provided seroprotective antibody levels to the H3N2 subtype.

I did not describe it, but there is a published study comparing in a head-to-head fashion an MF59 and alum adjuvanted H9N2 vaccine.  The MF59 adjuvanted vaccine provided significantly higher antibody titers than the alum adjuvanted vaccine.

MF59 allows for antigen sparing, and two doses with the clade 1 Vietnam strain provided putatively protective antibodies to the homologous strain, as well as to strains in a number of other clades.

A strain-change study, in which a clade 2 antigen will be evaluated, is under way.

The durability of immunity was seen in a study in which recipients of a clade 0 adjuvanted vaccine were boosted six to eight years later, and rapidly, within seven days, produced neutralizing antibody responses at high levels, not only to the antigen in the booster vaccine, but also viruses in all clades that have been associated with human illness.

Finally, a single dose primed for homologous and heterologous responses upon receipt of heterologous booster one year later.  This was achieved without interference in responses to the seasonal vaccine antigens.

Novartis also has in development an MDCK cell-derived H5N1 vaccine.  A trial currently is ongoing in young adults.  Upon its completion, a pediatric plan will be developed.

In summary, children, including in the U.S., are at risk for infection with pandemic influenza viruses.  There is a substantial clinical experience with MF59 adjuvanted vaccines that thus far has not indicated a safety signal, although data in children remain limited.  MF59 adjuvanted H5N1 influenza vaccine in adults and in children has been immunogenic and well tolerated.  Further clinical studies of pandemic influenza vaccines in U.S. children, we believe, can be justified as contributing to a better understanding of vaccine prevention against a serious problem affecting the health or welfare of children.

DR. MODLIN:  Ted, thank you very much.

Dr. Tsai’s presentation is open for discussion.  Questions?  Dr. Debold?

DR. DEBOLD:  I just have a technical question.  On slide 28, where you have the titers for the toddlers, children, and adolescents, on the far right, does that mean that in the adolescent group the average titer was 67 or is that the number of --

DR. TSAI:  It’s a bit confusing.  This is geometric mean fold rise.  With the second dose, there was a 67-fold rise in titer compared -- and so the geometric mean titer for that group, adolescents, was 344.

DR. DEBOLD:  Okay.  The standard that I keep seeing applied -- is it 40 that you would consider as seroconverted?  Is that correct or not?  So does that mean that, on average, adolescents’ titer was 344?

DR. TSAI:  That’s correct.

DR. DEBOLD:  Is that a lot more than what we need?

DR. TSAI:  As was discussed earlier, we really don’t know what levels of antibodies are protective against H5N1 infection.  We have essentially extrapolated from the seasonal vaccine criteria, where a hemagglutination inhibition HI antibody titer 1 to 40 is considered to be protective.  There are some clinical data and other data that would support that level, that immune correlate.  But for H5N1 vaccine, there is very little data that I’m aware of that would inform identification of a threshold of protection.

DR. MODLIN:  That’s absolutely right.  And, of course, these titers don’t persist.  In other words, these are titers that are obtained within a month after completing the last dose.  Of course, there is going to be a decline in antibody titer, and so you have to factor that in as well when you are talking about seroprotection rates and correlates of protection.

Dr. Eickhoff?

DR. EICKHOFF:  Ted, you mentioned a lot about absence of safety signals, but you didn’t show us a whole lot of safety data.  Could you describe the surveillance protocol you used following receipt of one of the experimental vaccines?  Was it active or passive surveillance?  What did it take to generate a safety signal?

DR. TSAI:  I don’t know the details of those specific protocols, but, in general, in vaccine trials solicited adverse events are collected through diary cards.  Depending on the age of the subject, those events can be different.  For example, in the adolescents, you could ask about pain, whereas in infants the adverse event is tenderness.  Those adverse events would have been collected for some interval -- I believe it would be seven days -- after each vaccination.  Then other spontaneously occurring adverse events would be collected after each visit.  Serious adverse events that would be reported spontaneously would be entered into the database as well.

So, in general, in vaccine trials you have solicited events that you deliberately try to collect and then all the spontaneously adverse events also are collected.  Of course, among them, those that are defined as serious are analyzed separately.

DR. EICKHOFF:  If I can pursue that one point further, what did it take to generate a safety signal?  Some significant deviation from the appropriate control group?

DR. TSAI:  The data monitoring board that oversaw the pediatric trials was given a rather broad definition and ability to stop the trial, with any significant safety concern.  There were no stopping rules, for example.  Sometimes in a clinical trial you will specify a specific stopping rule.  There were no stopping rules.  But the overall database was examined, and the safety monitoring board would have had the liberty to interrupt the trial.

What I meant by safety signal, I was really referring, more so, to this database that has been collected, as well as to the pharmacovigilance database, which, admittedly, consists of spontaneously reported events.  So the pharmacovigilance database, comprising reports that cover the interval where 27 million doses were distributed, found no increases in specific adverse events that were examined -- GBS, other neurologic diseases that potentially could be of an autoimmune nature, and others.  I can’t remember the specific events.  But the incidence rates in that database did not exceed baseline expectations.

What we have with the data master file, however, is an ability to compare the safety experience of the MF59-exposed subjects with controls who received the same vaccine without the adjuvant.  This provides for a much more powerful way of looking at the safety of the adjuvant itself.

We are in discussions with the FDA about this analysis, so I don’t feel I can really comment at this time.  But I can say, our own analysis is very reassuring.

DR. MODLIN:  Jack?

DR. STAPLETON:  I was interested in following that a bit also.  In this clinical database, have you systematically looked at people who have received annual or repetitive doses of MF59?  If not, have you systematically done that in your Fluad recipients in Europe?

DR. TSAI:  There was a meta-analysis done.  Two thousand subjects received Fluad, and there were 1,500 comparators in the first year, and then smaller numbers in subsequent years.  The reactogenicity was followed after these three consecutive seasons.  In fact, whereas in the first year there were significant differences in local reactogenicity -- pain, erythema, induration -- those differences actually disappeared after the second year and third year.  So the difference in reactogenicity actually was reduced with a second dose and a third dose.

DR. STAPLETON:  I’m interested in whether you saw any signals toward autoimmune diseases, in particular?

DR. TSAI:  That wasn’t -- the object of this study was to look at acute reactogenicity.

DR. MODLIN:  Ted, the adjuvant produces extraordinarily high levels of antibody.  Are the heterologous responses independent of the titer of antibody?

DR. TSAI:  This is an interesting question.  Do we see these heterologous antibodies because a rising tide lifts all boats?  There is, in fact, a study -- I don’t know the details of it, but it’s a collaborative study that our group in Siena has with CBER scientists, in fact -- to look at that.  There is a suggestion that we are actually seeing reactivity to other epitopes that are not seen with an unadjuvanted vaccine.

DR. MODLIN:  It’s obvious that you are priming for heterologous antigen, which I think is critically important.  But I think that would also be an extraordinarily important issue to get at in terms of primary protection, particularly with one dose?

Roland?

DR. LEVANDOWSKI:  There’s another study that has been done looking at reimmunization many years later with two different H5N1 antigens.  I don’t think you mentioned that.  John Trainor (phonetic) for NIH did some studies.  In 1998, he did a study immunizing lab workers and others with the original Hong Kong 97 H5N1, the clade 0 vaccine, as Ted said.  In that study he was actually looking at dose ranging, so he had all sorts of unusual combinations of very low doses to very high doses of the antigen.  Some of those -- not all of those -- people were available eight years later and were immunized with Vietnam 1203 vaccine.  The first vaccine was prepared by Protein Sciences.  The Vietnam 1203 vaccine was prepared by Sanofi.  So they were totally different antigenically and also by the way they were manufactured.

The bottom line is that those individuals who had been immunized with Protein Sciences’ vaccine eight years earlier also had extremely unusually, we would say, high antibody responses after a single dose of vaccine.  Whereas the primary immunizations were getting geometric mean titers that, let’s say, were about 30, those who were reimmunized years later were getting responses that were about 60.

It’s not quite the same kind of study, but I think there is maybe a principle there that heterologous priming may occur even without addition of an adjuvant.

DR. MODLIN:  Dr. Klimov?

DR. KLIMOV:  Ted, just a technical question.  I believe that in slides like this one for seasonal vaccine, the HI titers being measured using turkey red blood cells or chicken red blood cells, while for the H5 vaccines the titer is being tested probably with the horse red blood cells.  This is correct?

DR. TSAI:  Yes, I think that’s correct.  I have tried to show you micronute data in most cases.  But, yes, in the HI cases it would have been horse erythrocytes.

DR. MODLIN:  Dr. Gilbert?

DR. GILBERT:  A comment.  The data on the heterotopic responses is useful, but my comment is, I find it hard to interpret the results without seeing some map relating the different target viruses to the vaccine virus.  There might be some interesting work to do in statistical methods of looking at the antigen relationships and displaying those in some way, to set an interpretation framework.

