I would like to request that any media inquiries be directed to Ms. Shelley Burgess from the FDA Office of Public Affairs. Also we would like to request that everyone please check your cell phones and pagers to make sure they are off or in the silent mode.
Now I need to read into public record the Conflict of Interest Statement for today's meeting. The Food and Drug Administration (FDA) is convening the February 25, 2011 meeting of the Vaccines and Related Biological Products Advisory Committee under the authority of the Federal Advisory Committee Act (FACA) of 1972. With the exception of the industry representative, all participants of the Committee are special government employees or regular Federal employees from other agencies and are subject to the Federal conflict of interest laws and regulations.
The following information on the status of this Advisory Committee's compliance with Federal ethics and conflict of interest laws, including, but not limited to 18 US Code 208 and 712 of the Federal Food Drug and Cosmetic Act, are being provided to participants at this meeting and to the public. FDA has determined that all members of this Advisory Committee are in compliance with Federal Ethics and Conflicts of Interest Laws.
Under 18 US Code 208, Congress has authorized FDA to grant waivers to special government employees and regular government employees, who have financial conflicts when it is determined that the Agency's need for a particular individual's service outweighs his or her potential financial conflict of interest.
Under 712 of Food Drug and Cosmetic Act, Congress has authorized FDA to grant waivers to special government employees and regular government employees with potential financial conflicts when necessary to afford the Committee of their essential expertise.
Related to the discussion of this meeting, members and consultants of this Committee have been screened for potential financial conflicts of interest of their own, as well as those imputed to them, including those of their spouses and minor children and, for the purposes of 18 US Code 208, their employers. These interests may include investments, consulting, expert witness testimony, contracts and grants, CRADAs, teaching/speaking/writing, patents and royalties, and also primary employment.
For topic one, the Committee will discuss and make recommendations on the selection of strains to be included in the influenza virus vaccine for the 2011-2012 influenza season. This is a particular matter involving specific parties. The Committee will also hear an update regarding the H5N1 and H9N2 surveillance and virus characterization. There is also an additional update that was added on febrile seizure signal after influenza vaccine. These are non-particular matters.
Based on the agenda and all financial interest reported by members and consultants, no waivers were issued under 18 US Code 208 B3 and 712 of the Food, Drug, and Cosmetic Act. Dr. Margaret Rennels is serving as the industry rep, acting on the behalf of all related industry. She is employed by GlaxoSmithKline in Washington D.C. Industry representatives are not special government employees and do not vote.
In addition, there may be regulated industry and other outside organization speakers making presentations. These speakers may have financial interest 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 are not screened by the FDA for conflicts of interest.
This Conflict of Interest Statement will be available for review at the registration table. We would like to remind members, consultants, and participants that if 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 affected firms, their products, and, if known, their direct competitors. Thank you. Dr. Romero, I will turn it over to you.
DR. ROMERO: Thank you and, again, good morning and welcome to the 125th VRBPAC meeting. Why don't we begin by introducing ourselves, the members and the consultants, and then state where you are from. I will start with myself and then move to Dr. Debold.
I am Jose Romero. I am from the University of Arkansas for Medical Sciences.
DR. DEBOLD: I'm Vicki Debold with the National Vaccine Information Center.
DR. EICKHOFF: I am Ted Eickhoff, University of Colorado, Denver.
DR. COX: Nancy Cox, Centers for Disease Control and Prevention, Atlanta.
DR. SCHOOLNIK: Gary Schoolnik, Stanford Medical School.
DR. WHARTON: Melinda Wharton, Centers for Disease Control and Prevention, Atlanta.
DR. MCINNES: Pamela McInnes, National Institutes of Health.
DR. LEVANDOWSKI: Roland Levandowski. I am an infectious disease physician here in Bethesda, Maryland. Formerly with the US Public Health Service, but for the last several years I have had no institutional affiliation and work as a volunteer for public health and other not-for-profit organizations interested in control and surveillance of flu.
DR. GRAY: Greg Gray, University of Florida.
DR. DURBIN: Anna Durbin, the Johns Hopkins School of Public Health.
DR. AIR: Gillian Air, University of Oklahoma Health Sciences Center.
DR. WEIR: Jerry Weir, the Division of Viral Products, CBER.
DR. BAYLOR: Norman Baylor, Office of Vaccines, CBER.
DR. TACKET: Carol Tacket, University of Maryland.
COL HACHEY: Wayne Hachey, Department of Defense.
DR. GILBERT: Peter Gilbert, Fred Hutchinson Cancer Research Center.
DR. DESTEFANO: Frank DeStefano, Centers for Disease Control and Prevention, Atlanta.
DR. RENNELS: Margaret Rennels, GlaxoSmithKline, industry representative.
DR. CHEUNG: Ambrose Cheung, Dartmouth Medical School.
DR. GELLIN: Bruce Gellin, National Vaccine Program Office, Department of Health and Human Services.
DR. ROMERO: Thank you all and welcome. We will begin with Dr. Weir, Jerry Weir giving introductory remarks.

Back to top

Agenda Item: Introduction
DR. WEIR: Thank you and good morning. I am going to give a very brief introduction to the Advisory Committee meeting today. I will try to be really brief considering I guess you have noticed on the agenda we do not have any lunch schedule. I will try to plow through this.
This is our annual strain change selection meeting. The purpose of the meeting is shown on the first slide. It is to review influenza surveillance and epidemiology data, antigenic characteristics of recent virus isolates, serological responses to current vaccines, and the availability of candidate vaccine strains and reagents. Then to discuss and make recommendations for the strains of influenza A H1N1, H3N2, and influenza of B viruses to be included in the next years, 2011-2012 vaccine in the United States.
There are a couple of considerations for vaccine strain selection. I think almost everybody in the room knows this, but vaccine effectiveness depends on the match between the hemagglutinin (HA) and the neuraminidase (NA) of the vaccine and the HA/NA of circulating strains of virus. That is because antigenic drift of the HA and the NA are continuous for influenza A and B. That occurs yearly and that is what necessitates the revisiting of what strains should be in the vaccine. Antibody to HA is correlated with vaccine efficacy.
The other consideration for vaccine strain selection is the timing. The timelines for vaccine production are relatively fixed. The reason that we have this strain selection meeting in February for the next year is because it is necessary to do it at this schedule for the availability of vaccine for the Northern Hemisphere winter season of the upcoming year.
Ample vaccine supplies and timely availability depends on several factors that are influenced by strain selection. Two of these are listed on this slide. The growth properties of the strains and the available reassortants used for vaccine production and the availability of strain-specific reagents for inactivated vaccines to ensure the potency of the vaccines that are produced.
Type of the analyses that are used in the vaccine strain selection and what you will hear today from the various speakers are shown on this slide. First, the epidemiology of circulating strains and surveillance data from the U.S. and around the world, you will hear presentations from CDC on this topic.
You will also hear about analyses of the antigenic relationships among contemporary viruses and candidate vaccine strains. I have several presentations for today from CDC, the Department of Defense, as well as CBER. Some of these types of analyses will include hemagglutination inhibition test with post-infection ferret sera, HI tests using panels of sera from humans receiving trivalent influenza vaccine, possibly some virus neutralization tests, antigenic cartography, which shows the relationships between various viruses, phylogenetic analyses of HA and NA genes, and some vaccine effectiveness data.
You will also hear both from CBER and vaccine manufacturers, a little bit about manufacturing considerations, including the availability and characteristics of vaccine strains and high-growth reassortants and the availability of potency reagents for inactivated vaccines.
To quickly review what we did this time last year, we had a strain selection meeting on February 22nd last year. The recommendation that came out of this meeting is shown on this slide. The Committee recommended the following strains for the current season, 2010-2011. One was an A/California/7/2009(H1N1)-like virus. There was a replacement of the current vaccine following from the 2009-2010 season. This followed the emergence of the novel H1N1 pandemic virus during 2009. The Committee also recommended an A/Perth/16/2009(H3N2)-like virus. This was also a change from the previous 2009-2010 season. The B component was recommended to be B/Brisbane/60/2008. This was not a change from the previous 2009-10 season.
Recently, the World Health Organization met last week and made recommendations for the strains that should be included in trivalent influenza vaccines. The following is what they published shortly after their meeting. It is recommended that the following viruses be used for influenza vaccines in the 2010-2011 influenza season (northern hemisphere winter): an A/California/7/2009(H1N1)-like virus, an A/Perth/16/2009(H3N2)-like, and a B/Brisbane/60/2008-like virus. You will notice this is not a change from last year's recommendation.
As always, they note that in previous years, national regulatory control authorities approved the composition and formulation of vaccines used in each country. That is the role of this VRBP AC, this Committee, to make that recommendation for the U.S. vaccines.
Today, the Committee will discuss the various types of data that will be presented over the next couple of hours. They will discuss which influenza strains should be recommended for the antigenic composition of the 2011-2012 virus vaccine in the United States. The data will include things that I just mentioned on the previous slide -- epidemiology, antigenic characteristics of viruses, for example. Then the Committee will vote on the strains that should be in the vaccine.
I will come back later after the discussion and we will walk through these and set up the voting. It will essentially be some version of what you see here, to retain the current strain or recommend some other possibility. I will stop there.
DR. ROMERO: Thank you very much, Jerry. We will begin then with the U.S. Influenza Surveillance. It will be presented by Dr. Lisa Grohskopf from the CDC.

Back to top

Agenda Item: U.S. Surveillance
DR. GROHSKOPF: I am going to start by presenting some characteristics of virus isolates that have been recently, over the past season, submitted to and tested at CDC.
Since October 1st, 2010, the approximate beginning of the US season, 644 U.S. virus isolates have been tested in CDC laboratories and antigenically characterized. Among these, 85 or 13 percent were characterized as Influenza A (H1N1), of which 100 percent were further subcategorized as A/California/7/2009-like, which matches the component of the current trivalent seasonal vaccine.
348 or 54 percent were characterized as Influenza A (H3N2), of which nearly all, 99 percent, were further sub-characterized as A/Perth/16/2009-like. 211 or 33 percent as Influenza B, 94 percent Victoria lineage, 99 percent of these B/Brisbane/60/2008-like. The majority of the isolates characterized therefore matched these strains that were recommended for inclusion in the currently available northern hemisphere seasonal trivalent vaccine.
With regard to resistance to antiviral agents, high level resistance to the adamantanes, rimantadine and amantadine, continues to be noted among Influenza A isolates. A(H1N1), A(H3N2) and B virus isolates tested remain susceptible to the neuraminidase inhibitors.
The next three slides are geographic maps summarizing the spread and distribution of influenza throughout this season so far. I have three of them. This one is from close to the beginning of the season. It is week 41 of the year, the week ending October 16, 2010. During this week, we had relatively low influenza activity with a majority of states reporting -- approximately half the states, those denoted by the diagonal lines, reporting no activity, and the other half, denoted by the cross-hatched lines showing sporadic activity.
Moving forward into week 48, about seven weeks later, December 4th, as we might expect, we have some increase in activity. At this point, almost all states are reporting at least sporadic activity and we also have twelve states denoted in yellow reporting local activity. This week and the week preceding were the first time we started to see some regional activity reported. In this case, by this time, it was Georgia, Kentucky, and Virginia.
This slide is for week six of the current year, 2011. It is the week ending February 12, 2011. As we might expect, since it is February, we have an appreciable increase in activity with a majority of states reporting at least regional activity and most reporting widespread activity.
This slide summarizes data from the U.S. Outpatient Influenza-like Illness Surveillance Network or ILINet. The graph denotes percent of outpatient visits reported by participating sites that are due to ILI. ILINet includes about 3,000 health care providers and captures about 25 million patient visits a year.
The National baseline for ILI activity is represented by the dotted line and is calculated from the proportion of visits reported by these providers that are reportedly due to ILI during the previous three seasons during non-influenza season weeks. At this period of time, currently it is calculated as about 2.5 percent. The most recent week for which the data had been published, the current percent of out-patient visits for ILI is 4.5 percent, which is somewhat above the national baseline.
I should also mention that the definition of ILI for the purpose of this surveillance system is not laboratory-confirmed disease, but is fever defined as either temperature of greater than 100 or 37.8 -- 100 being Fahrenheit, 37.8 being Celsius and either cough and/or sore throat, for which no other explanation is available other than influenza.
This slide summarizes data on virologic typing from the 80 WHO collaborating laboratories located in the US, as well as 60 National Respiratory and Enteric Virus Surveillance System collaborating laboratories. These laboratories will type influenza isolates into A and B and many also further sub-type As into A (H3) and (H1). The most recent data on this graph is for the week, again, ending February 12, 2011. It is represented by the right-most bar on the graph.
Approximately 21 percent of isolates were types as B. These are represented in green. Approximately 40 percent were typed as A (H3). Those are in red. The yellow are type A for which no sub-typing was performed. One trend notable on this graph is that if we look at the orange portion of the bars, which is up at the top of the graph, there has been something of a steady increase in the proportion that are of the A virus isolates that are further characterized as A (H1N1), particularly if we look at December and January.
This slide summarizes data on pneumonia and influenza or PNI mortality from the 122 U.S. Cities surveillance system. The epidemic threshold and seasonal baselines are calculated using regression modeling techniques from data from the corresponding week's period from the preceding five years. Those are the sinusoidal black curves that you see on the graph.
For the current week, you can see at the very far right, for this week in 2011, we have some activity -- some increase in pneumonia and influenza mortality above the epidemic threshold -- actually, for the past couple of weeks.
This is the last data slide. This slide summarizes pediatric mortality. Deaths in children confirmed associated with influenza, that is children under the age of 18, have been considered a reportable event since 2004. Those events are reported to the system. A total of four seasons are represented on the chart from the 2007-2008 season all the way off to the left to the current 2010-2011 season in the right most section.
To date, so far this season, a total of 35 pediatric deaths were reported to the system. The ones that are represented in the light blue color actually were the ones that were reported for this week. You can see there is something of a lag as those light blue sub-bars are not all for the current week.
Among these 35, 13 were reportedly associated with influenza B. That is about a little more than one third. Nine or approximately one quarter were A(H3) and seven, approximately twenty percent were A(H1N1). The remaining six did not have any further sub-typing reportedly done.
In summary, Influenza A(H3N2), A(H1N1), and B strains continue to co-circulate in the U.S. Recently characterized strains appear to be well matched to the recommended 2010-2011 seasonal vaccine. Recently circulating viruses remain susceptible to neuraminidase inhibitors. In the U.S., influenza activity remains high in recent weeks with an increase in proportion of influenza A viruses identified as A(H1N1) since the beginning of the U.S. season. Thanks.
DR. ROMERO: Thank you very much. Any questions for Dr. Grohskopf? Very good. Thank you. Next we will hear from Dr. Nancy Cox from the Centers for Disease Control on the world surveillance and virus characterization.