DR. TSAI:  That’s an important question.  I’m not involved in this research, but, as was mentioned yesterday, this antigen cartography allows for a very accessible way of understanding those antigenic relationships.  Derek Smith at Cambridge University has made such a map for the H5N1 clades.

DR. MODLIN:  Roland?

DR. LEVANDOWSKI:  One more question, which I guess I should have asked the other speaker as well, because it would probably apply to both situations.  A lot of the studies that are being described relate to comparing vaccine with and without the adjuvant.  The benefit is as important as the risk, I think, in order to figure out what the risk/benefit ratio is.  For those vaccines that have adjuvant in them, how difficult is it to do the standardization, to do the potency testing?  Are there any special tricks that are necessary?  When you are trying to compare one vaccine with another one -- they are not formulated exactly the same -- how much certainty do you have that you are actually comparing the same amount of antigen?

DR. TSAI:  I’m not sure I’m the right person to answer that question, to be truthful.  Maybe it’s a rhetorical question.

DR. MODLIN:  Maybe we can find somebody to answer you offline, if that’s okay, Roland.

Bob Daum and then Dr. Klimov.

DR. DAUM:  I don’t know if you are the right person to ask this question, but you said it, so I’ll ask you to clarify briefly.  If you could put the last slide up for just a moment, which starts with “Summary.”  The first statement, that children in the U.S. -- I’m paraphrasing -- are at risk for infection, which I happen to agree is an underlying assumption as to why we are sitting here -- it strikes me that we should then approve research under 50.52.  But you began your presentation by saying it is approvable under 50.54.  I don’t know if you appreciate the difference in bureaucratic intensity those two approaches involve.  So you may not be the right person to ask.  But they are huge.  I wonder if you would comment on that statement versus 50.54.

DR. TSAI:  I guess we were trying to take a more conservative approach, in the sense that there might be skeptics that H5N1 virus was not a direct threat, because we have had no isolations in the U.S.

But I personally do believe that there is certainly, as someone has said, a non-zero risk -- perhaps it was you -- and that children may, in fact, be at higher risk, for reasons that Dr. Vaughn mentioned.

I think I may have forgotten to comment on a point in my slide.  Children also may be at increased risk inherently because of the distribution of the cellular receptors for avian influenza viruses in the respiratory tract.  They seem to have a greater expression of the sialic acid alpha-2,3-galactose receptors in the respiratory tract compared to adults.  That could be an inherent biological risk factor for increased risk of avian influenza viral infections in children.

DR. KLIMOV:  Just coming back to two remarks before.  During the last big WHO discussion about the vaccine strain selection for the seasonal influenza, also the status of H5N1 vaccine development and vaccine strains was discussed.  This is going to be published, if it’s not published, very soon.

I have this data about the antigenic relationship between different clades.  If necessary, I can show this very briefly, if this will help the discussion.

DR. MODLIN:  Thank you.  Dr. Joffe?

DR. JOFFE:  I just wanted to make a comment that I could have raised with the previous speaker as well, and perhaps this will come up this afternoon.  In interpreting the safety or adverse-event data, the comparisons between the adjuvanted vaccines and the non-adjuvanted vaccines or, for example, the H5N1 vaccine versus Fluad, I find it difficult to interpret those comparisons, because, in a sense, the most interesting comparison, which I don’t think we have, is the H5N1 adjuvanted vaccine versus a true saline placebo, with respect to the adverse events.  What you find when there is an active control, as there is in, I think, most or all of the studies that have been presented, is that you get, potentially, an increased risk of adverse events in the control group, and then the question is, is the experimental vaccine in these comparisons elevated above that baseline?  But that, in a sense, is an artificially inflated baseline for many of the comparisons.

So thinking about study designs going forward -- and, again, perhaps this will come up this afternoon -- I wonder if, from the point of view of adverse events, the favored study design might not be a true placebo as opposed to some sort of active placebo, which I think makes it very hard to interpret adverse-event data.

DR. MODLIN:  That’s a great point, and I think we can discuss that this afternoon.  You can easily see arguments on either side of that.  There’s no question about that.

Other questions?

(No response)

If not, let’s go on to the next speaker.

Ted, thank you very much for a very informative presentation.

The final speaker before lunch will be Dr. Richard Gorman, from NIH.  Dr. Gorman is associate director for clinical research in the Division of Microbiology and Infectious Diseases at NIAID.

Agenda Item:  NIH Presentation

DR. GORMAN:  Good afternoon.

I’m well aware, after this discussion this morning, of the risks and benefits of being the last speaker before lunch, so I’ll try to be brief.

The NIH experience with pediatric H5N1 trials consists of two trials which I hope to discuss in some detail.

The first trial is a randomized, double-blind, placebo-controlled Phase I/II study of the safety, reactogenicity, and immunogenicity of intramuscular inactivated H5N1 vaccine in healthy children aged 2 through 9 years of age.  That is titled “DMID 04-077.”  In our nomenclature, “04” is the year in which the protocol is developed.

The second was an open-label study of intramuscular inactivated influenza vaccine in healthy children aged 2 to 10 years.  That was a follow-on study to the 04-077 study, where all human subjects were offered the opportunity to receive the study vaccine.

The rest of this discussion will center on 04-077, with some comparisons to adult trials with similar vaccine products.  All the products that we are going to be discussing today are unadjuvanted and all the same product.  The placebo that we used in ours is, in fact, a saline placebo.

The study design for 04-77 was a multicenter, with three participating centers.  The centers were in the University of Maryland, St. Louis University, and UCLA, all in the United States.  It was randomized five-to-one between the active agent and a placebo control, double blind.  The population was healthy children.  Children between the ages of 2 and 9 were enrolled in two strata, 2 to 5 and 6 to 9.  This study was designed in 2004, and the standard seasonal influenza recommendations for pediatrics did not extend below age 2 at that particular time.  That’s why these age ranges were chosen.

The vaccine was an inactivated subunit of the influenza A/Vietnam.  The dose, volume, and route:  45 µg of hemagglutinin in 0.5 mL given IM.  The manufacturer was Sanofi Pasteur of Swiftwater, Pennsylvania.

The procedures for this study were:  Two IM doses of the vaccine or placebo given one month apart.  The third dose was offered at six months to vaccine recipients only, and it was optional.  So all the participants got two vaccines and the people who got the study agent were offered the opportunity to get a third vaccine at that particular time.

Looking for reactogenicity, we had a memory aid for the first seven days, which was gone over by telephone on day 7.  There were clinic visits at regular intervals, both for the vaccines and the for the serologic draws during which reactogenicity and safety data was collected.  There were telephone follow-ups from day 56 until month 12, looking for serious adverse events.

For immunogenicity, we did sera for both HI and MN, both before the vaccinations and one month after each dose.

For the immunogenicity data set, there were 125 children enrolled, 23 into the placebo arm and 102 into the vaccine.  One hundred seventeen received two doses.  There were 113 subjects with all sera available, 21 placebo and 92 vaccines.

The first question in terms of safety and reactogenicity:  Was the dropout rate different in the placebo and the vaccine group?  The answer was no.  The ratio of five-to-one maintained in the dropouts as well.

The booster dose at six months:  Fifty-eight vaccine recipients were boosted, and 55 of those had sera at 28 days.

The gender of this population was 54 percent boys, 46 girls; 89 percent Caucasians.  The median age in the study was 6 years of age.  The cohort between 2 and 5 had 61 subjects; the cohort between 6 and 9 years of age had 52 subjects.

There were two serious adverse events.  Both were deemed unrelated.  One was a case of rotavirus diarrhea 13 days after second vaccination.  There were other family members ill, and a rotavirus ELISA test from the stool was positive.

The second case was a case of rat bite fever.  It was 55 days after the second vaccination.  It was purportedly due to Julie, who was the newly purchased pet rat.  The subject was hospitalized for persistent high fever and received IV antibiotics.  Other immunological tests looking for other diseases all turned out to be negative, except for rat bite.

Continuing on with the safety drill-down, there were 141 adverse events:  61 percent were mild, 12 were deemed related; 37 were moderate, two deemed related; and 2 percent, or three total events, were related as severe.  All of those were considered to be unrelated.  Of the three events in the severe group, two were in the placebo control and one was in the study group, the active vaccine group.  The one that was in the study group was an intercurrent non-influenza-like illness that also shows up in our reactogenicity data.

Drilling down further to the reactogenicity of this antigen, there were no fevers over 103.  There was one mild and two moderate after dose 1, two moderate after dose 2, and one moderate after dose 3.  There were no severe reports of injection-site pain.  The redness was mostly mild, up to 20 mm.

I’m going to show the reactogenicity data in two different ways, first this way, in a graphic form, and then in a more visual form, for those of you who are graphic learners versus those of you who are visual learners.

In this particular slide, we see the reactogenicity data after dose 1 for all of the reactogenicity events that exceeded 5 percent.  If you look at the slide, we have the vaccine participants versus the placebo.  In the white are the total number of any severity of reports during the first seven days and in the red are any that are moderate or severe.