Back to top

Agenda Item: World Surveillance/Virus Characterization
DR. COX: Thanks very much. As usual, I am going to give a presentation mainly concentrating on the virology of the current strains that are circulating globally. I really tried to incorporate data from other WHO collaborating centers to give you a more global feel for the kind of data that we actually examine at the WHO meeting in Geneva.
The first slide I would like to show you is just a snapshot of pretty much where we are now. The intensity of the color is indicative of the level of influenza activity that is occurring in the various countries. The darker the color, the more intense the activity.
This is just a snapshot of one week at the end of January, but we do receive reports from a wide variety of regions and countries throughout the world. Of course, this is during the northern hemisphere winter so there is a lot more activity occurring in the north than in the southern hemisphere.
The pie charts simply demonstrate the percentage of the viruses isolated that are either A(H1N1) 2009 pandemic viruses -- we are still working on new nomenclature for these viruses -- A(H3N2) viruses, shown in blue, and Influenza B viruses shown in the dark brown. You can see that depending on the country/region of the world you are looking at -- this is just for a one week snapshot, but if you look for longer periods, you see a similar pattern of a lot of variability with respect to which viruses are actually predominating in different regions of the world.
Here we are looking at a pie chart showing the percent of viruses by type and subtype that have been reported to WHO's Global Influenza Surveillance Network, which I will refer to periodically as GISN as I go through my talk.
If we look globally, about 27 percent of viruses that have been reported to WHO are A(H3N2), 21 percent Influenza B. There were only a few -- a handful of former -- what we are calling now former seasonal H1N1 viruses, those that have been displaced or almost completely displaced by the H1N1 pandemic 2009 viruses, represented here in yellow and comprising 31 percent of the reported viruses.
There are a number of Influenza A viruses that are not sub typed because they have been detected by either indirect immunofluorescence assays or by rapid influenza detection tests. During this time period, up to January 29th, there were ten H5N1 viruses that had been reported.
If you look at this slide, you will see the temporal distribution of the different types and subtypes of influenza. You will see that, generally speaking, H1N1 viruses tended to predominate mainly in Europe, while H3N2 viruses predominated in some other countries, including the US and Canada.
This slide shows the number of specimens received by CDC by type and subtype. You can see that because we had a predominantly H3N2 year, starting off early -- although we have had an increase in the proportion of pandemic H1N1 viruses in recent weeks -- we have more H3N2 viruses than B or H1N1 viruses to look at in our WHO CC.
So now if we go into more detail about the viruses, themselves -- I just wanted to show, first of all, where H1N1 pandemic activity had occurred throughout the world. There was a lot in Europe, including Russia. Essentially the epidemic spread from east, occurring first in the UK, and then spreading westward across Europe and into Asia. India also reported quite a lot of pandemic H1N1 activity during this period. Regional pandemic H1N1 activity occurred in the US. Local activity in many other countries.
In terms of the summary, global summary of H1N1 epidemiology, we can see that sporadic former seasonal A(H1N1) viruses were detected in very few countries. I will name them later. H1N1 pandemic '09 viruses co-circulated with A(H3N2) and B viruses. In the northern hemisphere, there were widespread or regional outbreaks of these viruses in Asia and Europe, in particular North Africa and in parts of North America.
In the southern hemisphere, in general we can say that there was very low activity in Southern Africa, South America, and Oceania. However, sporadic to widespread outbreaks did occur in some countries in South America in the very beginning of our winter season, the end of their winter season.
Here is where I will summarize our findings for the former seasonal influenza A(H1N1) viruses that had been represented in the previous vaccine by A/Brisbane/59. There have only been rare detections, as I said before, with only six viruses analyzed. They were identified in China, Malaysia, the Russian Federation, Tunisia, and one from Virginia in the U.S.
These former seasonal viruses were antigenically and genetically similar to A/Brisbane/59/2007 and they belong to genetic clade 2B. So there is nothing really unusual about these viruses. They were resistant to oseltamivir, as we had noted for those previous seasonal H1N1 viruses, but they were sensitive to zanamivir.
Moving on, we had A(H1N1)pdm09 viruses from about 100 countries around the world to analyze by the WHO collaborating centers. We did the typical testing that Jerry mentioned.
If we compare what has gone on in previous seasons, we will just look at 2009 here, which is where these viruses first began to emerge, causing summer activity in the northern hemisphere and wintertime activity in the southern hemisphere during this period. This is the peak of activity in the northern hemisphere. You can see the scale indicates that there were really a very large number of detections of pandemic H1N1 viruses.
In 2010, we were having a decrease of activity, starting at the beginning of the year. It had gone down here, continued to go down, and there really was not much pandemic H1N1 activity in the southern hemisphere, but then it began to increase in the winter of 2010 and here we are during the first few weeks of January.
For those of you in the back of the room, you certainly will not be able to see these numbers. I will just describe verbally what we are seeing in this hemagglutination inhibition test of H1N1 pandemic viruses.
Here we have our vaccine strain, represented by California/07/2009. We have a homologous HI titer of 2580. What you can see is that the viruses -- the test antigens, which start here, and they are representative of viruses isolated in a number of different regions of the world, the U.S., the southern hemisphere, here, Europe, here, and Asia, here. The majority of these viruses are very well inhibited by post-infection ferret sera to the California/07 virus. There really has not been much antigenic variation observed at all.
Here is a similar hemagglutination inhibition test that was performed by the WHO collaborating center in Beijing, which became a WHO collaborating center only recently. This is the first time that this entity has participated fully in the WHO vaccine strain selection process.
They are using some ferret sera produced by CDC and some ferret sera that they have produced by themselves. For these China viruses that were isolated during mainly 2011 -- so these are recent viruses from a variety of different locations around China. You can see that they are very well inhibited by antiserum to the California/7 virus.
It is much easier to actually look at antigenic cartography when you are considering the antigenic relationships of all of these different viruses. Basically, what you are doing is actually putting all of the data that you saw in those previous slides into a two-dimensional map. The mapping is done much like a map is drawn to determine the relationships and positions of different cities or different landmarks on a map that we are so accustomed to reading as we navigate from place to place.
Right in the center here you see the A/California/7/2009 virus. There are many, many, many influenza viruses represented here and actually dots upon dots upon dots. These squares, open squares, represent different sera that have been used in the HI assays. You can see that the viruses are in the main, clustering right around the A/California/7/2009 vaccine virus. This is a very nice way to depict that, although there are a few outliers. The majority of the viruses are, in fact, A/California/7/2009-like. This is a compilation of data from all five WHO collaborating centers.
Some of these outliers have been looked at in detail. We know precisely what amino acid changes have occurred in the hemagglutinins of these viruses that cause them to be outliers.
Here is a frequency table compiled by us at the CDC -- our WHO collaborating center at CDC. If we look at the previous period of March 2010 to August 2010, you see we only had one percent of viruses that were low to the California/7 vaccine strain. During the current period, September to January 2011, again we see a very small number and very low percentage of the total number of viruses, 215 that we characterized in detail, that are low reactors.
For our collaborating center, those low reactors have all had changes in the hemagglutinin in certain positions, ranging from amino acids 154-158 in the hemagglutinin.
We like to compile the information from all of the five WHO collaborating centers for human influenza. This is the data for our WHO CC. This is for the Chinese collaborating center in Beijing. This is for the Japanese collaborating center in Tokyo. This is for the British collaborating center in London. And this is for the Australian collaborating center in Melbourne. These acronyms will be used throughout.
We can see that overall almost a thousand viruses have been analyzed in great detail. Only about five percent have been low reactors. This is a very reassuring picture.
For those of you in the back of the room, you cannot possibly see this. You do have this evolutionary tree in your package. The most important thing to note is that -- there are two important things to note. One is that there is increasing genetic diversity in the hemagglutinin.
The second thing to note is that we have a number of recent viruses, which we have sequenced. We have both the HA and the NA sequence data shown in your package. The January viruses are shown in pink. Those isolated in December are shown in orange. Those isolated in November are shown in green. So we have quite a number of viruses.
There are occasional low reactors that are designated by an LR after the strain designation. They scatter throughout the tree. As I mentioned before, they are primarily indicated by changes -- you can see this on the tree -- an indication that they have had changes in the 153-158 channel.
You will see that there are some recent viruses down here that are a bit different from the majority, but antigenically we cannot really distinguish viruses throughout the tree.
The neuraminidase phylogenetic analysis really mirrors that for the HA. So these are the viruses that were shown -- the neuraminidase is the viruses shown at the top of the tree for the HA and so on. Here is out California/7 vaccine virus.
As you have heard from Lisa, she gave you a distillation of our results for antiviral susceptibility testing for the different types and subtypes of influenza viruses circulating. About 2,000 samples of H1N1 pandemic 09 viruses were analyzed. There were 30 oseltamivir resistant viruses, all with this particular signature mutation in the neuraminidase. But only one had been detected at CDC since October 2010 and that was from a treated patient. No zanamivir resistant viruses were detected. The picture for the M2 blockers was similar to what we had seen before.
In summary, what we can say is that the former seasonal A(H1N1) viruses continue to appear to have been replaced by the newer A(H1N1)pdm09 viruses. Increasing heterogeneity has been observed among the HA and the NA gene sequences of these viruses. Recent viruses tend to form three HA gene subclades, which I pointed out. The vast majority are antigenically closely related to California/7. A minority of viruses had reduced titers with post-infection ferret sera.
Moving along to the H3N2 viruses, which we have fondly called our problem children this year, we see that for H3N2 viruses we had widespread activity in Chile in September and then we had widespread activity caused by H3N2 viruses in Canada and the U.S. and quite a bit of H3N2 activity in Mexico, as well.
H3N2 viruses have co-circulated with H1 and B viruses. There was, as I said, widespread activity in the U.S. and Canada, regional activity in Cambodia and China, Hong Kong SAR in September and October, and there has been sporadic activity in some countries in north Africa and Europe.
I mentioned the widespread activity in Chile in September. Although activity has remained very low during this period in all countries except Chile in South America, H3N2 viruses have been predominant. So it is low level activity with H3 predominating in South America. Then sporadic activity has occurred in other countries.
If we look at the number of H3N2 viruses detected in the Global Influenza Surveillance Network, we want to go back to 2009. This is the winter of 2009, prior to the emergence of the pandemic H1N1 viruses. You will see that we had quite a few viruses identified in the network.
This was the southern hemisphere season in 2009. We had a couple of peaks. As the 2010 season began, fairly low levels of activity -- sorry, this is the southern hemisphere; this is the northern hemisphere 2010. This is where we are in 2011. Still quite a bit of H3N2 activity.
Here we are looking at the number of H3N2 viruses characterized by each of the WHO CCs, again, using the same designations as I mentioned before. You can see that the WHO CCs in Atlanta and in Beijing had the most influenza A(H3N2) viruses to characterize.
Here we have a very typical H3N2 hemagglutination inhibition test carried out at CDC. Because the H3N2 viruses do no agglutinate turkey red blood cells, we have transitioned to using guinea pig red blood cells for all of our hemagglutination inhibition tests and for our post-vaccine human serological assays.
What we can see here is the Perth/16 recommended vaccine virus. This is the vaccine virus that is Brisbane/10-like, the previous H3N2 vaccine component. We can see with a homologous titer of 320, we have really pretty good coverage going down here. We do have one virus, which is eight-fold reduced. By and large, there is good coverage.
Here I have a hemagglutination inhibition test performed by the WHO collaborating center in Tokyo. Here is their homologous titer. They have used the Victoria/210, Perth/16-like virus here. You will see that there is really very good coverage. There is only one virus that has a four-fold reduction in titer compared to the vaccine virus. These were relatively recent viruses, most of which were identified in Japan. We saw a very similar picture whether we looked at data from the U.S., from Japan, and so on.
If we look at our frequency table, you will see that we have a very low proportion of viruses, which have an eight-fold or greater reduction as compared to the homologous Perth/16 titer in our test using post-infection ferret sera.
If we look globally -- again, we have all of the different WHO CCs represented here -- the number of viruses that are Perth-like and those that have an eight-fold or greater reduction in titer to the homologous Perth/16 titer, you will see that overall only about between five and six percent are reduced in titer to the Perth/16.
I would like to point out that there is one outlier here in the WHO collaborating center in London, which had analyzed a relatively small number of H3N2 viruses using hemagglutination inhibition assays because they were having particular difficulties with hemagglutination with their guinea pig red blood cells. When they changed and got different lots of guinea pig red blood cells, they actually go better results. So there is a lot of biology, as you know, going on in these multi-component hemagglutination inhibition assays. Their CC struggled, in particular, with this issue.
If we look at the antigenic cartography of the A(H3N2) viruses, this is a time series for data generated by our WHO CC in Atlanta. These antigenic maps, I should say, are produced by the University of Cambridge Group, headed by Derrick Smith.
In February 2010, a year ago, this is what we were seeing with the Perth/16 vaccine virus. Here, this is the A/Brisbane/10 previous vaccine virus. The circle indicates those strains which are antigenically matched to the Perth/16.
Moving on to September 2010, this is the picture that we are seeing. Still the majority of viruses are well matched to the Perth/16 vaccine virus. Then this is the latest in the time series where the viruses remain well matched.
If we do a compilation of data from a number of the WHO CCs, this is what you see. The old Brisbane/10/2007 vaccine virus is here. The Perth/16 is here. There are two different genetic subgroups within what we have designated overall when it emerged as the Perth/16 genetic group. The Perth/16 viruses -- the ones that fall into the Perth/16 clade are shown in red. Those that fall into the other genetic group, which I will show you on the next slide, Vic/208 clade, are shown in green. If anything, it looks like those viruses have moved antigenically a bit closer to the Brisbane/10 than the Perth/16-like viruses.
Because we and other groups had had, as I mentioned, some difficulties with some of the viruses not agglutinating red blood cells particularly well, we did a very careful series of tests using the same viruses, precisely the same viruses, not passaged more or less in hemagglutination inhibition and in neutralization plaque reduction assays and also in micro-neutralization assays. We were trying to really understand whether or not there had been antigenic variation that we were not able to detect using our typical hemagglutination inhibition assays by supplementing those tests with additional functional assays, which, as I said, were the neutralization, plaque reduction, and micro-neutralization assays.
Here we have the Perth/16 virus, another Perth/16-like virus represented by Wisconsin, we have an egg isolate shown here, and its MDCK counterpart shown here, which is representative of what is circulating. As many of you know, sometimes when you pass viruses in eggs, which you must do for vaccine production, you do get slight changes in the way that those viruses induce antibodies in post-infection ferret sera -- sometimes differences in the way the viruses, themselves, behave. But this Wisconsin virus is very representative of the Perth/16-like virus group.
What we can see is that there is very good coverage, whether we are talking about hemagglutination inhibition or plaque reduction micro-neutralization. So we are not getting a huge drop-off in titers here. This was very reassuring.
I wanted to include some additional data that was produced by the WHO collaborating center in London where they did plaque reduction neutralization. They did not use the micro-neut test as we had done, but, again, you do not see a dramatic drop-off. There is some reduction in titer here, but not a dramatic drop-off as you see here with the old vaccine virus.
Because I know that Zhiping will not have time to show the micro-neutralization test that we did with the pre- and post-human vaccine sera, I thought I would show you that here. Again, we were trying to back up our hemagglutination assays with a functional assay so we took serum from individuals vaccinated in Japan, the UK, and China and we really wanted to compare these two columns here, the pre-vaccination GMT -- once again, I want to emphasize this is a micro-neutralization titer -- and then we look at the post-vaccine titer and we compare the homologous titer to the Perth/16 vaccine virus to those obtained for the more recent viruses that have been circulating.