Not surprisingly, injection-site pain differs between the two groups, and perhaps decreased activity.  But other than that, the groups look remarkably similar.

After dose 2, the results look remarkably the same — again, decreased activity, again the gap narrowing, injection-site pain continuing to show a difference between the vaccine and the placebo.

After dose 3, there were no placebos, because this was only given to the vaccine recipients, and these were their responses.

For the visual learners, we look at it this way after dose 1.  On the y-axis you see the percent of all who complained.  Inside the graphs you can see severe, moderate, and mild, in three different color representations.

After dose 2.

Then after dose 3.

As one of the considerations this group is going to have over the rest of this session is the risk of this group versus other groups, I thought it might be useful to show the parallel nature of this versus some of our adult data.

This is a slide comparing the pediatric and adult reactogenicity of children and adults who receive the same antigen in the same dose, again unadjuvanted.  You can see that the panels look very similar, except for elevated temperature, which appears on the pediatric side but not on the adult side.  But when you use 45 µg in both pediatrics and adults, the reactogenicity post-vaccination looks remarkably the same.

When you go up to 90 µg, the dose approved for adults, you will notice that the charts now look slightly different, with adults complaining of a lot more pain, which is consistent with the data that we have received over the years that the more antigen you put into a vaccine, the more pain there is at the local injection site.  But if you look at the systemic signs of elevated temperature and body aches, they remain the same in both groups as well.

Talking a little bit about the efficacy or the immunogenicity of this particular vaccine, we would like to look at the microneutralization results.  In the blue boxes you will see the placebo results.  There is no change in microneutralization; in the brown boxes, ages 2 to 5, fourfold increases after doses 1, 2, and 3; and in the yellow boxes, the microneutralization increase after doses 1, 2, and 3, for ages 6 to 9.

The hemagglutination inhibition results:  Again, you see the same sort of results, with the placebo having no increase; ages 2 to 5, after 1, 2, and 3, going from 6 to 53; and the age group 6 to 9 going from 41 to 64 percent.

Again, because this group is going to be looking at and talking about the risks for pediatrics, as well as the benefits for these types of studies, we wanted to show you the results of this particular study, 04-077, versus two studies that we have done in adults, one with healthy adults and one with stable-health-condition elderly, all receiving the same doses.

In this particular time, post-vaccination 1, there was 7 percent increase in the peds, 21 percent increase in healthy adults, and 7 percent in the elderly; post-vaccination 2, 38, 33, and 23; and post-vaccination 3, 58, 38, and 17.

This is looking at the same dose, looking at the dose that is presently approved in adults.  We have gone from the 45 µg that we have used in pediatric subjects to the 90 µg that we have used in adults.  You can see that the initial response after dose 1 is better for the adults, with pediatric subjects only receiving a 7 percent increase, while adults receive 23 or 25.  But by vaccination 3, the resulting increase in hemagglutinin inhibition looks pretty similar across the three groups.

In summary, two IM doses of 45 µg of unadjuvanted inactivated H5N1 vaccine in children is well tolerated, leads to immune response comparable to adult responses, and there is a slightly better response in 6- to 9-year-olds compared to 2- to 5-year-olds.

Thank you.  I’m available for questions.

DR. MODLIN:  Terrific.  Dr. Gorman, thank you very much.

Are there questions for Dr. Gorman?  Yes, Dr. Klimov?

DR. KLIMOV:  Again, this technical question.  The slide on page 20, HAI data, is this with use of horse red blood cells or turkey red blood cells?

DR. GORMAN:  I can’t answer that question because I wasn’t involved in the assay work, but I can provide that data for you before the end of the meeting.

DR. MODLIN:  Ann says it’s horse.  Dr. Joffe?

DR. JOFFE:  The various IRBs that approved this protocol -- do you happen to know what subparts they were approved under?

DR. GORMAN:  The communication from IRBs goes to the principal investigators.  The essential documents that we collect inside of NIH are that they were approved.  I perhaps could go back and get that information.  I do not know that off the top of my head.

DR. JOFFE:  Just to follow up, was this approved by NIH IRB or was it all local IRBs?

DR. GORMAN:  This was all local IRBs.

DR. DAUM:  But isn’t it true that if the IRBs approved, it wasn’t 50.54?

DR. GORMAN:  That is correct.  They were approved at the local level.

DR. JACKSON:  Just a point of clarification.  On these slides, does this mean that the measure of a post-vaccination titer greater than 1-to-40 is the same as the definition of a fourfold rise?  The footnote indicates post-vaccination titer greater than 1-to-40, but the title is “Fourfold Rise.”

DR. GORMAN:  They had to meet both criteria.  They had to have a fourfold rise and the final titer had to be greater than 1-to-40.

DR. MODLIN:  But the data in the boxes are fourfold rise?

DR. GORMAN:  That is correct.

DR. MODLIN:  Other questions?

(No response)

If not, Dr. Gorman, thank you very much.  We certainly appreciate the presentation.

We have a lot of work to do this afternoon in a relatively short period of time, so I’m going to ask all the committee members to return on time, ready to begin the discussion.  We will start at 1:15 on the dot.

(Whereupon, at 12:15 p.m., the meeting was recessed for lunch.)

AFTERNOON SESSION

DR. MODLIN:  Good afternoon.  I would like to reopen the meeting.

I know that a number of committee members and others have flights this afternoon.  Even though the agenda says that we are going to 4:00, we have a target finishing time of 3:00.  I’m hoping that we will be able to conclude by then.  However, we do have a lot of work to do, so I’m going to ask people to try to recognize that reality.

The next item on the agenda is the open public hearing.  I understand we have at least one individual who has signed up for the public hearing.  Before starting that, I need to read the following statement.

Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making.  To ensure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is important to understand the context of an individual’s presentation.  For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or your oral statement to advise the committee of any financial relationship that you may have with the sponsor, its product, and, if known, its direct competitors.  For example, this financial information may include the sponsor’s payment for your travel, lodging, or other expenses in connection with your attendance at this 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.

I understand that we have at least one speaker who has signed up, Dr. Mary Kathryn Reeves-Hoche.  Dr. Reeves-Hoche, would you like to begin?

Agenda Item:  Open Public Hearing

DR. REEVES-HOCHE:  Thank you, Dr. Modlin, committee members.

I’m Dr. Mary Kate Reeves-Hoche.  I work for Sanofi Pasteur, where I am the director of the pandemic influenza program in the R&D stage.  So that’s my financial disclosure.

Sanofi Pasteur is very proud to have partnered with HHS and the DMID of NIH in licensing the first H5N1 influenza vaccine for the United States in April of 2007.  Since that time, our company has moved on to test lower dosages of antigen with different adjuvants -- namely, aluminum salts and our proprietary oil-and-water emulsion adjuvant.  Our vaccine has been tested in a stepwise fashion, first unadjuvanted, as already presented by Dr. Gorman from the NIH, and then with the aluminum adjuvants, and now we are testing with our novel adjuvant.

We have had great success in our adult program using this novel adjuvant and have achieved a very high immune response at the lowest dosages of H5N1 antigen reported to date.  A vaccine containing only 1.9 µg of antigen generated a high level of seroprotective immune response in over 70 percent of the participants in a clinical trial.  In the same clinical trial, vaccine containing 3.75 µg of antigen generated a high level of seroprotective immune response in over 80 percent of the participants.  Again, these were adults.

Based on this experience, we are advancing our clinical program into the elderly and into pediatric populations.  In fact, as already mentioned at this meeting, we have an ongoing pediatric trial in Thailand using H5N1 plus the aluminum adjuvant.

We recognize the vital importance of pediatric populations and the importance of protecting that population, but we also recognize that they are considered a special population, with additional regulatory safeguards.  Sanofi Pasteur believes that by working with committees such as this, as well as working with the FDA, we can safely evaluate H5N1 vaccines in pediatric populations.

We would like to acknowledge the support that we have received from the U.S. government in our collaborations with the NIH, as well as our collaboration with other manufacturers.

Thank you.

DR. MODLIN:  Thank you, Dr. Reeves-Hoche.

Let me ask if any of the committee members have questions for you.

(No response)

If not, thank you very much.

Is there anyone else who wishes to make a statement in this portion of the meeting?  Yes?

DR. CHU:  Thank you.  My name is Susan Chu.  I’m a physician by training.  Right now I’m contributing editor of Flu Wiki, which is an online Web site on pandemic flu.  I’m also a cofounder and president of Ready Moms Alliance, a nonprofit, grassroots effort for promoting pandemic preparedness.

I don’t have any connection with any of the manufacturers in the room or any in industry in general.

I want to, first of all, thank the committee and everyone who has worked so hard on it.  Coming from where I’m coming from, I’m fully cognizant of the threat of a pandemic, and the threat to our children.  Right now, as we speak, in San Diego, Ready Moms Alliance is presenting in an exhibit the issue of the risk to children of H5N1 because of the high mortality.  As far as I’m concerned, I think that we ought to go as quickly as we can, in any way we can, to promote any studies that would lead to a pandemic vaccine being available, if and when the pandemic breaks out.