What we are looking for is a 50 percent or greater reduction in titer and we do not see that for any of these viruses. These are panels of serum taken from immunized adult populations in these different countries.
Then if we look at serum taken from elderly volunteers, who were vaccinated with the Perth/16 vaccine virus, again, we see a similar thing where there are not dramatic reductions in post-vaccine titers for any of the viruses.
We also wanted to look at pediatric populations because pediatric populations often behave a bit like ferrets. If there is a reduction, we are much more likely to see it. We had two panels of serum taken from pediatric volunteers, one if the US and one in China. Again, we see that there is really a good antibody rise if you look pre-vaccine to post-vaccine GMT. We do not see any dramatic reductions, certainly not at all in the panel from China.
I will not go through this in detail but basically we have compared, here, our results that we obtained using hemagglutination inhibition with our human serum and our results obtained using the micro-neutralization assay. What we are looking for is a sense that there is a greater than 50 percent reduction. So we would be looking for viruses that fall below this black line, here. We were very reassured by the micro-neutralization human serology.
As I mentioned, there are two distinct genetic groups -- actually, more subgroups, but basically there is the Victoria/208 genetic group here, outlined by this bracket, and then we have the Perth subgroups here. We sub-classified these as Perth A and Perth B.
The majority of viruses circulating in Europe and North America have fallen into the A/208 group, but I would like to emphasize that in China they saw something quite different. A majority of their viruses were circulating down here in the Perth A or Perth B groups, although there was some mixing and matching.
You can see that there is quite a bit of genetic diversity, but as I was able to demonstrate before, even though we have looked extremely extensively using a variety of different assays, we are not able to detect clear cut antigenic differences.
The phylogenetic analysis of the neuraminidase gene sequences really mirrors that of the HA sequences. So for the comparable virus, you have the neuraminidase falling either into the Victoria/208 group or the Perth groups A and B.
We really, in addition, tried to look at the hemagglutinin structure and understand where the various changes have occurred and whether there may be something that we have missed in any of our assays that we might be able to discern structurally or if there are any other kinds of tests that we should be doing. I would just like to point out that on this panel here we have the Perth/16 clade viruses represented and here we have the Vic/208 clade viruses represented.
The two signature changes here at the top of the molecule that have defined and differentiated both of these genetic groups from the previous Victoria/10 vaccine virus were right here at the top of the molecule shown in yellow. Changes at these sites and of this particular nature have tended to cause antigenic variation in the past. We can very clearly see why we were able to distinguish the Brisbane/10 viruses from the Perth/16 viruses.
Within the genetic group, you can see that there are these different subgroups that are outline here. But when we look carefully at these changes, we do not see signature changes that have, in the past, really denoted major antigenic change. Once again, it was somewhat reassuring that we were not seeing anything when we looked structurally at the HA that would indicate, based on past experiences, that we had something going on that we were not able to detect in our assays.
I think Lisa has really summarized the antiviral susceptibility of A(H3N2) viruses so I will not mention anything in particular about these results here.
In summary, unlike in Europe, the majority of viruses that were detected in the U.S. and Canada were A(H3N2) viruses. The majority of these, globally, the H3N2 viruses were antigenically related to Perth/16 virus. A subset of H3N2 viruses were selected for further characterization in micro-neutralization and neutralization plaque reduction assays and these results supported our HI data.
The H3N2 viruses continue to fall into these two major genetic groups, which we had seen develop for the southern hemisphere vaccine strain selection process. But they are essentially antigenically indistinguishable.
Viruses that have somewhat reduced titers to the Perth/16 antiserum in HI assays are scattered throughout the phylogenetic tree and do not share a common signature amino acid changes.
I know that was a lot of information on H3N2. Hopefully the Influenza Bs will be quite straightforward. We have seen quite a bit of Influenza B activity during the recent period, September to January. You can see that while we had local activity in the US and Canada, really the majority of Influenza B activity occurred in some places in Scandinavia, southern Europe, and quite a bit of activity in China and Russia.
Influenza B viruses co-circulated. The B/Victoria lineage viruses predominated in many parts of the world, however, B/Yamagata lineage viruses did predominate in China. I mentioned the B activity in Canada and the U.S., which was increasing during December and January. Influenza B predominated in Norway, the Russian Federation, and the Ukraine. There was regional or widespread activity reported to WHO by Algeria and Israel in December and January. Then there was just sporadic activity in the southern hemisphere.
This shows the number of Influenza B viruses reported to the Global Influenza Surveillance Network. You can see that for the current period -- so for the 2010 season, you see quite a bit of activity represented by this fairly large number of viruses reported. Then into January we see a lot of Influenza B viruses being reported.
Again, we are looking at the number of viruses that were characterized by the various five WHO collaborating centers. CDC and Melbourne had the largest number of Influenza B viruses to actually look at.
For the B/Victoria lineage viruses -- these are the viruses that are represented in the current vaccine strain -- we have the Brisbane/60 --we have too many Brisbanes -- the Brisbane/60 vaccine virus represented here. Viruses tend to be well covered. There are a few exceptions. Here we have a virus from Saskatchewan, which is quite low. I will show you where that falls on the phylogenetic tree. You can see the difference -- why that virus may react differently. But by and large, in our hands, the viruses are B/Brisbane/60-like.
Here is another hemagglutination inhibition assay performed at CDC. You can see here is the Brisbane/60 homologous titer here and a variety of viruses, most of which are fairly recently isolated from a variety of areas around the world. By and large, they are B/Brisbane/60-like.
Our own frequency table demonstrates this. For the Brisbane viruses that are outlined here in green -- so this is the Vic lineage viruses outlined in green here -- we have a small proportion of viruses with an eight-fold or greater reduction in titers compared to the homologous Brisbane vaccine titer.
The Yamagata lineage viruses are actually shown here in yellow. We are representing this lineage of viruses by the B/Wisconsin/01/2009 reference virus.
If we just concentrate on the Victoria lineage viruses on this particular slide, you will see that we have characterized in great detail over 600 viruses of the B/Victoria lineage and only two percent have an eight-fold or greater reduction in titer from all of the WHO CCs.
The antigenic cartography for Influenza B viruses is just a little bit messier. Here we have the B/Brisbane/60 vaccine virus shown here. Then we have some viruses, which are actually a bit closer to the previous B/Malaysia vaccine virus. There is one particular amino acid change at 165, which causes these viruses to cluster more closely.
Here is our evolutionary tree for B/Vic lineage viruses. You can see that we have these different groups: one, two, three, four, five, and six. That virus that I pointed out that was a low reactor from Canada, from Saskatchewan is down here in group five. This is a clade of viruses that we have seen relatively rarely. Viruses in this clade tend to be low reactors, but they do not seem to be spreading in any great numbers. The majority of the viruses are here and related to the vaccine virus shown here in red.
We see a fairly similar corresponding evolutionary tree for the NA genes. The one difference being that we see intra-clade reassortants, shown here, where you have the hemagglutinin from one sub-lineage and the neuraminidase from another. Basically, the majority of the neuraminidases are falling into this large group up here.
Moving right along to the Yamagata lineage viruses, which we certainly must not forget about, as they have predominated in China, we have had fewer viruses at CDC to look at, but what we can say is that the B/Yamagata lineage viruses have moved on from the previous B/Florida/04/2006 vaccine virus, shown here with a homologous titer of 1280 and reductions of eight-fold or greater for these test antigens.
You can see that these test antigens, recent viruses are much better covered by post-infection ferret serum to the Bangladesh/3333 and these other viruses, which are similar to it. We have the Wisconsin virus shown here and we have good coverage by all of these viruses that are more contemporary.
Here we have a hemagglutination inhibition test performed by the China CDC. It is a little hard even for me to see this. They are seeing not such great reductions with their ferret sera to Florida/4, but there is an indication that they are picking up a number of viruses with reduced titers, whereas the Bangladesh/3333, which is typical of the viruses that I was pointing out to you on our HI test, does cover contemporary viruses very well.
If we now concentrate on the Yamagata lineage, what we can see is that -- here we are comparing not to the B/Florida/4, but to the B/Bangladesh/3333, which is Florida/4-like -- we have viruses that are well inhibited generally and a very small percentage of viruses that have reduced titers.
Here is the antigenic cartography for B/Yamagata lineage viruses. What you can see is that the Florida/4 viruses are shown here. This is over a longer time period. But the majority of recent viruses are really clustering around the Wisconsin/1/2010 strain. There has really been antigenic movement within the Yamagata lineage viruses.
There was a lot of debate about whether we would be seeing as much B/Brisbane/60 circulation by next year at this time, but that is the predominant lineage. It was very clear that if we were to have chosen a B/Yamagata lineage virus, it would have been Wisconsin-like strain.
Here we can see the evolutionary relationships among the Yamagata lineage viruses. Here is the old B/Florida/4 vaccine virus down here. This group three is the Bangladesh/3333 group. We have our Wisconsin virus and a number of reassortants made by New York Medical College. Those viruses may be mentioned later on.
We also have another vaccine candidate here. The Hubei virus, which was sent also to Doris and she has made a number of reassortants using that master strain. There is not much to say about the neuraminidases. We see a very complementary NA phylogenetic analysis. So the viruses are falling mainly into this group three.
In summary for Influenza B, these viruses have circulated in many countries. The B/Victoria lineage viruses predominated globally except in China where Yamagata lineages did predominate. Most B/Victoria lineage viruses were antigenically and genetically closely related to the vaccine virus. Others fell into minor clades. I mentioned the clade five viruses.
For B/Yamagata lineage viruses, most recent viruses were antigenically closely related to these three viruses that I pointed out on the trees and in the tables: Bangladesh/3333, B/Hubei, and B/Wisconsin/1/2010. Genetically, most of the recent viruses were distinct from Florida/4 and did group with these viruses in group three. And no oseltamivir or zanamivir resistant influenza B viruses were detected.
As mentioned before, these were the recommendations that WHO made. I just wanted to include in my package some web links to recent publications by the WHO group. Then, finally, I would like to acknowledge all of my colleagues in the Influenza Division, in particular, Dr. Sasha Klimov, who is here today in the audience, but everyone here and many others who have contributed to the data presented, as well as to the U.S. State Public Health Labs, the National Influenza Centers scattered around the world, the other WHO collaborating centers, the Essential Regulatory Labs, and many, many other collaborators. Thank you very much.
DR. ROMERO: Thank you. Any questions for Dr. Cox? Comments?
DR. GELLIN: I am always impressed by what a body of work that is. Two questions. One of them, you had mentioned about the low reactors and the biological issues at the lab in London, is that the same issue for the H1 as well as the H3? Because they showed, I think, that there was a higher percentage of H1 low reactors in the London lab, as well. The question really is that a lab problem or does that signal some other evolution, since they have had more H1 this year?
DR. COX: Yes. What we know for H1N1 viruses is that if you isolate them in a particular cell line, you tend to get more of the variance with changes within that region that I mentioned -- the 153 to 158 region. They had been using the particular cell line that pulls out that particular variant or that selects for that particular variant.
They have subsequently switched to normal MDCK cells. I think what they are seeing is a combination of the two phenomena. Plus, they receive a lot of viruses that have been isolated COT, the special kind of MDCK cells. I think it is a combination of they are receiving viruses with changes that cause that reduction in activity and then having some additional low reactors, but they are scattered throughout the tree. There is no corresponding molecular changes that would signal that we have something new.
DR. GELLIN: A second, very different question, in Jerry Weir's slides, he talked about the WHO recommendations as you did and commented that there is always the caveat that local national authorities can decide what they want. How did the China conversation go at the WHO meetings, given the predominance of that other B strain in China? How did they handle that? Or was that just acknowledged that China would have to decide what they are going to do?
DR. COX: Yes. The discussion was quite interesting. It is very good to have China -- I would like to go on record saying the WHO collaborating center in China has contributed a great deal of information. They feel that they will have to make their decision based on the epidemiology in their country. Because we were able to look at data very carefully and say, yes, those Yamagata lineage viruses have moved on from the previous Florida/4 vaccine virus, they would have a clear pathway forward should they decide that they wanted to make that selection to the Yamagata lineage.
DR. CHEUNG: Do you have correlation of mortality with individual strains -- mortality data?
DR. COX: You probably noted that there are fatal cases designated on some of the evolutionary trees. Your question is probably more general than that. What we know is that on a population-basis, more severe seasons -- that is seasons that are characterized by greater numbers of influenza-related excess death and excess hospitalizations -- occurred during H3N2 virus circulation.
However, for the pandemic H1N1 viruses that circulated, we do know that those viruses disproportionately affected the young. It varies by type and subtype. It varies according to whether it is a pandemic or newly emergent virus versus a virus that has circulated in the population for a while. In general, we have more severe seasons when H3N2 viruses predominate.
DR. CHEUNG: Secondly, do you have any data on those strains, for instance, within H1N1, as well as H3N2 and B? I know the percentage that are covered is over 95 percent, but do you have any data on mortality on those that are not covered by the potential vaccine? There were several that you did not cover -- three percent or somewhere around there.
DR. COX: Those viruses are isolated not from fatal cases, just through normal surveillance. Unless it is actually designated on the table, it is not from a fatal case. But we tend to, whenever we do get viruses isolated from fatal cases we tend to subject those viruses to a much greater degree of scrutiny.
DR. DURBIN: I have a quick question. On the H1N1, you are seeing more cases more recently in the United States. Do you think that is just a characteristic of the virus where it comes up later in the season? Or do you think that it is sort of outperforming the H3 virus at this point in the United States?
DR. COX: That is an unanswerable question. I have been asked and many others have been asked why did the UK have so much H1N1 activity when they had a lot of activity last year, in particular, they had quite a lot of activity and hospitalizations and even some deaths in London where they had experienced quite a severe season before -- not severe, severe, but they had seen a lot of H1N1 activity in that population? They do not have an answer for it.
I think it is just a matter of virus seeding in the population in conjunction with antibody and, if you have a lot of vaccination or have had a lot of infections previously, you are much less likely to have H1N1 in those areas. But there are still pockets of the population which have a lot of susceptible so that is where you are going to see the outbreaks. I think in general that is what we can say.
DR. ROMERO: Dr. Eickhoff?
DR. EICKHOFF: Nancy, these two B strains have been co-circulating for what -- about 20 years now?
DR. COX: Yes, longer.
DR. EICKHOFF: Is there any historical data to suggest that when one strain emerges in one part of the globe, as B/Yamagata did in China this year, does that presage or predict what might happen the next year more globally?
DR. COX: We have really tried to look at this thoroughly. What is very frustrating to us about the Influenza B viruses is that there is not a distinct pattern. We had a period during which the Influenza B/Vic viruses circulated only in China and then they sort of crept out to Japan, but they really did not affect North America or Europe. We had a ten year period where Vic viruses did not circulate beyond Asia. Then they suddenly popped. Then they predominated for a period of time. Then the two have co-circulated.
It has become increasingly difficult to predict with any certainty which of the two lineages will predominate. It is a huge dilemma for us. We have started looking at various modeling tools and so on, but we do not have anything that really is ready for prime time yet.
DR. LEVANDOWSKI: I have a question related to the H1, as well. With the shift from H2N2 to H3N2, there was a period of time during which there was reasserting going on between the H2N2 strains and the H3N2 strains. Is there any evidence that there is any kind of reassorting going on between the previous H1N1 strains and the pandemic strain?
DR. COX: Good question. I think I can answer it very clearly. We have developed a technique that can screen for reassortants and we are not seeing reassortment between the old H1N1 and the new H1N1 or between the new H1N1 and H3N2 human viruses.
DR. ROMERO: Any other questions or comments? Thank you very much. Next, we are going to hear from Dr. Kevin Russell, from the Department of Defense on Vaccine Effectiveness Report.

Back to top

Agenda Item: DoD Vaccine Effectiveness Report
DR. RUSSELL: Good morning, everyone. Thank you very much for this opportunity to share what the Department of Defense has been doing with regard to influenza this past season.
I represent the Armed Forces health Surveillance Center that was established in early 2008. I think particularly since the establishment of the Armed Forces Surveillance Center, we are able to bring together many of the different facets and laboratories that work with influenza. I will describe briefly at the beginning of this why that is so.
I want to thank, quickly, as well, Major Ron Burke, who is in the audience, who helped collate a lot of this information from many of our different partners throughout the world.
What I am going to describe quickly is the different systems that contribute to the data that I am showing, both from a laboratory sense and from a healthcare utilization sense in our many, many clinics throughout the world, both ambulatory and in-patient.
I am not going to spend a lot of time talking about strain circulation or molecular analyses, although I will show some slides on what we have contributed to the CDC. 100 percent of these sequences are shared with the CDC. They request a variety of interesting samples, actual samples for further analysis. We might push a few that we have particular interest in to the CDC. We do not do this work in a vacuum within the DoD or within the U.S. government. We are very pleased with the support that the CDC gives us and thankful for that.
I am going to spend a little bit more time on the vaccine effectiveness studies that are done with a variety of partners because I think it is a little bit less represented here and we have some unique databases that can contribute to that.
Quickly, looking at the information where we do pull information from, at the Armed Forces Health Surveillance Center, we have a Division of Epidemiology and Analysis. 1.4 million active duty records -- I will show you a graphic next that really outlines this more clearly, but a lot of information feeds into our Defense Medical Surveillance System. It is a very robust system for doing analyses on because the denominator is known. It is built on a backbone of the demographics of our active duty, not just the healthcare encounters so we know a lot about the people that we are talking about. That is very powerful.
The Global Emerging Infections Surveillance and Response System, known as GEIS by many of you, has operated since 1997. It is now a part of the Armed Forces Health Surveillance Center. It provides a wonderful opportunity to synergize what we are seeing throughout the world in emerging infections, influenza being one of the priority pathogens, to synergize that with that healthcare utilization and our uniforms that are also throughout the world.
Nearly 500 locations this year, 75 countries globally represented by that partner network. As I said, we do share these sequences 100 percent with the CDC. Some of the laboratories send samples directly to them. Most of them work through the United States Air Force School of Aerospace Medicine in San Antonio to help process these.
Last influenza season, we processed over 81,000 because of the concern of pandemic H1N1. This year we have processed over 13,000 to subtype determination and a subset of those for actual molecular analyses.
This is the Defense Medical Surveillance System I said I would give you a graphic of. Quickly, just pointing out the very large number of individuals that contribute immunization records, important for this meeting, medical encounters, ICD9-focused, serums in our DoD serum repository -- I might touch on a study that recent events have led to the formation of -- reportable medical events, influenza in the DoD is one of those, and, again, the inpatient/outpatient encounters, all on a backbone of the demographics, again, a very robust system.
The laboratory network throughout the world is kind of concentrated through our overseas labs and in Europe in the military health system hospitals there. The United States focused on, out of two laboratories, United States Air Force School of Aerospace Medicine in San Antonio and the Naval Health Research Center in San Diego. USAFSAM does the sentinel surveillance throughout the world, often in collaboration with our overseas laboratories.
NHRC does a very robust system of surveillance of febrile respiratory infections in our recruits. They also do shipboard, Mexican border -- as everyone knows was instrumental in the first identification of the pandemic H1N1. That is, again, a very robust system from the standpoint of known populations, highly vaccinated, and, again, looking at those influenza breakthroughs in those populations is of interest to this group and we share robustly with the CDC.
This is a snapshot that we update weekly with regard to the number of samples processed throughout the world regionally, share with partners in the DoD throughout the world, and the distribution of subtypes. I will be showing this in a different form. Again, the major overseas laboratories are represented here and our military treatment centers throughout the world, as well as the U.S. reference centers.
This is just a quick look at an epidemiological week and the distribution of influenzas that we are seeing. This is just a quick look, as you saw from the CDC, to share with you on whether or not there is a large difference in recent influenzas that we are seeing. This is from the Global Network, not U.S. The Global Network, the orange and the red are a bit difficult to see, but the orange represents all As, of which you will see that in recent weeks almost all of them are composed of the H1, and 100 percent of those we have seen are the pandemic H1, and fewer contributions by Bs or H3s, although earlier in the season we did see quite a few of those, again, on a global scale.
From the standpoint of the U.S. as USAFSAM has reported, we are seeing the same thing as has already been reported to you. It is about one-third, one-third, one-third between pandemic H1, H3, and Bs. There are regional differences in virus circulation. You have seen some of this. I will show you quickly from our perspective.
European sites, a lot of H1. We are seeing a dramatic increase in pandemic H1N1 among service members, particularly in our recruits. I will share some of that data. Relatively low activity right now in African sites, but an upward trend of influenza in our service members since week 47.
This graphically demonstrates those samples of those 13,000 or so this season that have been sequenced -- this is the HA1 gene -- and the distribution of those chosen for sequencing around the world, giving emphasis to locations where there is perhaps under-representing in other places. Many of you may know we have a very active laboratory in Cairo, Egypt. They have sampled from this region, but they do share their samples directly with the CDC so it is not represented here.
Real quickly, again, because this information was included with what Dr. Cox shared, looking at the pandemic H1N1. What I would like to show here is this triangle because I am concentrating in my talk about individuals that were previously vaccinated and their influenza experiences. These red triangle, which you can pick out in this otherwise unreadable graphic to your right, demonstrates those viruses obtained from individuals previously vaccinated.
Some clusters right in here. This is a homologous sequence of recruits, as we first started seeing the H1N1 in recruits. Basically, what I want to point out is they are kind of spread out throughout. There is not any greater than three percent heterogeneity among those samples -- 97 to 100 percent homology with all of these samples.
Of these 23 that are depicted here as previously vaccinated individuals, twelve have been sent to the CDC. We do have results on four of those, showing that antigenically they are very similar to the vaccine strain. Despite these apparent infections in previously vaccinated, there is no clustering that I can point out to you, per se.
Looking at H3s, again the HA1 gene, about 82 percent of these in our global system were Victorian lineage. Again, you see the triangles here spread relatively throughout. There cannot be more than -- again 96.5-99.4 homology throughout this phylogenetic tree.
Lastly, quickly, our Bs, again, seeing some previously vaccinated individuals spread throughout. 99 percent of the Bs that we have acquired are with the Victorian lineage.
I am going to present now about six or seven different looks at vaccine effectiveness. The first comes from USAFSAM. I pointed out that they, in the DoD, function in the sentinel surveillance throughout the world. It is very robust in the U.S., but very robust throughout the world, as well.
This is looking at kind of the peak of the season, weeks 48-4. 132 individuals of influenza-like illness and looking at the vaccine status and laboratory confirmed influenza. Overall, vaccine effectiveness was right at 60 percent. I do not have this broken down -- I cannot break this down by the type of vaccine they received. As you will see, I will be able to do that with some of the others.
Another look at some of the USAFSAM data, this is over the whole season. This is concentrating only, out of about 1,200 that they captured as ILI, influenza-like illness, 66 percent of those they knew the vaccination status. Of those 66 percent, 541 they knew got the vaccine. The other 200 and some did not get a vaccine. They knew that vaccination status.
So if we look at these two two-by-two tables, knowing that all of these 541 received a vaccine, we look at all flu in this left one. These individuals received the LAIV. These individuals received the TIV. You see relatively equal distribution of all flu cases between the two influenza vaccine status. In this two-by-two table, you will look only at pandemic H1N1 infections. Those that received the LAIV, 24, TIV, 12, indicating that you were about 2.7 times more likely to have received the LAIV, if you had a pandemic H1N1 infection.
Looking at Naval Health Research Center data and some of their surveillance network and what they can do from the standpoint of vaccine effectiveness -- they have two difference systems, as I mentioned. The recruits is very robust, but this is looking at dependents and civilians in the southern California area.
Of 568 individuals, 427 had a known, unambiguous vaccine status, having received or not received and ,again, laboratory-confirmed diagnoses here. 158 of those were positive for influenza. 91 were Bs. 61 were H3s. And, unfortunately, in this area, only 6 were pandemic H1N1s.
The vaccine effectiveness, using the classic one minus odds ratio, showed for the B component a 61 percent vaccine effectiveness, for the H3, 64 percent vaccine effectiveness, and with those only few cases, but demonstrating essentially no coverage for the pandemic H1N1. Again, just another piece of a lot of different evidence that I am showing you here, from different ways of calculating.
The military recruits, again, a captive population, essentially 100 percent immunized and, for the purposes of this slide, 100 percent immunized with the LAIV. Two sites that did receive TIV were excluded from this analysis.
We had a very remarkable increase in infections with pandemic H1N1 in recent weeks in our recruits. That is very unusual for us. Highly vaccinated, we do see influenza in that highly vaccinated population. We take those seriously, in case they are breakthroughs to see if there is drift of the circulating strains from the vaccine. But this is very unusual to see so many of late.
Of these 200 febrile respiratory infections -- we call that within the DoD, it is the same definition as the ILI -- enrolled between December and February, 45 percent were positive for influenza. Again, we usually see onesies and twosies at these sites of influenza laboratory-confirmed cases. 67 of these were H1, 14 H3, 3 B and some co-infections.
Overall, the vaccine effectiveness is 81 percent. I am going to quickly say do not get overly -- we know that is high when it comes to influenza vaccine effectiveness calculations and there is a variety of reasons for that. 100 percent vaccine coverage in this population, as well as the fact that infections generally tend to hit in the first couple of weeks after they get to camp, if they get infected -- so that is going to give you a bias towards a high vaccine effectiveness because, again, like other definitions, they are not covered by vaccine until 14 days after vaccination.
Again, a classic vaccine effectiveness calculation, but I want you to compare this with other years. That is what is important. Because of the bias in this population, what is important is to compare this one calculation year after year in that population with the same laboratory methods, same captive population, same immunization status of 100 percent, and see when you have outliers of years.
This year 81 percent seems a bit lower than usual, but not too much different. Last year, 54 percent. What that 54 percent represents is the vaccine effectiveness of last year's trivalent against the pandemic H1N1 because that is what was circulating in our recruits exclusively, like the nation. So 54 percent last season is the vaccine effectiveness for the heterologous trivalent with the pandemic H1N1 in circulation.
This year, overall, 81 percent -- but here is what I am wanting to get to -- for the pandemic H1N1 in these populations, of which there are robust numbers of 41 percent vaccine effectiveness.
So let us look at this in a little bit different way among our military recruits again. What we have here is divided up by subtype. We are comparing this to the CDC region four south east United States as far as the distribution of influenzas seen in that region. As I have stated and as you have heard from others, that distribution is basically one-third, one-third, one-third, with the three different subtypes of influenza.
What is remarkable here is -- with these influenzas -- a very high percentage of the H1N1 among our young recruits. Yes, the young are afflicted with infections a little bit more than other populations, but we have very tight visibility to this in the DoD and it certainly is not to this extent. There is a five to ten percent difference in rates between different age groups within our DoD surveillance. So this is quite profound. Within the H3 and B components, a smaller percentage contribution to the influenzas being seen. So, again, another piece of evidence that there seems to be something less protective with the TAIV H1N1.
The other thing I want to point out here is normally we see the median for the infections with influenza in our recruits in the first couple of weeks, as you see here, more typically. This is very atypical. What this is essentially saying is the infection with pandemic H1N1 is evenly distributed across the recruit trainers.
They arrive in a staggered manner and they stay for 12 weeks, 14 weeks -- it changes. It is a little bit different for different services. But you are seeing a much different median week of infection than we typically see with influenza in our recruits.
Now looking with the Armed Forces Health Surveillance Center and Defense Medical Surveillance System that I showed you a diagram of the healthcare encounters, this is a methodology that was published in PLoS One last year. It is a case control analysis of our reportable medical events. These are laboratory-confirmed influenzas, as well, within the Department of Defense.
What you are seeing here is, among those individuals that received the TIV, around a 60 percent vaccine effectiveness. But among those that received the LAIV, again this is across all age groups, about a 30 percent vaccine effectiveness. This is not broken down by subtypes, just the vaccine.
Another look, also published in JAMA in 2009, a different methodology. This is looking at ICD9 codes based on inpatient and outpatient visits, ambulatory and hospitalized. It is looking at all pneumonias not otherwise specified or an ICD9 clinician that gave an ICD9 of influenza as the cause of that visit.
What you see here, again, broken down by vaccine received, LAIV, is essentially an incidence rate ratio of one or adjusted with confidence intervals around one, where there is no difference in whether or not you got influenza -- no evidence of coverage. The TIV has some amount of coverage there represented. When you compare the two it is about 25 percent more effective than the LAIV.
In summary, the DoD does maintain a global laboratory-based influenza surveillance program made quite robust in 2006 with additional funding. We also have the advantage of leveraging that laboratory network with healthcare utilization throughout all of our uniformed members in a population that is very well known with good denominator data.
Molecular analysis indicates good agreement between the current vaccine components and the circulating influenza strain. You have heard a lot of this. Our concern with some of these different looks at the effectiveness of the vaccine is that there appears to be poor protection against the pandemic H1N1, particularly with the LAIV.
With that, I thank you. I thank the many different partners that have contributed to this. I think it has really been phenomenal, their cooperation on DoD-wide level of giving this data and providing this data. Thanks to all of them.
DR. ROMERO: Thank you very much. Dr. Gray?
DR. GRAY: That is tremendously interesting data, Dr. Russell. I seem to recall from one of these papers that the explanation for the difference in the LAIV versus TIV was the repetitive nature of the vaccinations. In other words, among the more seasoned personnel, who had received a number of vaccines and they did not respond to the LAIV because of pre-existing immunity. But now you have a situation where you are seeing a problem when they very first enter the military. I think that is a little troubling.
DR. RUSSELL: By and large, I think you can say these are influenza vaccine naïve individuals. That is what is troubling to us. I will point out that we had an influenza death of a recruit just at the end of this last week, who received the LAIV five weeks prior. Circumstantial. It does happen, but it is very unfortunate.
DR. ROMERO: Any other questions? Dr. Levandowski?
DR. LEVANDOWSKI: Going back to the Influenza B information on epidemiology, you mentioned that seven percent of the strains that had been isolated were B/Yamagata-like. Was there any geographic clustering to those that you could tell us about?
DR. RUSSELL: I am going to read the bullets here, myself. I did not give a lot of time to this. Can anyone from USAFSAM comment on that?
DR. CONNORS: My name is Benjamin Connors. I am with USAFSAM. I am a molecular biologist. For the Bs, we got a great deal of cooperation from some of our other partners. A good percentage of our Yamagatas were from AFRAMS, which is based in Thailand. A great deal of them were early season in Nepal and Bhutan. We got a few from our partners in Peru. We had only seen a single case in the U.S. up to this point, as well as one in one of our deployed soldiers in Afghanistan. That is the total of our Yamagatas throughout the season. Most of them did come from other parts of the world and very early -- July and August. That is when we found those.
DR. LEVANDOWSKI: Could I just ask a follow-up to that then? In terms of the Victoria lineage, have those replaced the Yamagata lineage in areas that you were originally finding Yamagata-like strains?
DR. CONNORS: There were quite a few more Victoria across the board. There were probably -- I think it was around 93 percent Victoria overall. In the United States and the northern hemisphere, where it is our normal one, it was more around 97 percent of the ones we sequenced -- North America and most of our normal military sites, as opposed to one of our sentinel sites in Asia and South America.
DR. LEVANDOWSKI: But the question I guess I am trying to get at is whether subsequent to the early season Yamagata-like strains in areas that AFRAMS was getting their isolates from, were those then replaced by Victoria-like strains later on?
DR. CONNORS: Yes. There were quite a few more throughout the season with the Victoria strains.
DR. RUSSELL: Again, what he is saying, I think, is there is no evidence that the Yamagata has out-competed and has taken over the Victoria lineage in any place.
DR. CONNORS: Correct. None that we have seen.
DR. ROMERO: Any other comments or questions?
DR. RUSSELL: That is Benjamin Connors from USAFSAM. They did a lot of this work so thanks to him. Thank you very much.
DR. ROMERO: Thank you. We are on time. We are scheduled for a twenty minute break here so we will go ahead and do that. Sorry, fifteen minute break. Excuse me. We will reconvene at 10:35 AM, please. Thank you.
(Break.)
DR. ROMERO: Why don't members of the Committee take their seats please?
DR. JEHN: I would just like to make a brief announcement. If those who intend to speak that have registered with me for the open public hearing, if they have any slides, if they could please see the folks at the AV table and have those loaded now so we do not have a disruption of the meeting? Thank you.
DR. ROMERO: Thank you. Okay, we will proceed. Next, Vaccine Responses by Dr. Zhiping Ye from the FDA, please.

Back to top

Agenda Item: Vaccine Response
DR. YE: The antigenic properties of the representative viruses mentioned by Nancy was confirmed and tested by using human sera, rather than the ferret sera to analyze their antigenic properties.
In this study, the recent H1, H3 and B isolates were analyzed by using 12 human sera panels from four or five countries. Each was immunized with H1N1, H3N2 and B using 2010 and 2011 formulations.
The sera was analyzed for the presence of antibodies to the HA antigens of recent virus isolates. The method was using HI assay to the recent viruses and then to compare with the vaccine strain. A subset of the sera was analyzed by using micro-neutralization, which was already mentioned by Dr. Cox.
This slide shows the sera panels from the five countries. Most of them contain the sera from adults, elderly, and two children sera come from China and the U.S., as mentioned by Nancy in her presentation. As here shows, the vaccine strain is using California/07. H3N2 is Victoria-like, which is Victoria/210, which is Perth/16-like virus and B/Brisbane/60. All the panels were immunized with a similar vaccine strain. By the way, for the children from China, they were five to three -- three to five years old, from the U.S., they were six to 24 month old.
Now we start from H1N1. I think Nancy already showed this slide. I think for the serology studies it is quite tricky to pick up which strain is going to be used in human serology studies. We tried to use the virus that is more representative to the virus that was mentioned by Nancy in the surveillance studies.
The virus, basically we chose the sum of the virus in the center, close to California/07 -- some outliers just to represent the overall virus and analyze the antigenic properties.
Right here I just spend a little bit of time on exactly how we choose. For the vaccine strain, we have to use the positive control, which is California/07. There are two kinds of the virus that we use for serology studies. One is the wild type. One is the reassortant.
I think that Nancy did not have time to illustrate -- mention the reaction of the reassortant to the ferret sera versus the wild type. The reassortant usually gives you a relatively higher HI titer versus to the wild type virus. For the representing virus, we are choosing not only geographically, but also represent the mutations or substituting mutations to represent the virus -- the mutation on HA.
One I wanted to point out is the Kentucky/9/2010. There is a severe mutation in amino acid 155. Here it is right in the critical region of 155-158. This is outliers and the rest of them represent the virus in the center or close to California/07. Also, we used the former seasonal H1N1 as a control in that study.
This slide shows the study from five laboratories, including the lab in CBER and also CDC Australia, UK, Japan, and China. This is the combination of all the studies. The Y bar shows the percentage of the GMT titer, I think as Nancy already mentioned, and X bar shows the viruses that we used for serology studies.
The tricky one is that only you can use a few viruses in this study. So we only can choose a few viruses, but among the labs we try to share which virus would be common and which one is representative to the different countries. Here you can see that. In this, the green bar shows adults. The blue bar shows the elderly. The purple bar shows the children.
Here, is wild type. You can see if we set wild type as 100 percent, now you can see that the HI titer from reassortant 179A or 181 have a relatively higher titer, as I mentioned. However, the rest of the viruses, the titer from those testing viruses compare with the wild type. Now you can see that 50 percent of the GMT, as mentioned in Nancy's presentation, and if above 50 percent we consider that similar to the vaccine strain and lower than 50 percent we consider that one not covered well by the vaccine strain.
Here you can see the majority of the viruses covered pretty well by the sera immunized with California/07. Only a few of them had a relatively lower titer. One thing I have to point out, the Kentucky -- the low reactors are covered pretty well by three populations, which are adult, elderly, and children. Also, you can see that for former H1N1, it is way below 50 percent. However, you still have some residual antibodies against the former H1N1.
In this, the black bar shows the overall GMT for adults, elderly, and children. I think it is covered very well by the vaccine strain to California/07.
Now we go on to the H3N2. Here, I think Nancy showed this one, too. It is the former H1N1 -- this is the pandemic H1N1 -- and here is two genetic clades. One represents Victoria/208 and one is Perth/16, but antigenically they are overlapped. The virus -- the majority of the viruses are chosen in the green and some in the red.
This, again, is the vaccine strain. It is a wild type and reassortant, which is used for manufacturing of the vaccines. It is also representative of viruses. Here, I just highlight some amino acid substitutions of those viruses. I wanted to point out some low reactors, which have an asterisk and also some MDCK cell isolates. I think Nancy also mentioned that because the trouble that depends upon what kind of cell you use and what kind of red blood cell you use, especially for H3N2. It is analyzed in this study.
Here shows the 100 percent of the control, which is the vaccine-like antigen. Compare that with the rest of the testing viruses. Here you can see that the majority of the viruses covered very well in the population of adults, elderly, as well as in children, except some of the virus -- that is three viruses cell isolates. They did not cover well by the adults and the elderly. These three viruses were not tested in children.
However, the response in the overall study among adults, elderly, and children, you can see that the majority of the viruses covered very well by the vaccine strain against the Perth/16-like virus.
Now, to go on to the B strains, this slide shows the antigenic cartography of the Victoria-like. As Nancy mentioned, the current strain is Brisbane/60. This is the previous vaccine strain. Here is the outliers, which is B/Singapore/616/2008. That virus was also included in serology studies.
This slide shows the Yamagata. This is the former Florida/4 virus and this shows the recent -- the viruses represent the current Yamagata virus and also mentioned by Dr. Cox.
Here shows that the vaccine strain used is B/Brisbane/60. We used the wild type of the virus, itself. This blue shows the Victoria-like. We tried to cover the viruses by choosing clade six, clade one viruses in this serology study. This one shows the Singapore/616 is the outliers and a low reactor. That is being used in this study, as well. This red one shows the Yamagata lineage. Here is the Florida/4 and also Wisconsin/1, which is similar to the Singapore -- no, I'm sorry this is representing the recent viruses.
In the B strain, we include B/Victoria in the left panel and also B/Yamagata boxed with the red. Here you can see that, regardless of the population, all of them except one covered very well by the vaccine strain against the B/Brisbane/60.
The antigen against using Yamagata-like is different from population to population. For the adults, you can see that some of them still have a reasonable antibody against those Victoria-like viruses. Some have lower. Some of them covered very well by the vaccine against B/Brisbane/60 in adults and the elderly population. However, in children, all of them did not react well to the vaccine strain against the B/Brisbane/60. For the Yamagata in the children, they do not have a good cross-reaction between Yamagata and Victoria.
Here, the black bar shows the antibody against the overall summary of the population against the Brisbane/60. You can see that covered very well across the different populations. However, for the Yamagata lineage, only adults and the elderly covered reasonably well. Children are not reacting well to the antibody against B/Brisbane/60.
With that, this is the conclusion. As I mentioned, for H1N1, representative recent pandemic A(H1N1) virus was covered well by the vaccine containing A/California/07-like virus. For H3N2, representative recent H3N2 virus was covered reasonably well by the vaccine containing Perth/16/2009-like vaccines. For B, in terms of B/Victoria lineage, the representative recent viruses were well covered by the current vaccine. However, for the B/Yamagata, it covered less well, especially in children population. Thanks.
DR. ROMERO: Thank you. Any comments or questions? None? Thank you. Next, we will hear from Dr. William McCormick from the FDA, talking about availability of vaccine viruses and reagents.