Having said all that, when I listen to the subtext that’s going on, I think the biggest concern is the safety, specifically with regard to adjuvants.  Instead of going to the details, I want to take a 30,000-foot view of that and ask questions.

When you put an adjuvant and an antigen into a child, the reason why you would put that would be because the antigen is a little bit too weak to produce the immune response that you want.  But at the same time, when you put that in a child, the child’s body also has a whole lot of auto-antigens going around that are weak, that are not causing problems, that are being suppressed into producing disease because of tolerance.

I’m not a research scientist on that aspect.  I don’t know whether, particularly with broad-acting adjuvants that act across different vaccines, their effect on the antigen that you are putting in, the vaccine antigen, would also have the same effect on the auto-antigens in the child’s body.

There is a second level to that question, when we look at adverse-event reporting.  If you give a child a vaccine, how do you know he or she is protected?  We don’t know until you expose the child to the disease.  But if you don’t have that, what you do is, you monitor for antibodies.  But in a case of auto-antigens and similar kinds of induced response, what we are doing right now that I’m observing is waiting for much further downstream, when they actually have symptoms, for them to report back.

What I’m looking at is, there is a disconnect.  The parallel of the antibody testing for vaccine antigen does not exist -- we don’t have the ability yet to test accurately whether the child is having those kinds of responses.  Before we have that ability, I would say, on the one hand, it’s good to have vaccines as quickly as possible, but on the other hand, I would say, go slow, because we don’t know what we are looking at.

The other part of the story is, I would suggest that it is very important to tell the public where the areas of uncertainty are, what it is that you don’t know, what it is that you can’t work on.  My experience of talking to the public online for three years is that the vast majority will understand complex decisions, where you can’t have the perfect solution.  You have to tell them and explain to them, and they can get it.

Thank you.

DR. MODLIN:  Thank you, Dr. Chu.  Perhaps during the next hour or so, we can address a couple of the issues that you raised.

Are there any other members that wish to speak during this portion of the meeting?

(No response)

If not, we will go on to the committee discussion and recommendations to the agency.

Agenda Item:  Committee Discussion and Recommendations

I would like to lead off by reminding us -- we have heard a terrific, very thorough, careful explanation of the regulatory and the ethical frameworks for our discussion today -- I want to remind the committee that our purpose here is really not to discuss the ethical issues per se.  Rather, we are constituted as a technical advisory committee.  I think the agency would prefer that we devote most of our attention to the technical issues, overlaid over the ethical framework that we discussed earlier this morning.  That’s really our purpose here, and I’m going to try to keep the discussion focused in that way, if at all possible.

The questions that we have presented to us, I think, help in that respect.  It may very well be that the best way to begin our discussion would be focusing on the first question here:  Please discuss whether clinical studies in one more pediatric age groups should be conducted using inactivated pandemic influenza vaccine candidate as part of pandemic preparedness.

Of course, in my opinion, one of the issues at the very heart of this question is the issue that Dr. Daum raised earlier, which is, just what is the risk of pandemic influenza in a population of U.S. children?

I don’t mean to say that this is where we have to start, but I think it might be a reasonable place to start.  So I’m going to open this up for committee discussion.

Bob, do you want to start off with any further comments around that issue?  I think you made your point clear -- oh, one other thing I do need to say is that we will not be voting on any of these issues.  It’s really the discussion part that is, I think, the critically important part of the meeting this afternoon.

Bob?

DR. DAUM:  I can just reiterate in one sentence.  I think that at some point an influenza pandemic is going to occur.  There is no information about what form it will take, when it will come, or what virus will be the perpetrator of it.  But I believe very strongly that when it does happen, children will be involved in helping to spread the virus and be involved in the epidemic, if that’s the right word to use.  Therefore, in thinking of preparedness and plans to get ready, I believe very strongly that children need to be included in understanding their immune responses to vaccines and preventive measures that we as a scientific community want to develop.

I guess I don’t understand a reason to exclude them.  I don’t see a rationale for directly extrapolating from adults information on what to do with dosing and regimens in children.

Beyond that, I think I have made my point clear, and I’ll stop talking.  But I’m a strong proponent that children should be included in influenza preparedness, by understanding their responses to the vaccines we heard about today.

DR. MODLIN:  Dr. Jackson?

DR. JACKSON:  I agree.  I think it could help us to consider that the question could read “should continue to be conducted,” since studies of pandemic vaccines have already been conducted in children in both the U.S. and elsewhere.  The critical issue seems to me to be the degree to which you can extrapolate from older age groups to pediatric age groups.  I don’t think you probably can for these vaccines, and so we need pediatric-specific data in order to develop vaccines and schedules that will be most likely to protect the pediatric population.

DR. MODLIN:  Ted?

DR. EICKHOFF:  I totally subscribe to Bob Daum’s point of view, as expanded on by Lisa Jackson.  I think there’s no other reason we are here today.  We are worried about children.  History teaches us over and over again that children are at risk if there is going to be a pandemic.  We can argue about whether it’s going to be H5N1 or something else, but that’s really a nonproductive argument at this point.

DR. MODLIN:  Pablo, do you feel any differently?

DR. SANCHEZ:  No.  I absolutely agree.  I think as pediatricians, we always complain that we have to treat our infants and children based on adult data.  We have to look at it in pediatrics.

DR. MODLIN:  Dr. Gilbert?

DR. GILBERT:  I agree with everything that has been said.  Nothing more to add.

DR. MODLIN:  Pam?

DR. MCINNES:  I completely support it.  I think, in addition, we have to have a learning curve as the data are derived, in terms of understanding how we can perhaps start collapsing age populations, et cetera, and not assume that everybody will march down exactly the same path.  But we have to learn from the data as they accumulate.  I really think they need to be extremely robust in order to build a body of safety and immunogenicity data that we have a lot of confidence in.

I don’t know if you are going to get in the next question, but I have strong feelings about how rigorous the informed-consent process should be.

DR. MODLIN:  We can come back to that.  José?

DR. ROMERO:  I concur with everything that has been said.  I think this is a unique opportunity for us to really, as pediatricians, serve as the voice of our population of medical interest.  I think this is an opportunity to really move this vaccine development forward.

It’s intriguing.  There are lots of avenues that we can talk about.

DR. MODLIN:  Seth?

DR. HETHERINGTON:  I agree completely.  I think a question that might be raised -- maybe not today, but in the future -- is, how much information is going to be needed to feel comfortable with the use in children?  Just looking at some of the numbers that we have, we have the DMID study of 125 kids and the Novartis study of 472-plus kids.

Just to put it a little bit in context, if you have 300 patients’ worth of data over a period of time that you think is adequate and there are no serious adverse events, then you have a 90 percent certainty that any serious adverse event is going to be less common than 1 percent.  If you look at the other side of the coin, the risk that a pandemic would present, 10 to 40 percent infectivity among kids with a 60 percent mortality, it doesn’t take a whole lot of calculation to figure out that you are well into the benefit-plus column of doing this.

I think the key point, perhaps, is that we will have data on kids down to 6 months of age.  A question that should be raised is, do you need to go even younger than that?  I think some of the epidemiology we saw earlier is that under 6 months is a very at-risk population.

The usual limitation we have on seasonal flu vaccine, going down to 6 months, I think assumes that under 6 months you are protected by maternal antibody to some degree.  But I’m not certain that we have any information that says the same will be true with an H5N1 virus.  So we should give some serious consideration to pushing that age limit down to as low as possible, to cover the most susceptible patients.

DR. MODLIN:  That’s an excellent point.

Vicky?

DR. DEBOLD:  I think this question as stated is sort of -- the trials are already happening, so it’s not a matter of saying you do or don’t agree with them.

I think I’m less enthusiastic than what I have heard so far.  I’m not convinced that the information that I have about specific risks to American children, at this particular point in time, is compelling enough to go into large-scale trials with children, absent doing some of the preliminary early things that, for me as a member of the public, I feel are missing.  It seems like much of the discussion today has centered around a tension between the need for production efficiency versus the potential for creating unintended and potentially unnecessary autoimmune problems in some susceptible individuals.

I think we have to be very careful, because all of this is being played out now in an arena of challenged, I think, public trust as it relates to vaccination.  It’s very important that whatever is done here is done correctly.  If there is no pandemic and if it doesn’t ever come, but yet we have many sick children, there will be an extremely high cost to having made that choice.

DR. MODLIN:  Norm?

DR. FOST:  There are two separate questions on the screen.  One is whether or not a pandemic is likely ‑‑ or implicit questions.  I don’t have any expertise on that.

But the next question is, if a pandemic occurred in the U.S., would it be better to just vaccinate children based on adult data or would it be better to do studies?  I think it’s clear that we can’t wait until the pandemic starts.  I think we have seen evidence that it’s too late at that point to start studying it.  Therefore, if it’s going to be studied, it has to be studied before the pandemic comes.