Back to top

Agenda Item: Availability of Vaccine Viruses and Reagents
DR. MCCORMICK: Good morning and thank you. I would like to, first of all, explain who I represent. As the Division of Product Quality, we are CBER's product testing group. We do the lot release testing for the flu vaccines. We also function as part of the ERL or Essential Regulatory Laboratories that are coordinated by the WHO. We provide coordination and calibration of the potency reagents with TGA in Australia, with NIID in Japan, and with NIBSC in the UK.
These are the players. There is quite a bit of history. You know these have been around for a little while. The job that I have for you today is to tell you about the virus strains that are available to make these vaccines and to discuss and give you an update on the availability of potency reagents for these particular strains.
If we start with the H1 candidate vaccine viruses, we have a number of reassortants that have been made and have been pressed into play to manufacture vaccine: the X-179A from New York Medical College, the X-181 from New York Medical Colleges, the 121-xp, which is a reverse genetic strain from NIBSC. These three strains have been in vaccine in the past season.
The NIB-74 is a new strain from NIBSC. It is a classical reassortant. I believe it is from at least three different viruses that make up the genetic components of this virus. The HA is from A/Christchurch/16/2010.
These are the candidate viruses for making vaccine. From CBER's perspective, the reagents that are available for testing these viruses include a lot of reference antigen manufactured from X-181 -- from the HA from X-181, a Lot 69 from X-179A reassortant. We had an experience where we were required to prepare reagents for 121xp. We were able to recalibrate Lot 69 and have it perform effectively for 121xp. If manufacturers should select the NIB-74, then we will make a decision about manufacturing reagents for that.
The antisera for this -- I know the whole world is very grateful for the job that NIBSC did to prepare antisera initially. CBER now has antisera, has several sheep that are producing antisera, and it works very well. We now will be supplying antisera for this testing.
For the H3, there are two classical reassortants that have been pressed into service to make vaccine: X-183 and X-187. The reagents for these are derived from the HA of X-187. We have, again, two lots. The antisera for this was prepared from X-183. We have several sheep that are producing for this so we have a significant quantity of the antisera available for the H3.
For the B Candidate viruses, of course the wild type was used. There are several classical reassortants now that are available and are being evaluated. It is possible that the BX-35 may make it into some vaccines. If it does, we will make a decision about preparing reagents.
We do have reagents for the B based on the B/Brisbane/60 -- several lots. Our antisera is prepared from the egg adapted wild type. We have several lots available of that. As you can see, some of the lots of antigen are getting a bit low. What we will do is we will make sure we replace those lots and bridge them back to the original lot.
This is kind of a summary slide that just reminds everyone of the rules for use of potency reagents at CBER. Only CBER authorized reagents shall be used to test potency of vaccines marketed in the U.S. The authorized reagents will be either produced by CBER or adopted by CBER for use. CBER will collaborate in the calibration of any reagents that we do adopt.
CBER will verify the availability and acceptability and performance of compatible authorized reagents with each manufacturer's vaccine product. We do that at the beginning of each production campaign. CBER will make every effort to assure the availability of reagents appropriate for all strains selected for production of vaccine. That is really the bottom line message that I want you to take home -- is that we will do the best we can and scramble to prepare any reagents that are necessary for measuring potency. With that, I conclude. Thank you.
DR. ROMERO: Thank you, Dr. McCormick. Any questions? Go ahead.
DR. LEVANDOWSKI: Could you comment on availability of reagents for the B/Yamagata lineage strands? That was not included in your slides.
DR. MCCORMICK: We have not made any decision yet on that. When pressed for that, we will scramble and we will prepare whatever is necessary.
DR. GELLIN: I couldn't help, when you brought up the sheep, to think about -- last year the PCAST Report that looked at influenza vaccine technology improvements. Among the things to highlight, not only for pandemic preparedness, but for seasonal, were things that could shorten timelines. Given where we are this year, that is not as necessary, although it may be for Yamagata, as you just mentioned. But can you give us a preview of coming attractions of how some of these timelines might be shortened in coming years, both in this portion of it and also sterility testing?
DR. MCCORMICK: I could try to do that. That is not something that is really close to what I am doing. My job is to follow the tried and true steps and to get everything in place as quickly as possible. I defer to several other people in the room to do a better job of that. I can tell you that in light of the potential for changes in strains or for the Yamagata, we already have several sheep standing by. From a side angle, that addresses your question.
There is a quarantine period. We have already passed that quarantine period, but I know that does not get to the heart of your question. There are other ways being investigated of raising antigen that are cell-based or recombinant that do speed up the process of getting the antigen and then providing that to my people, who can put it into sheep. I still have to have sheep to be able to do that.
DR. LEVANDOWSKI: I just want to follow up further on the B/Yamagata aspect. Are there any strains that have been identified as potentially useful or appropriate for use in vaccine preparation for the B/Yamagata lineage viruses?
DR. MCCORMICK: I have been keeping an eye out. I think that what I am seeing here is the same thing that -- I have no more information than what you have, I think is what I am trying to say.
DR. COX: I think that we had provided to Dr. Doris Booker at New York Medical College the Wisconsin/1/2010 wild type, the Bangladesh/5945/2009 wild type, and the Hubei Wujiagang wild type. All of those are in that new antigenic group of B/Yamagata lineage viruses. I think Doris has made some reassortants. Those are being analyzed to make sure that they have not incorporated any antigenic changes during the laboratory manipulation.
Those wild type strains, I believe, had been distributed to vaccine manufacturers. If not, certainly they can be, but I am pretty sure they have. Then the potential high growth reassortants -- the BX-41, 37, 49, et cetera -- will be distributed, if they are appropriate.
DR. ROMERO: Any other questions or comments? Thank you very much.
DR. MCCORMICK: Thank you.
DR. ROMERO: Next, we will move on to comments from manufacturers, Dr. Samson Lee from Sanofi Pasteur.

Back to top

Agenda Item: Comments from Manufacturers
DR. LEE: Thanks. On behalf of all the manufacturers, I just want to thank you for the opportunity to share our perspective on the influenza vaccine season. I hope today to share with you three things: some feedback of the current season, from the 2010-2011 season, help walk you through the manufacturing process to help you understand the timing of strain selection and how that impacts the availability of vaccine, and then also give you some late breaking information about what the current status is for manufacturing.
By all accounts, 2010-2011 season has gone very well. At the beginning of the season, a manufacturing timeframe, it was a little bit uncertain because we were still responding to the 2009 pandemic. Most of the manufacturers will continue to produce that monovalent vaccine, but then as that was understood of how much of the pandemic vaccine was needed, we were able to then transition rapidly into the 2010-2011 vaccine.
The other complication was that the H3 component, even though that was carried forward, in terms of productivity, it was relatively moderate compared to past years. H3 is usually the highest yielding strain, but this year it was relatively moderate.
Two other complications later in the season. One was some late information regarding the tip caps for the syringes that for some manufacturers delayed approval and release of the first doses. A second complication, later, was also for some manufacturers that the potency, as measured by SRID, was inconsistent, depending on the source of the reagents that they were using for that assay.
Even with all of these complications, in the end, it was a very successful year. I think it was a combination of factors. One, that we were able to certainly carry over the H1N1 component because the pandemic strain was carried into the seasonal vaccine. A second is that the timing of the strain selection was made in a timely fashion by this Committee. That allowed manufacturers to respond and produce very high numbers of vaccines, despite the challenges.
In this next chart, I would like to show you how we have improved both supply and immunization over the past several years. Certainly, manufacturers are the ones providing vaccines in large quantities that allow then the public health sector to push in terms of immunizations.
If you look at the last several years, you can that in 2009 -- 2009 was the year where manufacturers also had to produce the H1N1 pandemic vaccine so those numbers were relatively flat. This chart only shows the trivalent seasonal vaccine. If you look at 2009, it does not show the full number of doses that were produced by manufacturers for that year.
2010 there was a significant growth in terms of vaccines that were produced and distributed, in combination with recommendations by ACIP for universal -- their universal recommendation for vaccinations and other campaigns, there certainly was a bit increase in terms of immunizations. The increase was across the board in all age groups.
I think for 2011, we are well poised to make big gains in terms of both vaccine supply, but also in terms of immunization growth. I think that immunization growth is important to allow manufacturers to sustain that high level of production capability for the seasonal vaccine.
I would like to move next then into the annual vaccine manufacturing timeline. This is a typical timeline, not suited for any specific year. So there may be small adjustments on a year to year basis, but this gives you an understanding of the different linkages between what happens, what the critical milestones are, and how it affects manufacturing.
As you know, surveillance occurs around the year. Usually, around the December/January timeframe, you start to get a picture of what the prevalent -- at least one of the strains that may be carried over into the next year. Manufacturers will look at that data that is provided by CDC and others and make a decision about a strain that might be used for production at risk.
At this meeting, the various data are considered. Typically, at least two of those strains can be selected here. In the past, the third strain sometimes is deferred. If at all possible, usually I know the Committee tries to make a selection for all three strains.
Based on that decision, manufacturers continue production. Then we will either switch to a second strain. If needed, if there is a new strain, for either the second or third, produce a working seed for that third strain and then at an appropriate time, switch to production of that third strain.
After a period that the strains are fixed, then manufacturers would also need to balance the number of doses that are produced of each strain. Each strain is produced independently and then, in the end, we need to balance the three different strains to formulate into the finished vaccine.
The other key aspect to this is production and standardization of the potency assay reagents. Depending on when the third strain -- or second or third strains are identified, receipt of those reagents could occur as late as May or early June.
That long leeway time primarily is for the production of the sera with the antibodies in the sheep and not necessarily the reference antigens, themselves. Producing that purified HA right from the start is the key point to start off the process. That process is relatively efficient, but we still need that reference strain in order to start the production of that purified HA and immunize the sheep.
Once the reagents are available, the manufacturers then can use that to finalize the potencies of the material that has been produced. That then allows the start of formulation. Now we can actually formulate the vaccine at the appropriate dosages.
After formulation, we proceed into filling and packaging. There is a key component here, which is the next milestone, and that is the timing of the annual license approval. This is important because it finalizes the packaging information -- the labels, the package inserts, all of the various components that go into that. While manufacturers can formulate and begin filling of the vials and syringes, they require that annual license approval to finalize the packaging and then can proceed into that packaging.
Finally, then, manufacturers move into distribution, which usually occurs in August/September. Vaccinations occur starting in the late fall and in through the winter.
I think, as you can see, overall it is a process that really requires a collaborative effort. I think that is probably the main message here today. It has been demonstrated year after year to work very effectively and it is a collaborative effort.
I would like to just highlight the two critical factors that Dr. Weir had mentioned earlier. That is the timing of the strain selection and availability of potency reagents that are critical to keep the process moving.
The timing of strain selection certainly helps ensure the timely availability of the vaccines. Understanding that it requires consideration of a lot of different factors, including the strain surveillance, the epidemiology, the timing of the selection, and the impact to vaccine availability. The later the decision of finalizing the third strain means the timing of manufacture is reduced and the total quantity of vaccine could potentially be impacted.
The timing of the strain selection also influences the final development of the working seed, as well as the reagents. In the flipside, if the decision could be made earlier, certainly that gives manufacturers more time to respond and work through the constraints that might be imposed by a low producing strain.
In terms of availability of potency reagents, I think you have seen in that chart, really it is a timing to produce any new reagents for any new strains that come out. That length typically takes about 8-12 weeks from the timing of availability of that strain.
I would like to then move on to the current manufacturing status. Many manufacturers have started production "at risk". The strains that have been started were the H1N1. The viruses that are used for production are the A/California/7/2009 X-179A, as well as the X-181. There is consideration of the new reassortment, the A/Christchurch/16/2010 made at NIBSC, the NIB-74. For the H3N2, most manufacturers are working with the A/Victoria/210/09, the X-187 reassortant.
Moving on then to the B strain, I know that at least one manufacturer has begun production of the B strain. That is working with the B/Victoria lineage, working with the B/Brisbane/60/2008. As you have heard, there is a reassortant produced at New York Medical College, the BX-35. Reagents are available for both of those viruses -- those strains.
For the B/Yamagata lineage, as you know, several manufacturers are working on a quadrivalent vaccine that includes the two A strains, as well as B strains from both lineages. The candidate virus that is being looked at right now and worked with by manufacturers is the B/Hubei-Wujiagang/158/2009 BX-39. Currently, there are no reagents available for this particular strain, as you heard in the previous talk. Having access to those reagents would be important to allow us to prepare a quadrivalent vaccine for clinical trial later this year.
I would like to summarize, again, just appreciate the opportunity for sharing with you today, but also to emphasize that this is a shared responsibility in the on-going collaboration that has been successful over the years to provide influenza vaccines each year.
I want to just emphasize the four components that have made that successful. One is certainly communication. Certainly being able to share today, but on-going throughout the year. Timely strain selection, as we have done each year and requiring the balance consideration of various factors.
Virus availability -- having access to the wild type viruses, but also the reference viruses, gives manufacturers the opportunity to evaluate the growth characteristics. This has occurred each year and has not been an issue, but I just do highlight that that is one aspect, as well.
Certainly, the annual vaccine approval and release process. The earlier the annual license can be approved, certainly that allows packaging to occur that much earlier and availability of the vaccine to the market that much earlier, as well.
The other point I would like to highlight here is that the in-season lot review -- so as each lot is produced and made available to CBER to review and release -- that on-going process is important to have as timely as possible as well. During the 2009 pandemic, the e-lot release process was launched and tested during that period. During the last year, 2010-2009 season, it was continued very successful. So I just wanted to commend CBER for that timely review and release of the annual lots.
I hope that gives you a good overview of manufacturers' perspective and I am happy to answer any questions.
DR. ROMERO: Any questions?
DR. EICHOFF: When -- I guess it is when rather than if a four-valent vaccine is ultimately approved by the FDA, is there -- I assume this will be a 60 microgram vaccine, 15 of each antigen -- is there a sufficient manufacturing capability to affect that one-third increase that is required?
DR. LEE: Well, you saw the big increases in recent years in terms of manufacturing capacity. Certainly that is anticipating some of the build up and investments that manufacturers have made in order to increase capacity. It will affect, certainly, because now instead of producing three strains, you are producing four strains within the same timeframe, but certainly manufacturers expect to be able to be able to produce the vaccines that are needed for a quadrivalent.
DR. GELLIN: Let me just add to that or build on that one. Sam, thank you for this presentation. Your slide that talked about the 19 percent overall growth, it is striking the amount of an increase in manufacturing by the many manufacturers between last year and this year. I know that reflects multiple decisions by multiple companies, but you are in a spot of representing industry so how do you give us some insights into the decisions that were made to increase the vaccine production for this year by about that 30 percent, it looks like.
DR. LEE: Certainly, those decisions needed to be made many years prior, working with growth within the U.S., as well as outside the U.S. In terms of increasing capacities, the big increase that you saw there between 2009 and 2010 was because in 2009 a big portion of that was used for making the pandemic vaccine. Even though 2009 it looks like there was a big step increase, if it was all trivalent seasonal vaccine, you would have seen a more gradual increase between 2008 to 2009 to 2010.
DR. GELLIN: If I could ask a second question, you may have answered my question in your comment about the quadrivalent strain. You said manufacturers are developing this strain. Then you said they were developing it for clinical trials later this year. Does that mean that you need advice from a Committee like this to figure out what you would need even for a clinical trial?
DR. LEE: Yes, that would be helpful. You are correct. Having understanding of what the appropriate strain to include in a quadrivalent for the fourth strain would be important.
DR. ROMERO: We will be sure to include that in the discussion. Dr. Debold?
DR. DEBOLD: This slide, again, that Dr. Gellin made reference to on the immunization growth, that, again, is just TIV? If we were to also be looking at LAIV, how might these numbers change? Can you give us some sort of an idea?
DR. LEE: I think this particular chart is total influenza vaccine and not broken down between inactivated and the LAIV. That is my understanding.
DR. ROMERO: Any other questions? Dr. Schoolnik?
DR. SCHOOLNIK: What is the optimal manufacturing capacity in terms of number of doses on an annual basis for the United States?
DR. LEE: I am not sure I am in the position to be able to answer that. Certainly, manufacturers make investment decisions based on the market needs. As the immunization rates go up, certainly manufacturers would then invest, as well. Maintaining high immunization rates certainly will help manufacturers then invest in maintaining capacities at high levels. Certainly, from a pandemic preparedness perspective, having high manufacturing capacity is helpful because you can then switch that over very rapidly to produce large quantities of pandemic vaccine.
DR. ROMERO: Any other questions or comments? Please, Dr. Wharton.
DR. WHARTON: Just a question about the quadrivalent vaccine again. I think seeing the information that was presented here this morning, it is making me think once again about a quadrivalent vaccine and how nice it would be to not have to make the guess about what the B strains are going to do, since we are not actually able to do that with any particularly good reliability.
I understand I am asking an unanswerable question to some degree because it requires guessing about a lot of things that are not certain, but if the planets were appropriately aligned, would it be possible to have a licensed quadrivalent vaccine for the upcoming influenza season for at least some part of the population?
DR. LEE: I am probably not well educated to answer that particular question because I am not sure about the progress of all the clinical trials from the different manufacturers. Perhaps the manufacturers could answer for themselves on the progress of their particular trials.
DR. GREENBERG: David Greenberg, Sanofi Pasteur. Our phase three clinical trials are ongoing. Our projection would be to have a quadrivalent vaccine available for the 2013 season.
DR. COELINGH: Kathleen Coelingh from MedImmune Astra-Zeneca. We have completed phase three trials in children and in adults, comparing trivalent safety and immunogenicity to that of the quadrivalent vaccine. Those studies are completed.
DR. RENNELS: GSK is also developing quadrivalent vaccine in both its Quebec and Dresden plans.
DR. MCINNES: I am wondering if any of the manufacturers would like to comment on increasing the amount of antigen per strain that they are putting into their vaccine. Are there any comments on that, potentially? Not by multiplying by four as opposed to three, but micrograms of antigen per H1, H3 --
DR. ROMERO: So per strain.
DR. GREENBERG: David Greenberg, Sanofi Pasteur. Pamela, I do not know if you are asking about specific age groups or specifically for what purpose. We have our high dose Fluzone vaccine, which contains 60 micrograms of hemagglutinin per strain for the three strains. That is where we have emphasized the higher dose for the population 65 years of age and over.
DR. RENNELS: Just speaking for GSK, our approach has been a different one. That is to use adjuvant to try to reduce the amount of antigen that is needed.
DR. BAYLOR: I wanted to make a clarification on a comment that was made a bit earlier in response to Dr. Gellin's question concerning the recommendation for a fourth strain or comments. In order to do a clinical trial, I think it is obvious from the comments that the manufacturers just made that some clinical trials are ongoing and a recommendation has not been made for a fourth strain. Just to clarify that, that would not be required to do clinical studies.
DR. ROMERO: Thank you for that clarification. Any other comments or questions? Very good. Thank you very much. Now we are going to return to Nancy Cox and ask her to present the update on H5N1 and H9N2 surveillance and virus characterization.

Back to top

Agenda Item: Update on H5N1 and H9N2 Surveillance and Virus Characterization
DR. COX: Thanks very much. This is really an informational talk. I think that it is very important for this Committee to understand that we are really moving ahead very aggressively with pandemic preparedness, in spite of the fact that we had a pandemic not too long ago. We know we cannot rest on our laurels and we have to really be on our toes in terms of detecting novel influenza virus infections of humans.
First, I am going to talk about swine origin influenza A virus infections of humans from 2009-2011. If we look globally and look at what has been reported by countries to WHO, there are seven swine triple reassortant H3N2 viruses that have been detected in the United States, human infections that have been detected in the United States, two human infections caused by Eurasian avian-like H1N1 viruses of swine origin in Switzerland, and one in China.
Because we detected quite a number of swine origin H3N2 human infections in the U.S., I thought it would be useful for me to just provide some summary information. As you can see, the majority of the seven infections did occur in children. We have noted this in the past for swine influenza infections. Generally speaking, children are more susceptible or children come to seek medical attention more frequently. We really do not know for sure.
That having been said, two of the infections occurred in adults. Case number five, a case in Pennsylvania, an individual with no direct exposure, but who lived adjacent to a farm where there were pigs. Then case number seven, an adult who visited a live animal market in Minnesota. That happened to be the same live animal market that the child of case number three visited in Minnesota.
As you can see, since September, we have a total of five cases that have been detected. We are becoming increasingly concerned about this increase in detection. We did not know if it was an artifact of the fact that there is a different algorithm for doing real-time PCR, which would allow the state health departments to pick this up, or whether there is an actual increase in the number of human infections.
Just to remind you, the reason we call these triple reassortants swine origin influenza viruses is that the viruses that are resident in pigs have a combination of genes originating from classical swine influenza viruses, the North American lineage, avian North American lineage, and seasonal human H3N2 lineage viruses.
When we did a hemagglutination inhibition test with the recent swine origin triple reassortant viruses from the human cases, what we found is that they were very well inhibited by antisera to the swine origin viruses that had been isolated previously. But they really were not very well inhibited by antisera to a variety of previous vaccine viruses.
They were better inhibited -- at least several of the viruses were better inhibited by antisera to human influenza strains, which had circulated between 1989 and 1992. This is a bit of a surprise to us because we had thought previously that these H3N2 viruses, the hemagglutination and neuraminidases had entered the pig population in about 1996 or 1997.
So we truly expected these swine origin viruses to be better inhibited by sera to these viruses, Wuhan/95 and Sydney/97, but there was very little inhibition. That was of concern to us because we figured that would indicate that we would probably have a high degree of susceptibility within the population as there would not be that much antibody.
If we look at the evolutionary trees of the hemagglutinins of these viruses that were isolate from humans, they are shown in the color red for the recent viruses and then the other human isolates in green, you can see that they are scattered throughout this part of the tree. These are H3 seasonal influenza viruses. These are the older strains and these are the newer strains. But you can see there is a large genetic difference. All of these are amino acid changes located along these fairly long branches of the evolutionary tree.
The most recent viruses have tended to cluster up here. There is a Pennsylvania virus down here. We can understand why there might be antigenic differences among these swine origin influenza viruses, although they are not great.
The neuraminidases showed a somewhat similar pattern with one exception. That is for this Minnesota virus. You note this very long branch on the evolutionary tree with all of these amino acid changes on it. This was sort of a red flag to us because whenever we see there is a great deal of genetic distance, including a lot of amino acid changes, we wonder where this neuraminidase came from.
We immediately contacted our agricultural colleagues in Minnesota, actually at the University of Minnesota where they are doing some swine influenza surveillance, and they have been able to provide us some sequences of N2 neuraminidases from viruses isolated from swine. That helps fill in this gap, but it does not completely close the gap.
Because we saw these differences and because there was some evidence for illness in family members in a couple investigations where we had detected these virus infections, we decided that we would try to do serology for family members if there had been influenza-like illness at the same time as the index case was ill. Often times these studies are done retrospectively and it really can be quite difficult.
We also decided that it was necessary for us to do transmissibility in the ferret model to determine if these viruses were transmissible as the human influenza viruses are in the ferret model. The ferret model is really an excellent one to explore this question. Then we decided to look at zero prevalence in different age groups. We also sent several of these viruses off to New York Medical College. Doris Booker and her team have already produced one high growth vaccine -- well, we hope it is high growth -- one vaccine candidate. She is working on others.
There is a very interesting story around one of these most recent cases, which will be published, but I do not have time to go into the details.
Now, I will move onto H5N1 viruses, which have been causing problems in a number of countries for what seems like forever to me. H5N1 activity is not always reported by countries in which the virus has become endemic in birds so the reports actually under represent what is actually going on.
We believe that the viruses are endemic in poultry in Indonesia, in Vietnam, possibly in China, and certainly in Egypt and potentially other countries. These are the countries where outbreaks have occurred or human cases have been detected. You can see that we have the genetic clades represented and have them listed where we know the characteristics of the viruses.
It looks like fairly complex nomenclature. We really had to work on a commonly accepted global nomenclature so that we would all know what we were talking about and so publications would use this nomenclature. It has been published by WHO and is updated periodically as the viruses continue to evolve.
If we look at just infections in humans that have occurred during the period October through February, you can see that Egypt has reported the most cases. We have two cases and a possible third one reported in Cambodia. The cases are continuing in the countries where the viruses are endemic in poultry and where there is a lot of interaction between people and birds.
This just is a very quick summary of the evolution of the HAs of these different H5N1 viruses. We do not know if the evolution is being driven by some of the vaccination in birds or just due to the fact that these viruses are infecting such a wide variety of different species of birds, which, as you may know, are quite evolutionarily distinct and there may be different selective pressures going on in the different avian hosts.
Needless to say, the picture is very, very complex. These large triangles just indicate the clades and sub-clades of viruses that have been detected in recent years.
We are trying to keep up the WHO group in conjunction with our partners from the agricultural side, in particular with a group known as OFFLU, which is the FAO/OIE influenza group. We are pooling data and really trying to put together, as complete as possible, a picture of what is actually going on in birds and humans.
The red viruses are the ones that may be familiar to some of you. These are the clade one viruses that were identified back in 2003/2004 and were produced -- actually, bulk amounts of vaccine were produced and clinical trials were carried out using the Vietnam strains. I think those data are quite familiar to many of you.
The clade one viruses have continued to evolve and we have chosen an additional strain, shown here in red, as a vaccine virus. Then there are some reference viruses here in blue. We put these viruses into ferret so we can really keep track of both the antigenic and the genetic characteristics of these viruses. I will not go into a whole lot of detail on the antigenic analysis, but the clade one viruses have remained, I would say, relatively stable.
The clade 2.2.1 viruses are of quite a bit of concern to us simply because there have been so many human cases detected in Egypt. These are the viruses that are circulating in poultry in Egypt, primarily in Egypt. Here we have the group C viruses. Within the clade 2.2.1, there is a lot of divergence and we have groups A, B, C, et cetera, through F.
The human infections have occurred in group C. These are viruses that are circulating primarily in the backyard flocks in Egypt, while these viruses up here represent viruses that are circulating in commercial poultry in Egypt. The human infections are occurring, as one might expect, as a result of individuals coming in contact with their home farmed birds.
The clade 2.2.1 viruses have remained relatively stable, but we are continuing to fine tune and update the vaccine viruses because we are seeing so many infections in humans. The ones that have not really been covered well by our vaccine viruses are in some of these other groups that really have not caused the human infections.
Clade 2.3.4 viruses have been circulating in China and have been detected in Hong Kong over the past few years in birds, wild birds, or birds that have washed up on the river banks -- sometimes poultry, sometimes wild birds. Also, we have a very good collaboration with folks from the Ministry of Agriculture in Vietnam. We have been receiving viruses from them for analysis, along with viruses from the human health sector.
You can see there is quite a bit of diversity. We have a number of vaccine candidate viruses that have already been developed. Here is the antigenic analysis that the clade 2.3.4, unlike some of the previous HI tables, you can see here that the viruses are more variable and some of the antisera do not cover these viruses very well.
Here is our Anhui vaccine virus, which does not do a terrible job with these more contemporary viruses. This is an '05 virus and we do have bulk vaccine in the U.S. stockpile.
Clade 2.3.2 H5N1 viruses have become a bit of a worry to us, primarily because it does appear now that these viruses may be spreading -- may have been spread to South Korea and Japan where they are having a lot of problems getting rid of these viruses through stamping out processes in their poultry. We believe they have been spread by wild birds to Japan and South Korea.
These are the new viruses that we have analyzed in recent months. Here is the vaccine virus, here. We are very interested in having an updated vaccine candidate.
Here is the antigenic analysis. Here is one of the older viruses. It seems to cover some of the viruses reasonably well, but certainly not all of them. We are looking to make, as I said, a new vaccine candidate.
Here is clade seven. We are not paying too much attention to this clade, but just noting that it is still around. To my knowledge, it has only caused one human infection and that was quite a long time ago in 2003.
If you can just look and get the gestalt from this, you will see that there are a number of institutions making these candidate vaccine viruses. They are all made by reverse genetics because you have to excise the multi-basic cleavage site from the HA so that you can have a safe vaccine virus to be used for manufacture. We have quite a number of different vaccine viruses, some of which have been used to make bulk vaccine for stockpiles. They represent a number of these different clades.
These are more recent reassortants that have been produced, representing clade 2.3.4. This was made by the WHO collaborating center in St. Jude that works on the animal side of things. Then we have a clade seven virus that is produced at CDC. We are expecting that they will be available. When they are available, this information will be posted on the WHO website.
These are viruses that we proposed for preparation of candidate vaccine strains. We have an updated Egypt virus, which is representative of those group C viruses, but then clade 2.2.1. I apologize for the miss-spacing here. This A/Hubei virus is a clade 2.3.2 virus, which was shared with us by our WHO collaborating center colleagues in Beijing. Then St. Jude in conjunction with the H5 reference in Hong Kong is going to also make a vaccine candidate using another one of the clade 2.3.2 viruses.
I will not talk too much about H9N2 viruses, but they are very concerning because they are widely prevalent in bird population throughout Asia and the Middle East. These viruses have some altered receptor-binding properties so that they bind better to human receptors than most avian influenza viruses do.
There are two groups on H9N2 viruses, both genetically and antigenically. We have the group of viruses sometimes referred to as the Chicken Beijing group. Actually, there are a number of subgroups there. I will not go through all of it, but there are some vaccine candidates and reference viruses and there are some human infections that have been caused by these viruses.
Then there is another group that has caused human infections, shown up here, and these are the G1 viruses. There are vaccine candidates either available or being worked on. These viruses are pretty heterogeneous. So the G1s are shown here and the G9s or Chicken Beijing virus is shown here. There is a bit of variation, but the bottom line is that we have these viruses that are available and then CDC is working on an A/Hong Kong virus from a human infection. We expect it will be available to be distributed in of May 2011.
In addition to H5 and H9 and swine origin H3N2, there are a number of other candidate vaccine viruses that are available of different subtypes. We have H7. We know these viruses have caused human infections and there are a variety of vaccine candidates available. H2, we are always vigilant about detecting H2 infections because H2 viruses are circulating in birds and H2 did cause a previous pandemic.
I think I will stop there and acknowledge all of the people I did before along with the OFFLU Network, the USDA, and our collaborators at NAMRU-3 and the Egyptian Ministry of Health. Thank you very much.
DR. ROMERO: Thank you. Dr. Gray?
DR. GRAY: Nancy, wonderful presentation. We heard a lot about vaccine candidates for avian flu. We know the recent pandemic was of swine origin. I have heard from OIE reports that the pandemic virus has now been detected in perhaps 20 different countries in their pigs and that at least four reports of where it is reassorted with enzootic viruses from pigs. I am wondering what we are doing with respect to developing vaccine candidates from swine origin viruses.
DR. COX: What you have said is absolutely true. We did bring together as much data as we could for the WHO consultation by working very closely with our OFFLU collaborators. We know that there are a number of reassortants that have been detected. There does not seem to be a predominate reassortant. So it is sort of mixing and matching of genes from viruses that are circulating in pigs. Pigs do seem to be an excellent mixing vessel.
I think what we are always most concerned about is the hemagglutinin and the neuraminidase components. What we have seen so far is that for the pandemic H1N1 viruses that have gone into pigs, there does not seem to be any appreciable antigenic drift in pigs, either. We would expect that if that virus were to reemerge, we would not see anything new. We have not seen any brand new subtypes -- that is to say an H2 or an H6 or whatever -- emerge from pigs with that internal gene cassette from the pandemic virus or the triple reassortant. That is what we would really be looking for.
What you have mentioned is definitely on our radar screen. We are working very closely -- more closely than ever before with our veterinary counterparts.
DR. ROMERO: Any other questions or comments? Very good. Thank you very much. We will next proceed to the open hearing and I will have Donald Jehn make some comments.

Back to top

Agenda Item: Open Public Hearing
DR. JEHN: As part of the FDA Advisory Committee meeting procedure, we are required to hold an open public hearing for those members of the public who are not on the agenda and would like to make a statement concerning matters pending before the Committee. Dr. Romero, will you please read the Open Public Hearing Statement?
DR. ROMERO: Yes. This is the open public hearing announcement for particular matters involving specific parties meeting. This is for the influenza vaccine.
Both the Food and Drug Administration (FDA) and the public believe in a transparent process for information gathering and decision making. To ensure such transparency at the open public hearing session of the Advisory Committee meeting, the FDA believes 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 state, to advise the Committee of any financial relationship that you may have with a sponsor, its product, and, if known, its direct competitors. For example, this financial information may include the sponsors 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. JEHN: We have received two requests to speak at the open public hearing. Have both of those speakers -- MedImmune and -- have they loaded onto the computer? Just one? We are missing one. Okay, there are two speakers, again, who have request: Kathleen Coelingh and Dr. Oleg Prokopyev of the University of Pittsburgh. Kathleen, do you want to go ahead and start.
DR. COELINGH: It is Kathleen Coelingh from MedImmune Astra-Zeneca. My comment has been addressed by the Committee.
DR. ROMERO: Thank you.
DR. PROKOPYEV: Hello. Thank you for the opportunity to speak. My name is Oleg Prokopyev. I am an Assistant Professor of Industrial Engineering at the University of Pittsburgh. The goal of my five minute presentation is to share with the Committee and the public some of our work, which was funded by the National Science Foundation on optimizing the annual influenza vaccine design. This is a joint work with Osman Ozaltin, a Ph. D student of industrial engineering at the University of Pittsburgh, Andrew Schaefer, my colleague there, as well, and Dr. Mark Roberts at the Department of Policy and Management.
The goal of this project, which is funded by the National Science Foundation, is to provide a more rigorous, systematic way, based on some of the mathematical models for the selection of strains. Basically, our model -- we needed to design a mathematical model, which determines which strains should be selected and when. The goal is to maximize societal benefit.
The underlying math is basically an integer program. I do not want to get into too much details, but this is a mathematical framework, which allows us to model logical integer decisions -- yes or no decisions. For example, if you are given a particular strain to be selected, yes or no.
Some toy examples, which many of you have heard about our traveling-salesman-problem -- so if you have a bunch of cities and you want to visit all the cities and get back to your hometown, which is the most optimal way to do it? There are a variety of application domains of this particular mathematical framework. For example, some of you who are going to fly back home are going to use air carriers, some Delta or whatever. When they schedule their flights, assign pilots to flights, et cetera, they actually utilize this particular kind of framework.
In the last probably 10 or 15 years, there was a push to apply this type of rigorous systematic models to problems in healthcare, computational biology, and bioinformatics. Our first paper on the topic has just been accepted for publication in Journal(?). So I have a copy of this paper for interested people.
Basically, we calibrated our model using publicly available CDC and FDA data from 2008-2009 season. One of the key issues here is that our model, besides providing the opportunity to answer the question to the design of the strain selection problem, it also provides the opportunity to answer some public policy questions.
Actually, we using this old data, we tried to answer some of the public policy issues regarding the flu shot design and manufacture. Some of our conclusions were as follows, for example, revisiting decisions more frequently might actually be beneficial. Cell-based manufacturing techniques may have a large impact because they provide more flexibility for the manufacturers.
Another issue, which has been discussed by the Committee, is the tetravalent flu shot. There are some obvious trade-offs: better immune response, more difficult to manufacture. And actually our model allows to basically evaluate the benefits of tetravalent flu shot versus a regular one. In many cases, our model shows a benefit to the tetravalent shot.
Here are some conclusions. Our model, we believe that the composition of the annual flu shot should be addressed along with the timing of its manufacturing. Our work, our model, which is still ongoing, allows also to incorporate uncertainties into this framework, including, for example, strain growth, prevalence of each strain, et cetera.
We are interested in working with the Committee to improve the model, to get some feedback from the interested people, and eventually provide it to the Committee as a decision-making framework. Thank you
DR. ROMERO: Thank you. Are there any other persons who wish to make a public statement? Apparently none, we will move forward. Now we will go to Committee discussion and recommendations.