I think as a general matter it’s preferable to study things systematically than to just give kids things off-label and hope for the best and see what happens.  I think we have too much history of things gone haywire that way.

So it all hinges on the likelihood of a pandemic.  If that likelihood was zero or close to zero, then doing studies would be exposing children to risk with no benefit, because nothing would ever come to help them.  If we thought it was an extremely high likelihood of a pandemic, then it seems to me self-evident that we need to do these studies now, before it comes.

So it seems to me the whole thing hinges on the likelihood of a pandemic, which is outside my area of expertise.  But as an informed lay person reading the literature, it seems to me likely enough and that the risks of doing these studies are low enough that it’s appropriate to do these studies now, before the pandemic comes.

DR. MODLIN:  Maybe I could just address a point there, since I’m next in line.  I was talking with Melinda Wharton and others during the break this morning.  This is, in some respects, déjà vu all over again, with respect to discussions that we had at the CDC and the ACIP around recommendations for smallpox vaccine about six or seven years ago.

I was told this morning that testing the smallpox vaccine in children does not create a precedent for this discussion that we are having today.  But I do want to point out that, as Bob and others have said, the issue of whether a pandemic will occur or not is really less of an issue.  It’s not a matter of if it will occur; it’s a matter of when.  That’s the lesson that history has told us.

With the smallpox discussion, it really was a theoretical discussion.  Nobody really knew whether there was any risk whatsoever, which made the decision making that much more difficult.

But I would like to point out that if we had an outbreak of smallpox, as bad as it would be, an outbreak of smallpox would be far easier to control and probably cause far less morbidity and mortality than, say, would an outbreak of a new strain of influenza virus, without a doubt.  We know that we can control an outbreak of smallpox much quicker and much easier than we can an influenza pandemic.  So in some respects, that’s where the comparison breaks down.  I think pandemic influenza is likely to be a more serious disease than smallpox may be.

I personally would agree.  I think it’s essential to do these studies, because children probably would be harmed in such a case if we didn’t have the information that would come from clinical trials, if we needed to apply a vaccine suddenly that we didn’t know how to use.

I’ll leave it at that.  I’ll pass things on to Dr. Joffe.

DR. JOFFE:  I’m inclined to agree with the majority sentiment that has been expressed so far that one ought to go ahead with these studies.  But there’s an issue that I think requires further discussion.  Part of coming to that conclusion, I think, is concluding that one can’t validly extrapolate from adult data to pediatric data.  That has sort of been asserted in the discussion thus far, but it hasn’t really been discussed in any detail.  I don’t claim any expertise to discuss it.  It’s certainly not my area of expertise.  But I think it is worth a discussion of why it is that we can’t validly extrapolate from adult to pediatric data.  I didn’t see in the presentations before any great differences between adult immunogenicity or safety outcomes compared with the pediatric outcomes that were presented.

I’m not trying to argue that one can’t.  I just think it’s an important issue that needs discussion, and I would like to hear more about it.

The other thing that I just want to raise briefly comes to the point of whether there ought to be studies in children younger than 6 months.  Thinking about the pre-pandemic situation, it seems to me not necessary -- I’m thinking ahead, if a recommendation might come out for pre-pandemic population-wide vaccination, based on studies that have been done -- it seems to me unlikely that one would require pre-pandemic vaccination in very young infants, because there would be very little cost in this elective setting to saying the recommendation could be to wait until they are 6 months old and then vaccinate them.  So I don’t know that one needs data in the pre-pandemic setting on vaccinating kids that are younger than some minimum age, whether it’s 6 months or something.

On the other hand, there might be a case to be made for studying those vaccines in young infants, thinking ahead to the pandemic setting, as opposed to the pre-pandemic setting.

DR. MODLIN:  Bruce?

DR. GELLIN:  In some ways, being in this position is easier because a lot has been said or harder, to try to identify some other issues.

We spend a lot of time talking about preparedness.  It’s interesting that the question is framed in the context of pandemic preparedness.  Bill Robb, who recently retired from the Department after 45 years, often defined preparedness as what you want in place the day before the pandemic.  I think that’s both in terms of what stuff you want, what you want people to know, and what other knowledge you have.

I think, in that context, we clearly know that just-in-time clinical studies are not going to give us the data we need, so getting ahead of that curve and knowing what we need to know ahead of time is really very important.

The other piece of this is that a number of people highlighted in their presentations the work that HHS, with Homeland Security -- and, frankly, across government -- did to try to look at the pandemic vaccine-allocation scheme.  It’s worth highlighting that the scheme that was presented is for severe pandemic.  There are slight differences in prioritization if the pandemic is deemed not to be quite as severe.

Nevertheless, when we took this issue to the public and tried to take the pulse of the public about what they felt, it was clear that an important value to the public was protecting children.  Protecting children is what this discussion is about, which is why I want to reinforce what Pamela said.  We really need to know this information.  We have to do it methodically, cautiously, and comprehensively, because we want to make sure that we are doing what the public really sees as an important value here.

Not to confuse us with a separate discussion -- it’s the one that Steven just raised -- we do have a stockpile of vaccine.  There was some mention made of that.  We continue to accrue vaccine into that stockpile.  The current plan is, when a pandemic is imminent, it would be used.

But there is a separate discussion that we are having, that the World Health Organization is having, and that essentially every country with a stockpile is having on whether or not you would use it before.

Again, if you start thinking about the size of a stockpile that could be more broad than select populations, such as occupational groups, this is information that we are going to need to know.  But again, as we said, it just needs to be obtained in a cautious and comprehensive way.

DR. MODLIN:  Thank you.  Roland?

DR. LEVANDOWSKI:  I would like to give an example that is a rationale for thinking that what happens in children may not be the same as what happens in adults.  Although there may be instances where everybody responds just the same -- H5 might be one of those where everybody who is immunologically naïve looks like they respond to the vaccine.  Perhaps there are subtle differences in local reactions and so on, but in terms of immunogenicity they look pretty similar.

But if you go to some other example -- I don’t know whether you consider it a pandemic or not -- when H1 reappeared, it was very obvious that there differences in immune responses between those who were old enough to have been exposed to H1 previously and those who were not.  It didn’t separate out into children versus adults, but it separated out into people who were younger than about 25 years of age and older.  Those who were younger required two doses of vaccine to get a tiny immune response, as compared to those who were older, who only needed one dose to get a very robust immune response.

We know very little about the other influenza A subtypes.  I don’t think we should forget that those are as important for pandemic preparedness as H5 and the other things that we have been discussing this morning.

I think that would be something to keep in mind, that we really don’t know exactly what to predict.  It might turn out that a lot of them are like H5, where everybody is naïve and the same.  But there could be some unexpected cross-reactivities that would in some way impact on how we try to approach things.  And we couldn’t know that without doing the studies.

In addition to that point, as Dr. McInnes raised yesterday about the opportunity to optimize, I don’t think we have fully optimized use of inactivated influenza vaccine in pediatric populations, in the first place.  I think there is a lot that we don’t know about the dose that would be optimal.  Again, it goes back to having not a lot of supply when these decisions were made.

The 6-month cutoff was somewhat arbitrary.  Studies were done that were set up to cut off at age 6 months.  That’s what information was available; that’s what FDA could support in the licenses that were administered.  If the studies had started at 2 months, I suppose, we would have a vaccine that we would consider for 2 months.

There are other questions that relate to whether younger children really respond effectively to inactivated vaccine.  There is that kind of discussion.  I’m not negating that at all.

Again, we have an opportunity to try to do something more.

The last point is that, whether it’s pandemic or interpandemic, influenza doesn’t really make that distinction.  The viruses are out there circulating somewhere.  Although we are compartmentalizing them, they really overlap a lot.  A lot of the information that comes out of the studies that are done for pandemic can help to inform how the interpandemic vaccines are used.  We may be able to improve utility of vaccines for pediatric populations.

In the other direction, too, if those optimization studies were done, in some sense, for the current seasonal vaccines, that could help in developing safer and more effective vaccines for use in children.

DR. MODLIN:  Thanks, Roland.  Jack?

DR. STAPLETON:  I would like to reiterate what you said, John.  This is different, I think, than smallpox.  There is H5N1 disease in the world.  The question is, will it hit North America and will it be the pandemic?  Given that there will be a pandemic -- or there will be a severe epidemic, at least, as it drifts -- if it does turn out to be H5N1, this will provide direct benefit to the children, which is something that we often don’t have the option to talk about in all these discussions.  If it isn’t H5N1, it will at least provide dosing of novel hemagglutinin types that will provide us some information that may help with whatever does evolve.

The second issue that we will, I’m sure, discuss a lot more is that the poor immunogenicity of H5N1 raises the issue that you probably are going to need an adjuvant.  That’s new for flu in the States, at least.  I think that’s something that we will have a lot more discussion about.  But it complicates the discussion.

DR. MODLIN:  Frank?