Back to top

Agenda Item: Committee Discussion and Recommendations
DR. ROMERO: Do we have the -- there it is. The Committee discussion will focus around which influenza strain should be recommended for the antigenic composition of the 2011-2012 influenza virus vaccine in the United States. Data to be considered include the epidemiology of circulating influenza viruses, the antigenic characteristics of influenza virus strains currently circulating in the human populations, the serologic responses to circulating influenza vaccine viruses of persons immunized with current influenza virus vaccines, and manufacturing considerations, including the availability of suitable vaccine candidate strains. We will begin. Any questions or comments?
I think I will start by asking a little bit of a follow-up to the DoD data that was presented with regard to efficacy of the LAIV strain. Does anybody have any information regarding the efficacy in younger populations, specifically children that are receiving that strain? Are we seeing similar reductions in the efficacy that we are seeing in the military population? You would expect that those are healthy, young adults that are essentially naïve to this vaccine.
DR. COELINGH: Kathleen Coelingh from MedImmune Astra-Zeneca. We do not have efficacy data for that particular formulation of that strain. However, at the recent WHO meeting in Geneva, the Centers for Disease Control presented some information for the younger population. I do not know if there is anyone from the CDC that can represent that information in the room.
DR. COX: I believe the question was really probably focused on young children and we did not have -- at least to my knowledge, we did not have the ability to stratify by age group down to that level with the number of infections that we had seen. Stratifying by age and by subtype, you get small numbers and then your confidence intervals are very wide. I do not believe we can address that question.
PARTICIPANT: MedImmune, as well. Tony Piedra has published some data, at least in abstract form, and presented it on the effectiveness of the H1N1 strain in children in the study that he is conducting. He showed it to be effective in all the age groups that he looked at, as well as they have some indirect effectiveness in adults.
DR. ROMERO: Could you summarize that? You said it has been published in abstract form.
PARTICIPANT: He gave a presentation at -- I will have to look up which meeting. I can certainly look to help provide that to the Committee after this.
DR. ROMERO: Thank you. Dr. Debold?
DR. DEBOLD: Could anyone speculate on what some of the mechanisms might be for decreased effectiveness with the LAIV?
DR. RUSSELL: I particularly appreciate that question. I will just point out a couple of issues from our perspective, as we have struggled with this.
Problems with cold chain has been ruled out with the LAIV. You would expect if there was a problem with cold chain, you would see an effect across all three subtypes, which we are not seeing. It is really focused on the H1N1.
This is very recent data, really within the last two weeks and much of it within the last week, getting some of these numbers. The virus, itself, we have no evidence has changed significantly to say there is a drift of the virus of that pandemic H1N1, but there is a lot that has not been analyzed yet with regard to sequence analyses, neutralization of the viruses that we are seeing in those previously vaccinated individuals.
I am really interested in other people's thoughts on what this might be representing, as well. Those are some of the things that we have struggled with in the last couple days, actually.
DR. DESTEFANO: I just have a question about the methodology that was used in the Department of Defense effectiveness studies. I understand that these were all vaccinated individuals and all of them had an ILI, at least, to get into the study. Has there been any kind of evaluation to see how this kind of methodology would compare with a more traditional effectiveness study of vaccinated and unvaccinated and laboratory-confirmed influenza as the outcome?
DR. RUSSELL: So there we about seven different methodologies presented. Several of them were traditional vaccinated/unvaccinated, with influenza/without influenza, and laboratory-confirmed. Several of the methodologies did present that way. There were a couple that were all vaccinated. The recruits are one.
It still used the classic vaccine efficacy calculation, one minus odds ratio, but those determined unprotected were within the first two weeks, 14 days after receiving the vaccine. The other data presented that was all vaccinated was looking at the relative ratio or distribution of the influenzas, both all influenzas and then the H1N1s by the two vaccines.
There is both there. I would be happy, again, to address specific questions with any one of those methodologies. They all tended to suggest the same thing, which is what was interesting from so many different perspectives.
DR. DEBOLD: Just a follow up question. Could part of the mechanism have anything to do with the fact that they were, perhaps, given other live virus vaccines simultaneously or not?
DR. RUSSELL: Certainly the multiple vaccinations received in our recruit setting has caused some concerns among the individuals that any one vaccine may have reduced immunogenicity antibody production, but I have no data to speak to that potential mechanism. I will point out, which I did not, that DoD has a serum repository. Every one of our recruits has a baseline.
Then a study is being started by Naval Health Research Center in cooperation with the serology branch of CDC there, looking at week four of recruits. You take serum from week four and then we have our baseline when they became our recruit in our repository, paired samples before and after exposures to look at the different components of the LAIV and the antibody response.
Dr. Cox was just mentioning to me, as we all know, that the antibody response for the LAIV is not like the TIV, but there is still some hope that something can be teased out. They are planning to do both neutralization and HAI in that study. That might speak to it a little bit, but how much the multiple vaccinations that they do receive in that setting and some other live viruses are contributing will not be addressed.
DR. ROMERO: Thank you. Dr. Wharton?
DR. WHARTON: A follow-up on the military data. Did you use the monovalent H1N1 vaccine during the previous year? If so, do you have any efficacy data on the monovalent H1N1 vaccine?
DR. RUSSELL: You saw the vaccine effectiveness figure, which I think was 51 percent in the recruits that was shared. I pointed out that that was largely an effectiveness with the trivalent, previous season vaccine on pandemic H1N1. When the monovalent became available and given, essentially there was no more transmission of H1N1 so we have no vaccine effectiveness of that monovalent with pandemic H1N1.
DR. ROMERO: Thank you. Any other questions or comments? Does the Committee have enough information on circulating B strains to offer a comment later in the vote? Or does that need to be discussed further? Okay, sounds pretty good. Let's move forward. Dr. Weir, do you have to summarize the vote question?
Apparently then we are going to do this -- I will take care of it. Sorry. I did not mean to pull you up here. We have several questions. We will be taking three essential votes, each addressing whether to change one of the strains. If there is unanimity in a decision to retain the current strain, then there will be no further discussion. If there is dissention about change, then we will discuss and make recommendations on that.
Let's begin. Option one, for the influenza A(H1N1), the question to vote yes or no is to retain the current strain A/California/7/2009 H1N1-like virus. If you want to retain this, vote yes. If you think we should change it, vote no. You may now vote.
DR. JEHN: So we have 15 yeses and one abstention and zero nos. In the case of an abstention --
DR. ROMERO: I do not know. We did not come up with a plan for the abstention. Do we need to discuss this further? I think we will just go around the table and talk quickly.
DR. JEHN: The abstention was Dr. Debold.
DR. DEBOLD: I abstained because I do not feel at this point that I have enough information to make an informed decision about this. I think there are outstanding questions about effectiveness. I know there are other issues about safety that have not been discussed and we were not asked to discuss that. I would feel more comfortable abstaining.
DR. ROMERO: Do you have any candidates that you might want to offer? Any other questions or comments? Okay, I think we can move on to the next one then? So the same question for Influenza A(H3N2). The question, vote yes or no, is yes to retain the current vaccine strain A/Perth/16/2009 H3N2-like virus, no if you would like to change it.
DR. JEHN: Okay, we have a unanimous 16 votes yes on that motion.
DR. ROMERO: Very good. Any comments from the Committee on that? Final vote question -- influenza B --
DR. WHARTON: I apologize. I think you asked previously if the Committee wanted to discuss the Influenza B candidates and nobody said anything, but I would like to raise that issue again. I am actually quite uncomfortable about this next decision. Not that I think talking about it will necessarily make it any better.
At least if my understanding of our history in recent years with these type B decisions are that, I think it is correct, in the period of time we have been alternating -- we have been moving back and forth between the Victoria and Yamagata lineage, we have only used candidate viruses from a single lineage from one or two years before we moved to the other lineage. I believe this will be three years of retaining the Victoria lineage, if indeed we retain the current vaccine.
Perhaps the H1N1 pandemic experience has so disrupted what happens with influenza that what has happened in the past will not predict what happens with the B virus this time. But I feel like that is what we are assuming in retaining the B lineage. I am sure these discussions must have taken place at the WHO meeting previously, but I really would appreciate hearing thoughts from others before I have to pick one of these buttons to push.
DR. ROMERO: Thanks a lot. Comments?
DR. MCINNES: I have a question. I am trying to clarify I think it was Dr. Baylor's response. The feasibility of incorporation of two Bs into this season's vaccine is zero, low, or possible? What is the answer?
DR. WEIR: I don't think he actually said --
DR. MCINNES: I am just clarifying. That is why I am asking. What did he say?
DR. WEIR: I think the manufacturers gave you a pretty good idea of the likelihood of when they might have -- when they hope to have a quadrivalent vaccine ready. I cannot remember all of the numbers, but I think they all gave you an indication. If I remember correctly, it did not sound like most of them were on target for the coming year, 2011-2012. Someone correct me if I am wrong.
What Norman clarified was that, of course, during the evaluation of whether feasibility of vaccine that a particular strain was not -- the manufacturers can evaluate the feasibility of the vaccine with current strains that are available. As they told you, many of those studies are ongoing. Does that help?
DR. LEVANDOWSKI: Could I follow up that with another question? There is kind of a middle road. What about a supplemental vaccine for specific populations, for example, pediatric populations or military or somebody else who might benefit from either their direct potential for exposure or their lack of immunologic priming that would -- in the situation with the pediatric population, for example, those children who have not been exposed to the alternate lineage really do not have any antibodies to speak of against those other strains. We are developing, with the two year model, a cohort of several million children, who may not be protected by receiving the vaccine with just the single strand.
DR. WEIR: I think that the idea of a supplemental, obviously the pandemic example shows that that is possible, if driven by the data. I guess we would have to go back to the experts and the data to ask whether that is warranted now. Clearly, our example shows that if the data suggested, we could assemble a VRBP AC and do that. It is always possible.
I do not know if Nancy or someone wants to comment on how much -- how strong the data is at this point to recommend something like that? I think the data was actually fairly clear, but it was your presentation.
DR. COX: I think that it is very difficult to predict what is going to happen with the influenza B viruses. We can only say what has happened over the last six months. Clearly, if we look globally, the B/Victoria lineage viruses predominate overall.
China has a little bit different story to tell, but towards the end of the season they were seeing about equal numbers, albeit very small numbers, of B/Vic and B/Yam lineage viruses circulating in China. For some weeks, it was actually nine Victoria lineage viruses and seven Yamagatas. It was vacillating back and forth. I think that we simply have to make the best judgment call that we can based on the information that we have at the time. That is what the WHO group did.
DR. WEIR: Can I follow up with that? I do not know if this is possible, but in follow up to Roland's question, can you comment at all on what might be a tipping point of how much data would sway you later in the year for the need for a different strain? Is there any way to quantify that?
DR. COX: I think that the solid information that we would need would not be available. It is not quite what you are asking, but the solid information that would help us make a decision would not be available until the southern hemisphere season is taking off. Depending on when activity occurs in the southern hemisphere, what viruses predominate and so on, it might be earlier or later in the season before you would know whether a switch had occurred and in Australia, New Zealand, South Africa, and so on you started to see the switch over to the Yamagata lineage.
It is a really difficult decision. I must say that we have not always been very successful, quite honestly, in choosing the lineage that would predominate based on data acquired a year prior to the season that we are trying to protect against with the vaccine.
DR. ROMERO: Any other questions or comments? Please, go ahead.
DR. DURBIN: To Nancy, the predominance of Yamagata in China this year, was that new from previous years or had you seen some activity in previous years as well?
DR. COX: There had been -- Yamagata had been circulating widely in China in previous timeframes. We are really looking at four to six month time chunks so we had seen B/Yamagata lineage viruses predominate in China for a while.
DR. GELLIN: You thought these first couple of votes would be predictive of the third. We have had this conversation before, but as we talk about a fourth strain with capacity in mind, would this be for all populations or would it be for some populations? In the past, it has been primarily pediatrics, which have been the concern. I just wanted to put that out there because we are also concerned about overall capacity, as well.
DR. ROMERO: So this would imply that you would have different vaccines for different age groups?
DR. GELLIN: Well, if you have different risks for different age groups, then maybe you want different vaccines for different age groups. I do not know the data well enough to know if that is right and if so, where you are drawing the line.
DR. COX: When we were -- a couple of years ago, I believe Tony Fiori gave a presentation where we were looking at the public health benefit of including a fourth strain, a second B strain in the vaccine. I have not looked at those data very recently. They have not been published because we had a pandemic in the interim, but from my memory the greatest public health benefit was in the children, in the younger age group. It makes sense in terms of the cross-reactivity of antibody.
I think that is something to consider. We do now have a vaccine that is -- the high-dose vaccine that is used exclusively in the older population and LAIV recommended for certain age groups. I think one of the things that we are always interested in is optimizing vaccine effectiveness. For influenza, it may be that we need to have slightly different vaccines formulated specifically for populations. It is a manufacturer's nightmare, but maybe that is what we need for public health.
DR. EICKHOFF: I think there are a lot of good reasons -- many good reasons to consider a separate pediatric formulation. The big bugaboo in the past with supplemental vaccines has always been confusion among various provider groups in terms of what to do.
Now we have an old folks vaccine. We have an influenza vaccine that is effectively a universal use vaccine. Children get an added benefit from a quadrivalent vaccine that is inconsequential to adults and the elderly.
It would be an additional cost for another separate vaccine, but I think one could make a very cogent argument that since this would be a year after year after year event as long as this co-circulation goes on, can make a strong case for a separate vaccine for children.
DR. ROMERO: Very good. Dr. Levandowski? Bruce?
DR. GELLIN: Let me just say -- and I am sure there is somebody here who can comment on this with authority, but the entry compensation program, I believe currently is hinged to a trivalent vaccine. If you change this, either a fourth valent or a separate, we have to accommodate that. This Committee is really looking at what strains should be available to protect the public and then there are subsequent issues, but that is a problem that I believe would have to be solved.
DR. ROMERO: Excellent point. Any other comments? Does the Committee want to move forward and vote? I know you do not want to but we have to.
DR. GELLIN: I guess the question is at the end of this vote, is that the end of the story or do we have another discussion afterwards?
DR. ROMERO: I think we may have another discussion afterwards.
DR. GELLIN: I was telling Melinda, here, we obviously have to make a decision on what a next component would be, but I think it is also related to what we talked about before about what other things might have to be in place like getting those sheep lined up and doing some other things, if, indeed, this is a route that somebody wants to go down.
DR. ROMERO: I think if the Committee has more comments or wishes to discuss it further, we can certainly proceed and discuss after the vote, if necessary. It is our time. Yes, please, go ahead.
DR. AIR: Do we have any sense of what proportion of the population are already protected against Victoria from previous vaccination years?
DR. COX: We do not do zero prevalence studies -- cross-sectional zero prevalence studies so we really cannot tell you what proportion of the population would be expected to be protected based on antibody. After vaccination, of course, the antibody does decline so you have that to take into consideration. Sorry I cannot answer that. Perhaps in the future we will be able to do more extensive zero prevalence studies that will help us make these kinds of decisions.
DR. ROMERO: So how late into this vaccine preparation season can we wait to make the final recommendation or can they switch over? If anybody in industry wants to comment?
DR. WEIR: Could you repeat that?
DR. ROMERO: Yes. Let's say we do not have a firm -- there is not a firm recommendation, but there is still some concern about one or the other strains being a problem, how late into the season could you shift or, in other words, say we need to change? Are we locked in at this moment?
DR. WEIR: I could give you my opinion, but you probably should ask the manufacturers what they think.
DR. LEE: I think you have -- probably the main challenge right now is that the particular candidate virus that is being considered, we do not have any reagents for it. We have a reference virus that has been reasserted and we can certainly generate and have some understanding of how productive it is, but the reagent would be the main thing.
In terms of timing, it is difficult to say because not knowing what the productivity is of the B/Hubei -- I cannot remember the full name of it, but certainly that particular strain, that is the one that is certainly the leading candidate for B/Yamagata lineage. I certainly would not want to wait more than a few weeks.
DR. GELLIN: But, Sam, the productivity is always an issue, but if it is not going to be a quadrivalent for the entire population then it is less of an issue because you do not have to wait for all of the four strains to match up with the other three, right?
DR. LEE: I think that would be correct, but still then you are counting on anticipating what immunization rates might be this year. In light of the abundant supply from this past year, if that were to now become limiting, we certainly do not want that to happen. Certainly, having the decision as early as possible allows manufacturers to produce plenty of supply for that.
DR. ROMERO: Any other questions or comments? All right. Thank you very much. Let us move to the vote. The question then is influenza B, to retain the current B/Brisbane/60/2008-like virus, B/Victoria lineage, vote yes. If you move to change it, vote no. Please vote.
DR. JEHN: All right. We have sixteen unanimous yes votes to retain the B/Victoria.
DR. ROMERO: Further discussion. Dr. McInnes, I think you had your hand up.
DR. MCINNES: So in addition to the sleepless night last night, I might have another one tonight. I think we make the decision we do with the data we have on the table, but I think, for the record, there is obvious discomfort and concern about this.
Every year -- I know I have not been here for all 125 of them, but I have been here for a lot of them and we are always struggling with this B issue. It goes all the way from when we did not have many isolates to look at and I think our colleagues have done a great job in increasing the number, but it has not completely helped with the clarity of the situation.
I think this quadrivalent is obviously a very good step forward in this direction. I am concerned because just historically, at least in the last ten years, we have done the two year, two year, two year switching between and I think about the birth cohort that could be at risk here. It is concerning.
I think it would be very helpful if efforts would continue on the Yamagata in terms of reagents, working with the virus, because I fear that there may need to be some response at some point and I would hate not to be in the position to be able to do that, even if it is late.
DR. ROMERO: Well said.
DR. GELLIN: You need the buttons that say yes, but and no, but. Then you would get a different color scheme on here. I think this is a yes, but. We have been through the pandemic. We have talked about preparedness lots in the last decade. It seems to lock in three and say thank you very much, we will go home and hope for the best is not really what we want to do, given the sleepless nights that Pamela gets and we all get.
I would think that this is an opportunity to lean forward and say, well, these are the three, however, since we do not know and we may not know until later, to act as though we might have to anticipate this and to start the mechanism to get the reagents ready. And send the signal that since there seems to be more production capacity currently than there is demand for trivalent to start to signal towards that.
I realize that there is not a licensed product that is going to be available, but I think that if there is a problem and we have a solution in the works, we can figure out a way to provide antigens to people to protect them. That is my yes, but.
The second is I wonder if the modelers have any predictions. They told us about optimizing so do the modelers have any insights into which way we should go on this B decision. If you had to pick between the Bs based on your models, do you have any insights? I can model four. I cannot model three.
DR. PROKOPYEV: We did not calibrate the data with the current -- we did not calibrate with this year's strains so I cannot really say, but what our results show in the old data is it all depends on the manufacturing capacity. If they have the additional capacity to do four --
DR. GELLIN: The question is if you had to pick based on your models. If you had to pick a B --
DR. PROKOPYEV: We can do it, but we need to calibrate on the actual data. We did not do it on this year's data.
DR. ROMERO: Let me have Dr. Weir, Dr. Rennels, and then Dr. Gray.
DR. WEIR: I just wanted to comment on Bruce's comments. I think we tried to get across earlier today that work is continuing on the reagents and the strains. I think somebody pointed out the wild type viruses for the Yamagata have been distributed. There is work going on reassortants and work will go with reagents. I do not think we have to worry too much about that. That will happen.
I think it is probably a safe bet that it would be going on anyway in preparation -- things are always trying to be ahead of the curve. They would be going on in case it was a southern hemisphere question. We do try to stay one step ahead when at all possible, when there is some data like this to suggest even a future candidate.
DR. RENNELS: Another thing it sounds like we need to be thinking of to be a step ahead is what has to be done to get either a monovalent or a quadrivalent eligible for coverage under the vaccine compensation system. Who does it? Who is responsible? When does that get done? I understand it is an act of Congress. I am looking at Bruce.
DR. GELLIN: I raised the question and I will have to figure out where to get the answer. I think it is a problem we will have to solve.
DR. GRAY: I am new here, but it seemed to me that Dr. Air's comment was quite cogent in the sense that she asked were there any data to suggest, in a national way, what the susceptibility to the two different B viruses might be. It would seem to me with a number of the population-based cohorts that we have identified across the country in a number of different settings that answering this some months prior to this meeting might be a practical way to give more ammunition, particularly if you are, in the future, going to wrestle with four different components, which four to use. It makes it even more complex.
DR. ROMERO: Any other comments? All excellent. Do you think the FDA has enough discussion on this at this point? Okay. Very good. If there are no other comments then let us move on to the final topic of the day, which is an update on the febrile seizure signal after influenza vaccine by Dr. David Martin from the FDA, please.