DR. DESTEFANO:  I agree with everyone that the possibility of a pandemic is not negligible, so I’m in favor of proceeding with such studies.  I would encourage studies that would focus, as some of the ones we have heard about, on either vaccines or strategies that will have as broad a protection as possible.  I don’t think we can predict when the pandemic will occur.  We are not sure it will be H5N1.  So strategies that will increase protection, whether through additional antigens or adjuvants or whatever, I would promote.

I would like to focus most of my comments on the safety concerns, the theoretical ones that have been raised about autoimmune disease, particularly from the newer adjuvants.  I would say it’s not likely that we will get good information on that until after licensure.  Even 25,000 adults that have been enrolled in some of these trials are probably not adequate to really evaluate these pretty rare conditions.

I support Dr. Chu’s suggestion of investigating the possibility of subclinical markers.  I think maybe someone ought to give some thought to that, if there are available markers that could be included in some of these clinical studies.

Also I think it’s worth looking at some of the data we have, particularly on the ASO adjuvant that has been used in 20 million adults in Europe.  I don’t know if they have been published, but I would like to see an analysis of what the pharmacovigilance data has shown, I would think -- the MF59 or the ASO, whichever one is in the Fluad.  I think that’s the one that has been used in Europe.

Anyway, if there are 20 million doses that have been used, I think the pharmacovigilance data on that could be revealing.  We have seen that, although it’s voluntary reporting, the experience in the U.S. is that serious autoimmune diseases, like GBS, can be detected through a voluntary reporting pharmacovigilance system.

In children, I think we ought to take a closer look at influenza vaccine in general, if that has any associated risks, along the lines of neurologic or autoimmune diseases.  There have been case reports of things like ADEM, acute disseminated encephalomyelitis.  The early studies have been done of the TIV vaccine from VAERS.  They were done early on.  There may not have been enough doses of the vaccine to really evaluate those.  I think, now that we have had more years of use in children ‑‑ I don’t know what the dosage may have been, probably in the millions now -- it could be worth revisiting the VAERS data and VSD data to look at these issues.

DR. MODLIN:  Thanks, Frank.  Melinda?

DR. WHARTON:  I do support the continued performance of pediatric studies looking at these influenza vaccines.  Just building on some of the comments of my colleagues around the table, I do think we have a chance to do it in a thoughtful and careful way -- I’m sure we will get into this subsequently in the conversation -- to take things one step at a time and in a thoughtful, reflective way, and particularly as we are thinking studies using adjuvants about which we don’t have as much experience, to have that process be a staged one that is really based on careful consideration of the basic science and preclinical data, the prior experience in adults, and use of other vaccines -- in a very thoughtful, reflective way, before we take those pediatric studies.

DR. MODLIN:  Thank you, everyone.

I think we have identified a number of questions that we need to come back to -- I think particularly Dr. Jost’s question about when data in adults is adequate, which we can do.  But I think probably our time is best served if we go on with the discussion.

If the recommendation is that studies should be conducted, please discuss your recommendations regarding ‑‑ and the first sub-bullet here is, which pediatric subpopulations should be considered.

I did hear Seth raising the possibility of infants under 6 months of age.  The obvious reason is that in a pandemic situation the absence of passive acquired immunity could be very important.  The disease could be quite different in that population than we currently see with seasonal influenza.

Another subpopulation to consider would be infants and children who have higher than normal risk from influenza of hospitalization, particularly children with chronic cardiac and pulmonary disease, children with chronic neurologic conditions, and these sorts of underlying health problems that we know put them at higher risk from influenza morbidity.

There are probably other populations that I can’t think of right now.  I think those might be important ones.

Rather than go around, why don’t we just open up this question for discussion about which pediatric subpopulations would be important.  Seth?

DR. HETHERINGTON:  Could I just ask a question first?  That is, how do people think of use of a pandemic vaccine -- in other words, if a case has been identified or a cluster has been identified, do people think in terms of just wholesale immunization of the U.S. population, or are we talking about geographically targeting an area first, including first responders?  That may affect how you approach some of these other questions.

DR. MODLIN:  Let’s let Bruce Gellin address that.  He’s the man of the hour here.

We heard some of that from the manufacturers today, about priorities for pandemic --

DR. HETHERINGTON:  Priority in terms of populations, but how about prioritization in terms of geographical location?  How do you perceive this?

DR. GELLIN:  Maybe some of the folks in the flu group have a better descriptor of this.  Often, when you look at seasonal influenza, it’s not like the weather map, where you watch something start in the West and work its way East.  The description is more like popcorn.  Once it’s around, it will show up in various places.  The assumption is that while there would be an effort, probably led by the World Health Organization to try -- I think containment may be optimistic -- to try to slow things down at the source, were that not to happen, then it’s just assumed that it will spread across a continent and across the world.

Therefore, efforts to try to hold this off in some area I don’t think are part of the current thinking, other than initially trying to slow something down in its tracks.

Is that your question?

DR. HETHERINGTON:  Yes.  Attached to that, then, is, what’s enough of a signal to trigger wholesale immunization of the population?  Is it a single case?  Is it multiple cases?  If so, how many?  What triggers immunizing everybody?

DR. GELLIN:  The World Health Organization has a series of phases.  We are currently in what is referred to as a pandemic alert, which is Phase 3.  The graduation of the stages will go when there is evidence of efficient human-to-human spread.  It’s a combination of the epidemiology and perhaps supplemented by some of the virological changes.  Essentially, the lead would be the World Health Organization to declare a change in the phase of the pandemic alert, where we are now, and that then triggers a whole cascade of things, both in the United States and other countries.

DR. MODLIN:  José, you have a comment?

DR. ROMERO:  In my previous life I worked as one of the physicians in the Public Health Department in Omaha.  One of the things that’s very important to keep in mind is the logistics of immunizing large numbers of individuals.  This can’t be done in a day, two days.  It has to be done over time.  There will be an early warning, but, as you said, there will be these indicators of when to start immunizing.  You don’t want to be trying to catch up.  You want to be ahead of the curve on that one.

DR. GELLIN:  Just to highlight, if the product requires two doses, then you have to build in not only the logistics time, but the time for the immune system to catch up to what it’s supposed to be doing.

DR. MODLIN:  Yet another reason to study it ahead of time.

Let’s get back to this question, though.  Are there subpopulations that should be studied, at least on a pre-pandemic basis?  How about kids under 6 months of age?  Pablo, do you want to address that?

DR. SANCHEZ:  I definitely think that it should be.  I think one of the problems currently is the fact that we don’t immunize those infants.  There is some data to suggest that they may respond -- unpublished.  But I think that relying on the fact that they may have maternal antibody is saying that that pregnant woman was immunized late in the pregnancy with the same vaccine.  We just can’t assume that in a pandemic setting.  So I really think that it should be taken gradually to the young infants, less than 6 months of age.

DR. MODLIN:  So you would support studies in younger infants.

DR. SANCHEZ:  Yes.

DR. MODLIN:  Jack?

DR. STAPLETON:  I think that’s definitely an age group where you can argue that you can’t use adult data to extrapolate.

DR. MODLIN:  And who may benefit the most, when you think about it, from what you learn, for a couple of different reasons.

DR. GELLIN:  And also to recognize what’s available from the medicine chest, antivirals are not currently available for children that age either.

This does raise the issue about pregnancy.  It’s not a subpopulation of children, but -- it has come up a couple of times here -- there is the question about maternal antibody and what impact it may have.  I think the kinetics of that are study-able -- obviously, not the impact of vaccinating pregnant women, but it will highlight that, while this meeting is all about children, when you look at that list, pregnant women are also at the top of the list.  In past pandemics, they have probably fared the worst.

DR. MODLIN:  Good points.  José?

DR. ROMERO:  A comment about antivirals.  There is a study going on right now looking at dosing kinetics in kids down to 2 months of age.  So that data will be coming forward.

Another group, I think, John, from a global perspective, is HIV-infected children.  That is a group worldwide that we need to look, because they, too, will be at very high risk for this.

DR. MODLIN:  Lisa?

DR. JACKSON:  Getting back to your earlier question about subgroups as chronic cardiovascular disease, pulmonary disease, it seems to me that it would be reasonable to extrapolate from healthy children to children with diseases that are not immunocompromising.  You would not necessarily have to conduct studies among those subpopulations.  That would be one area where you could, I think, have more generalizable information.

DR. MODLIN:  That’s a good point.  The only issue might be tolerability to the vaccine, as opposed to immunogenicity, in that group.  A safety issue could be somewhat different in that group -- maybe, maybe not.  That’s the only caveat there.

Seth?

DR. HETHERINGTON:  The other usual subgroups that come up here are children with nephritic syndrome or antibody-deficiency syndromes.  It might be interesting to know if the use of adjuvants helps those groups any compared to the naked vaccine.

DR. MODLIN:  Children with nephrosis are immunocompromised only when they have nephrosis.  It’s because they are losing antibody, which is typical only for a week or two during their exacerbations.

We could probably get into lots of different subpopulations here.  But what I was trying to raise earlier were the general ones that we recognize for children that are truly at increased risk for influenza.

Pam?

DR. MCINNES:  I guess my issue is a little more pedestrian.  In terms of looking at those 3- to 5-year-olds and 6- to 9-year-olds, it seems that a natural group is 2 to 9, if we are stepping down from adult data.  I didn’t see a need to stratify within those based on data.

Then I guess the next group that comes to mind is the 6- to 24-month-olds, because we do immunize children from 6 to 24 months old right now.  You would actually have some comparative data that you could look at.  Then the question of the 2- to 6-month-olds -- of course, the question, then, on concomitant administration of other vaccines and safety in that regard.  I think there would just be a whole host of parameters that might play out in that.

But from an age-group perspective, I can see kind of a methodical way of stepping down.

DR. MODLIN:  Good points.  Norm?

DR. FOST:  John, just going back to your general comment that the children we would want to study are the ones at highest risk, and coming back to a point I made earlier, that would be children in the most endemic areas.  Ideally, the earliest trials should be in the children who have the most to gain from it, which is not in the United States.

DR. MODLIN:  That’s an excellent point.  It certainly raises some of the ethical issues as well.  I think probably we should take a minute or two to discuss that.  I think the point was made that even though we haven’t seen any disease in the Western Hemisphere, the number of cases that have been seen in the tropics and the Eastern Hemisphere number in the hundreds, which makes it still an extraordinarily rare disease in these groups.  You can identify, obviously, high-risk factors as largely children who have been exposed to avian influenza through exposure to avian flocks, domestic avian flocks for the most part.

I’m thinking out loud here.  It raises major challenges, I think, in study design, in trying to study a vaccine in those groups, and whether or not that is technically feasible or possible to do, in order to address the ethical issue of considering them to be at high risk.

I’m not being very articulate here --

DR. FOST:  What design problems did you have in mind?  There are logistical problems and political problems, but --

DR. MODLIN:  I was including those.  Also I think there is this issue of whether they are truly at higher risk, which I think is probably open to interpretation.  Of course, that also would depend upon the stage at which you are in terms of -- but here, in 2009, I don’t think a child in Western Europe is at any higher risk than a child in the U.S.  A child in Turkey probably is, albeit the risk there is still extraordinarily low.

DR. FOST:  But, as you said, children who are raised in communities where flocks of chickens are in their backyards -- I don’t know how many children you would need for these trials -- that would seem to be the ethically ideal population in terms of risk/benefit ratio.

DR. MODLIN:  How do others feel?  Bruce?

DR. GELLIN:  I guess I’m not sure what this group is going to do with that kind of discussion, other than to highlight what I mentioned before, that there is an ongoing discussion at the World Health Organization, which is acquiring a stockpile, on how that stockpile might be used as well.  I would think that this question is going to be the same.  If you are going to be thinking about the use of a vaccine in a population, then what are the considerations before you get there?

I don’t know how much we need to weigh into that.  I think it’s an important thing to highlight, and maybe to transmit that to the World Health Organization, what our deliberations as we have considered this one.

But I’m not sure what we are going to do with that information.  The manufacturers may already have information on some of the studies that are going on in a whole range of these populations.  I took “risk” as meaning to be risk of complications from influenza, not risk of exposure in this.

DR. MODLIN:  Both.  I think that’s what Norm is getting at, and then the ethical issue of whether or not we shouldn’t be conducting it in someone who is at slightly higher risk of exposure.

But I think the point to be made is that the risk of exposure of children in this country is extraordinarily low, but it may not be a whole lot different than the risk of a child in Turkey.

DR. FOST:  You don’t want to study an HIV vaccine in Idaho.  You want to study it in a population where the risk of HIV is extremely high.  It seems to me it’s the same principle -- unless there are compelling reasons not to.

DR. MODLIN:  We’ll be careful.  We won’t go there.  That could raise a whole other -- Bob?

DR. DAUM:  It is interesting to sit and parse out which children play with birds and maybe generate a little cohort of children at slightly higher risk.  But they may be different.  They have more preexisting antibody, for example.  They may have a different kind of exposure.  I think it begets the question:  What do we want to know, and when do we want to know it?

My argument would be that there is going to be a pandemic of flu.  I have no idea whether it will be this coming season or beyond my lifetime.  But I’m in favor of ideas to get our population ready for such a pandemic, and “ready” includes antivirals, perhaps, and includes vaccines, almost certainly.  And I would like to know how those vaccines perform in healthy children.  By “healthy” in this context, particularly for Dr. Nelson’s benefit, I’ll say people who aren’t necessarily exposed to chickens and aren’t necessarily at high risk.

So I have no objection to doing studies in children in countries where they run around with chickens and where there are cases.  That’s fine.  Those are going on, it sounds like.  But I also think we want to know about people that are currently not in those high risks.

That’s the argument that I have been trying to advance, and I’m going to continue to stand there.

DR. MODLIN:  You are saying that’s the public health imperative.

DR. DAUM:  I think it is.

DR. MODLIN:  Any other comments about this?  Vicky?

DR. DEBOLD:  I disagree with enrolling sick kids, little kids, into these trials at this point.  I don’t think we know enough yet about developmental immunotoxicology as it relates to the adjuvants.  I think the story that we heard this morning about the kid who had elevated liver enzymes and developed subsequent issues -- many of the kids now are immune-activated.  They are allergic.  They have food allergies.  We know that the two adjuvants that we talked about today really turn on the immune system.  So what happens when you really turn it on?

I think there’s a lot of basic science that needs to be done before we go there.

DR. MODLIN:  Thank you.  Roland?

DR. LEVANDOWSKI:  One more comment about risk.  We have been very focused on the highly pathogenic avian influenza viruses, because they are very scary.  They do kill people when they infect them.  We haven’t been as focused on the possibility that a pandemic of influenza might start from some other source --

DR. MODLIN:  H2 or something.

DR. LEVANDOWSKI:  It could be H5.  We have non-highly pathogenic avian influenza viruses across the United States.  There has been H5 in the Southwestern United States and H7 in Virginia.  These things exist in the wild avian populations here.  There’s nothing that really tells us that there couldn’t be a reassorting event of some sort that happens in some child or person who lives on the farm or goes hunting or gets exposed to something while they are involved in nature in some way.

So I think, in terms of the risk, it’s not just risk because there’s something scary happening in Asia or Europe, and it doesn’t seem to be the same scary thing happening in the United States.  The fact is that there is the possibility that another event could originate a pandemic here, which we would still need to deal with in a similar way.

DR. MODLIN:  Jack?

DR. STAPLETON:  I would like to respond a bit.  I wouldn’t advocate that we use adjuvants to test the idea of autoimmunity.  But the opposite of that is the possibility that, actually, by immunizing with adjuvants, you may have a beneficial effect on asthma and autoimmune diseases.  I would say that one thing that should be built into these vaccine studies, where you have a definite outcome measure, is longer-term follow-up to look at both harm and good from these vaccinations.

DR. MODLIN:  I’ll let you and Vicky debate the hygiene hypothesis.  That’s exactly where you’re going here, which I fully understand.

Norm, in terms of pediatric subpopulations, what I’m hearing is that there is some enthusiasm for testing novel vaccine antigens in kids younger than 6 months of age, and probably less enthusiasm for children that we have put in other high-risk categories.  Dr. Jackson’s point was that unless they are immunocompromised, the expectation is that their ability to respond to the vaccine would be very similar to that of healthy kids.  The only question might be whether or not there’s a different adverse-event profile.  I think that would be something that would need to be taken into account on a case-by-case basis.

DR. BAYLOR:  Also we want to touch upon, is there anywhere in that pediatric population that you can extrapolate, instead of doing studies across the board in, say, zero to 18 years of age.  We used the word “subpopulation” to look at that as well, where there would be some age groups that you wouldn’t necessarily have to study, where you would be able to extrapolate to others.  I’m assuming -- maybe I’m incorrectly assuming -- that that would be the case.

DR. MODLIN:  You are talking about extrapolating adult data down to a certain --

DR. BAYLOR:  No, even within the pediatric age group.  If I did a study in adults, should I do studies in 16- to 18-year-olds?  Should I do studies in 15-year-olds?  If I do studies -- what you have already said you would like to see are studies in children below 6 months of age.  If I did studies in below-6-months-of-age, perhaps there are age groups above that I may not do studies is -- just getting into those subpopulations.

Pam, you made comments about the 2- to 6-month age group and the complications with that because of the concomitant immunization schedules in that timeframe.  You mentioned the 2- to 9-year-olds.

I guess what I’m getting at is, it’s not necessarily to study the whole range of the pediatric population.  There are areas where you could extrapolate within there.  I just want to get some confirmation of that.

DR. MODLIN:  I’ll let others weigh in here, but I would think that certainly kids that postpubertal, teenagers, could probably be studied with adults, or adult data could easily be extrapolated to them.

As everybody knows, the reason that we study the age range that we do now with seasonal vaccines is that up to at least 5 or 6 years of age, kids don’t respond as well to inactivated vaccines, at least to one dose, probably because they don’t have the immunologic memory to do so.  Once you get beyond 5 or 6, up closer to 8 or 9 years of age, you are more likely to see responses that we observe in adults.