Back to top

Agenda Item: Update on Febrile Seizure Signal after Influenza Vaccine
DR. MARTIN: Thank you for the opportunity to speak to you today about CDC and FDA efforts to deal with the febrile seizure signal after 2010-2011 vaccine. Extensive time was devoted to this at the ACIP a few days ago. This is meant to be sort of a recap of what was presented there. We understand that some individuals may not be familiar at all with this situation.
In terms of background, the FDA and CDC jointly issued public communications in January regarding a febrile seizure signal that had been found in association with trivalent influenza vaccine in children under two. Currently, there is only one product licensed and recommended in that age group, which is Fluzone.
A little additional background, febrile seizures are seizures that occur in children, who do not have other risk factors such as CNS inflammation, metabolic abnormalities, or other seizure histories. They affect approximately two to five percent of children in the United States. The peak incidence is between 14 and 18 months of age. They are associated with viral infections, genetic polymorphisms, and also whole cell pertussis containing, as well as measles containing vaccine.
In April 2010, an association was reported out of Australia between febrile seizures and a CSL product -- the southern hemisphere vaccine, which contained H1N1 antigens, as well as the other seasonal southern hemisphere formulation antigens. As a result, TGA conducted an investigation. They determined that vaccination for children under five would be suspended in Australia.
Subsequently, the ACIP took up this issue and Afluria, which is the cognate CSL product for the northern hemisphere, was not recommended for administration to children under age nine during this past season.
In addition, every single season, FDA and CDC work together through various systems, which I will discuss briefly in a moment, to generate hypotheses, look for new safety signals with vaccines. So in addition to the normal things that were done as part of this joint effort, the CDC conducted some enhanced medical officer review of all febrile seizure cases in VAERS.
Again, just for additional background here, this is kind of a view -- obviously since I work at FDA, it is slightly FDA centric -- but, of course, with vaccine safety we work hand in hand with the Immunization Safety Office at CDC. This gives you sort of an overall view of what we are trying to do.
Obviously, vaccine products are licensed based on safety and efficacy. This is sort of a general framework for all vaccines that FDA monitors. Even once they are licensed, we want to detect new safety signals through any means that we can. The primary method that we have, obviously, is VAERS, Vaccine Adverse Event Reporting System, which is, of course, a passive surveillance system. Its primary advantage is its national scope and the fact that you really have no a priori hypothesis when you are looking at it so you can potentially find new, unexpected, rare adverse events and investigate them further.
We obviously have some understanding of the safety profile of products from prior clinical trials that have been carried out. We also receive periodic reporting from sponsors. That enhances the broad FDA safety perspective, which we have because we deal with the safety profiles of all biologic products that we regulate. That is complemented by the sort of vertically integrated safety assessments that we receive from sponsors.
Other sources include obviously the medical literature and then the application of data mining methods to the VAERS database. All of this goes into signal detection. Once you have signals of concerns, you are able to potentially plug them into active surveillance systems where you can strengthen those signals.
Signal validation is connected primarily with certain limitations of active surveillance systems. But ultimately if a safety signal survives this entire pathway and remains on the table, then you need to take your a priori safety hypothesis derived from your signal detection activities and you need to test them in a formal epidemiologic study where you have adequate control groups and adequate power to find what you are looking for.
Most of what we are talking about -- again, just for perspective through -- through this talk today is operating right here in this area: signal detection and signal strengthening.
At the FDA, we use Empirica, which is a program that allows us to calculate a statistic called the Empirical Bayesian Geometric Mean. This quantifies disproportional reporting in VAERS.
I would like to emphasize VAERS is strictly reports from physicians or members of the public or sponsors or any other healthcare provider to the FDA and CDC, to the VAERS system. In a sense, you can think of VAERS as an extension of the clinician's intuition about what might be going on. We try to maximize the use of that extension of the clinician's intuition by looking for what we call disproportionality in reporting.
We cannot calculate rates in VAERS, but what we can do is say this is interesting that a certain vaccine and a certain event or health outcome seem to be reported at greater frequency within the database. Those reports of those paired make up a larger proportion of what we are seeing than we would otherwise expect. Obviously, this is an adjunct to the clinical intuition of our physicians, who are reviewing these reports.
To be slightly more specific, we do not use the point estimate, which is referred to as the EBGM. We use the lower bound of the confidence interval surrounding the EBGM as our cutoff. At the lower five percent bound, we look for that bound to be at least two so that we are reasonably confident that this disproportionality is at least twice what we would expect to see in the database.
I would just like to emphasize for everyone here that all of the limitations of VAERS apply to VAERS data mining. So there is nothing magic about this, but this is allowing us to quantify disproportionality that we might otherwise just suppose is occurring.
In early December, VAERS data mining signaled for the combination of Fluzone and the coding term febrile convulsion. The EB05 was greater than two with no stratification in the database. This signal strengthened when we used the 0-18 month age stratum. I would like to note that our age strata are shared with the Center for Drugs. These are not necessarily vaccine-specific age strata that would be used.
This signal has persisted to date. We have done some additional analyses in the VAERS database. Using the data mining system, we have found no interaction with other vaccines and Fluzone and we have found no other independent vaccine product/febrile convulsion pair findings to date.
The day after the FDA medical officer found this data mining signal we had an interagency meeting with Immunization Safety Office and the Division of Epidemiology at CBER. We reviewed available VAERS data mining findings, as well as preliminary vaccine safety data link rapid cycle analysis findings.
At that time, the primary analysis in VSD had not yet signaled, but there were indications from the secondary analysis that it was likely that the primary analysis would signal in the future. We made various joint decisions, but one thing that came out of it was obviously a joint complete VAERS review to ascertain all febrile seizure cases in VAERS. Really that just validated what the CDC had already been doing. We also made that analysis public in January with a communication both on the FDA website and on the VAERS website.
Now, I will briefly recapitulate what was discussed more extensively at ACIP. This is the CDC's vaccine safety data link system. It is a network of eight large populations from HMOs and managed care organizations in the United States. It covers approximately three percent of the U.S. population.
As we go forward in this talk, I will be briefly touching on the prospective surveillance, which was going on in the VSD for febrile seizures, the prospective surveillance for febrile seizures after a new product this year, Prevnar 13, as well as a joint signal evaluation that was conducted in a pool database to cover both influenza and pneumococcal vaccine, specifically Prevnar 13.
For those who may not be as familiar with the vaccine safety data link, this just provides you a little bit of background. The health event of interest -- there are various ones that are specified. Because of the southern hemisphere CSL situation, this code, 780.3, was specified as an outcome of interest this year.
This has an 80 percent positive predictive value in the literature for febrile seizures. There were rapid cycle analyses going on with this as the health outcome of interest. I should mention that that 80 percent predictive value applies to inpatient ED coding only, not to outpatient where the PPV drops to the teens.
The risk window was zero to one day. That was noted in the TGA report from Australia. The comparison window was 14-15 days to avoid confounding with measles containing vaccines. The primary age strata looked at were six months to four years and 5-17 years.
This is just a depiction of the primary analysis -- the self-controlled rapid cycle analysis. In this analysis, individuals who are vaccinated -- cases are ascertained that occur within day zero to one. Later, cases are ascertained if they occur in 14-15. Individuals serve as their own control, in order to control for time-fixed confounders.
This just gives you an idea of the size of the analysis. This was not the entire VSD. This was Harvard Pilgrim. These are by January 30, 2011. 200,000 TIV doses had been administered. You can see this goes from October to January. The statistic, which is calculated -- this is using maximized sequential probability ratio testing -- the log likelihood ratio is the statistic of interest for determining if there is a signal.
This red line going across here is the critical value, which determines when a signal has occurred. As you can see here, the signal was positive on December 26, 2010, just a few weeks after the VAERS data mining signal. This signal has persisted.
One of the things that was noted about the cases that make up this signal is that the majority concomitantly received either Prevnar 13 -- right here, 75 percent -- and/or DTaP-containing vaccines, 65 percent. Making this difficult, 11 children of these 20 received both products and possibly other concomitant vaccines, as well. Only 15 percent of the seizure cases received TIV along.
It was also of note that among the vast majority of TIV vaccinees who were not cases, there seemed to be less concomitant vaccination. You can see DTaP containing concomitant vaccination only 34 percent, only 36 percent Prevnar 13. This was of interest, but Prevnar 13 was of special concern since it was a new product this season. In addition, a prior VSD analysis looking only at the main effects of DTaP had not signaled in the past.
One of the next things that was done was there was an ongoing, as I mentioned on the prior slide, there was an ongoing evaluation of febrile seizures after Prevnar 13 in the southern California Kaiser database. That analysis used some slightly different methods, but it had not signaled. At any rate, it was decided that there was really a need for a joint signal evaluation with a pooled population and harmonized protocols.
A joint analysis was conducted. The preliminary results were presented at ACIP. These are divided into three different age groups: 6-11 months, 12-23 months, and 24-59 months. Same ICD-9 codes. Same assumptions there. The comparison window has now been harmonized to 14-20 days. The risk window is the same. What I will show you on the next slide are some attributable risks, which have been estimated from the self-controlled design.
The first thing is that there was no signal in the 6-11 month age group and there was no signal in the 24-59 month age group. In the 12-23 month age group there was no signal for Prevnar 13 alone, no signal for TIV alone, but there was a signal for joint administration.
To back up and orient you to this slide, in this joint analysis there was actually an effort to look at historical data, as well. There was a separate analysis conducted from August 2005 - April 2010. Obviously, it involved Prevnar 7 and it involved prior seasons TIV products. You can see visually what is there. There actually was a signal with a much lower attributable risk for TIV and PCV7.
On this sort of more relevant data involving current products, from May of 2010 - January of 2011, as I said before, we have a point estimate of an attributable risk of 61. It is kind of hard to read from the graph here, but I do have the number. The confidence interval ranges from 13-109. This can be compared to the excess risk of febrile seizures when MMRV is administered, as opposed to MMR plus V, which is 43 per 100,000.
In the VSD, most of the immunization for influenza has probably already happened. We do not know how much those confidence intervals may tighten up.
Just to summarize, again, we have a VAERS data mining signal for the combination of Fluzone and febrile seizures. There are no other independent products with the event being febrile seizure signals in the VAERS database. There is no apparent interaction between products within the VAERS database.
Obviously, in the VSD system, an active surveillance system, which would really be the next level up for evaluating this type of thing further, we have a signal in the self-controlled analysis. There is also a signal in the historical analysis, which was the type of analysis that was done that was already leading us to believe that the self-controlled would later signal in early December.
There is no signal independently for Prevnar 13 in the southern California Kaiser database. However, this VSD joint analysis has shown us a signal with concomitant administration of the two products with the aforementioned attributable risk.
The background for this is that we have previously described associations between whole cell pertussis vaccines, as well as measles containing vaccines and, starting in 2010, the southern hemisphere CSL TIV and febrile seizures. We also have the background knowledge that a prior VSD analysis found no association with DTaP.
The next steps include continued investigation in VSD. As I said, case ascertainment will still continue until the end of the influenza season. Of course, there will be additional vaccine administrations of all the concomitant vaccines, including Prevnar 13. These numbers are based on that 80 percent positive predictive value. That early chart reviews in the influenza analysis upheld that 80 percent positive predictive value.
Now, these additional cases that have been found in the joint analysis will all need to be reviewed. It is still only track, to date, with approximately 80 percent positive predictive value, but there could be a refinement in the attributable risk estimates if that assumption does not hold true by the end or if new cases are ascertained.
In addition, the role of other concomitant vaccines -- there are efforts to try to determine how to assess this, but it is complicated by the fact that the numbers are fairly small to address a lot of these other concomitant administrations.
At a practical level, this is being worked out at the ACIP. The General recommendations Working Group will consider additional information in coordination with the pneumococcal and influenza and other working groups as appropriate.
I would like to thank both Frank DeStefano and Grace Lee, who provided many of these slides at the ACIP just a few days ago and also the efforts of many colleagues at the Immunization Safety Office, Harvard Pilgrim, and Southern California Kaiser RCA working groups, as well as the CBER Pharmacovigilance Group.
Just for your reference at the very end of the handout, we also have the results of the VAERS analysis just for those who would like to know what was in the public release before.
DR. ROMERO: Comments? Dr. Debold.
DR. DEBOLD: Going back to slide 9, please? On this slide, you make reference to DTaP as well as PCV13 as potentially playing a role in increased risk. However, in the slides that follow, DTaP is not mentioned, in particular on slide 11. I am wondering how these confidence intervals change when you consider the effect of DTaP as well as PCV when given with TIV?
DR. MARTIN: I am not able to actually address that since at the FDA I do not conduct or oversee those analyses.
DR. DESTEFANO: I think that is the issue that David was getting at. Both the PCV13 and DTaP also tend to be given concomitantly. We focused on the PCV13 because it was the one that was most frequently given, more than DTaP. We had a previous study of DTaP that showed that it is not associated with febrile seizures. Also, as David mentioned, PCV13 was new this year. As he said, we are going to try to assess the contribution of other vaccines as best we can over the coming months. We just have not been able to do that yet.
I just want to make one minor correction to one of the statements that you made. It was a great summary in a few minutes of what I guess took us an hour to do yesterday. I think on your slide eight, I think you mentioned those data were from Harvard Pilgrim. Those are actually from eight of the VSD --
DR. MARTIN: Thanks, Frank.
DR. ROMERO: Any other comments or questions? Very good. Dr. Baylor, anything that we need to comment on before we adjourn?
DR. BAYLOR: No. I believe that is it. I just have to apologize for having to step out of the meeting, but it happens. I think that is it. Thank you so much, all of you.
DR. ROMERO: I think at this time the meeting is adjourned. Safe travels home.
(Whereupon, the meeting adjourned at 1:00 PM)

Back to top