That’s with seasonal vaccine.  With a novel vaccine, obviously, all bets are off, and we have to consider them immunologically naïve.

Jack?

DR. STAPLETON:  I think Pam’s 2-to-9 group -- if those data are very similar to adult data, then I would think you could extrapolate in both directions.

DR. MODLIN:  That’s where I was going.  Dr. Joffe?

DR. JOFFE:  I just want to say something again that I said when we were going around.  I haven’t yet heard a compelling rationale for studying, at least at this point, the younger-than-6-month-olds, if you are thinking use in the pre-pandemic setting.  I think the first question, the policy question, is, what do you envision the pre-pandemic population-based recommendations to be?  What would be the minimum age at which you would start to vaccinate infants in the pre-pandemic setting?  Then you want to be sure to study it up until that age limit at which the policy recommendations would kick in.

I think the issue of what you need to know in the younger-than-6-month-olds if you are thinking about use in the pandemic setting, not the pre-pandemic setting, is different.  There you might well say, whatever the facts are, the truth is, in the event of a pandemic, we are going to vaccinate everybody, and we are going to assume that we can extrapolate from, say, 6 to 12 months down to the younger infants, and maybe you don’t need those studies.

I think that’s a different discussion.  But I think in terms of what the lower age limit is, the first question is, what are the pre-pandemic policy recommendations going to be?  Then let’s make sure we have the information to guide administration of the vaccine within the scope of those policy recommendations.

I realize we can’t answer that, because the policy recommendations aren’t in place.  But that to me is the thought process.

DR. MODLIN:  I guess I would turn that around, by saying that I would guess that whatever science we have or data we generate will drive the policy, rather than the other way around.  That would be my guess.  I think that’s what we are talking about here.  I would hope so.

Norm, anything else about subpopulations that is important?  Have we had an adequate go at that?

The next question is the adult safety and immunogenicity data needed to support proceeding to pediatric studies.  We have been talking around that.

In your deliberations, please consider both the use of novel adjuvants and also whether other viral subtypes, other than H5N1, should be studied.

Again, we have had some discussion on both of those.  Why don’t we specifically focus on the adjuvant issue here?

It sounds to me like we have heard an awful lot this morning about the importance of adjuvants with novel antigens, in both adults and children.  It seems very hard to think that we are going to get away from not needing adjuvants, particularly since they are so antigen-sparing.

Let me ask how others feel about that.  Any disagreements?

DR. JACKSON:  No.  It seems clear that there is substantial benefit to adjuvants, and we would want to give the most beneficial vaccine.  You, of course, want to start studies in adults and move down, which has been the way it has been proceeding so far.  It’s hard to imagine that new studies would proceed any differently.

DR. MODLIN:  Any disagreements?  Ted?

DR. EICKHOFF:  That’s fine and good, and I would support beginning with adults and moving down.  But in the event of a threatened pandemic, you may not have the luxury of being able to derive adult data before you move into younger age groups.  So it depends on the assessment of how much time you have at your disposal.  But again, in the event of a threatened pandemic, you may not have that time.

DR. MODLIN:  Jack?

DR. STAPLETON:  I think, as Lisa said earlier, we are at that stage.  We actually have quite a bit of data in adults.  So I wonder whether we really need that much more adult data at this point to recommend going forward.  From what I have seen from this morning’s presentations, I would think that we have quite a bit of data in children and adults.  So moving ahead makes sense.

DR. MODLIN:  Christine, why don’t you put question number 2 up there, if you wouldn’t mind?

Any other discussion about viral subtypes?  Roland and others have raised some important issues.

Dr. Gilbert?

DR. GILBERT:  I would like to expand on a topic I brought up earlier, which is the use of a standardized panel of influenza isolates.  I mostly work on HIV vaccines, and in that field, it developed by different sponsors using their own strains to evaluate their vaccines, and on the basis of antibody levels or T-cell levels to those strains, they would advance the candidates up the pipeline in clinical trials.  So the field recognized that there was an urgent need to get several panels of standardized HIV isolates that would be shared, so that all vaccines can be compared using the same panel, so we can then interpret, in a head-to-head way, what the antibody or the T-cell data mean.

That might be something that needs to be part of the testing process here, to try to move toward a standardized panel that, in some sense, is representative of the antigenic types that putatively could lead to a pandemic.

DR. MODLIN:  We had a lot of discussion about testing of panels yesterday with seasonal influenza.  Obviously, WHO and CDC are tracking the ongoing changes with H5N1 and certainly have a panel of viruses.  I would guess, Dr. Klimov, that that panel is, for the most part, available to FDA and to vaccine manufacturers as well?  It’s probably changing --

DR. GILBERT:  In the HIV field, our goal is to update the panel periodically, regional panels, as well as subtype-specific panels.

DR. KLIMOV:  The panels, every six months, come together for seasonal influenza, and they update the situation with the H5N1 at this time.  We also have regular conference calls with HHS on the development and evaluation of influenza H5N1s in particular and the recent status of the vaccine candidates’ preparedness.

There are several groups, including several groups in the United States -- one of them is CDC; another one is St. Jude Children’s Hospital -- that are preparing vaccine candidates from different clades and sub-clades of influenza H5N1 viruses.  There is a discussion on the evaluation of recent H5N1 viruses, their antigenic profiles, antigenic differences between different clades, sub-clades, and what would be suggested as next potential vaccine candidates.

Right now vaccine strains, essentially, from clade 1 are available.  Vaccine from sub-clades 2.1, which is Indonesian virus, is available.  There are vaccine strains, a couple of them, from sub-clades 2.3.4, which are viruses which are mostly in China.  There are vaccine candidates from the sub-clades 2.2.  There is work done on the clade 7.  There was only a single case in China in 2003, but there was a human case of clade 7.  So there is work on clade 7 vaccine development.  Within, for example, sub-clade 2.2., which is the most widely geographically distributed sub-clade, there are several candidates, including, most recently, viruses from Egypt which seem to be antigenically and genetically more advanced.

So this work is regular and it’s under watch by HHS, by WHO, and by different groups.

DR. MODLIN:  Thank you.  Roland?

DR. LEVANDOWSKI:  Along those same lines, I think part of the question about standardization, if I understood what you were saying, is that you want to be able to compare results from different laboratories.  That’s actually not very possible, with either hemagglutination inhibition or microneutralization, for flu.  There has been a lot of work done to try to understand that.  There is no international standard for the different flu strains.  Although the strain itself may be the same, there are variations that may occur as it’s passaged in the laboratory to prepare it to be used in these different types of assays.  The assays themselves have some different features to them, the way that they are handled.

The point is, there have been studies that have done comparing different laboratories.  This information has been published for seasonal-type vaccine, and there is a study that is about to be published for H5 as well.  They say the same thing.  Even though each of the laboratories has what would be considered a validated assay -- internally, everything is consistent and it’s reproducible and reliable -- when you try to compare the results from the different laboratories, there may be fourfold, eightfold, even higher differences between the absolute titers.  You see that in the information that is provided for the committee here for strain selection every year.  Those panels that are used by the different labs to test the sera -- you see that the absolute titers are different.

Within that context, though, what you will see is that some labs tend to be higher, some labs tend to be lower, and if you tried to rank the different sera that you were looking at, you would see that they generally are ranked pretty much the same.  But the absolute titers are not that easily comparable.  If you were trying to look at, to use the term of art, seroprotection between different vaccines and different studies, it’s very difficult to do that sort of extrapolation.

DR. GILBERT:  I don’t want this to go on too long, but how, then, might you recommend that policymakers decide on recommendations for choosing one sponsor’s vaccine versus another sponsor’s vaccine, if you can’t interpret their data in a comparable way?

DR. LEVANDOWSKI:  I’m not sure I can answer that question directly.  I think what needs to be done for the vaccines is to demonstrate that they are immunogenic and safe.  I don’t think we actually know, particularly for H5, what the true protection level is for any antibody.  The correlation there hasn’t been done yet, so we don’t really have enough information to say.  But I think we can generally say is that more is more in terms of antibody against influenza.  The higher the titers, the more likely there is to be resistance to infection, less likelihood of complications and deaths.  It’s a matter of trying to increase that as much as possible, and therefore, a lot of the discussion about the adjuvants that we are having now.

DR. MODLIN:  That’s a perfect segue into this last question:  Please discuss what pediatric safety and immunogenicity data you would consider adequate to support licensure of inactivated pandemic influenza vaccine candidates for use in one or more pediatric populations.  Again, we are asked to consider both the use of adjuvants and other viral subtypes, other than H5N1.

Bruce?

DR. GELLIN:  This is a licensed pandemic vaccine?

DR. MODLIN:  That would support licensure for a vaccine to be used in one or more pediatric populations.

DR. GELLIN:  At what point in time?  At the pandemic or tomorrow?

DR. BAYLOR:  We’re not talking about, necessarily, tomorrow.  If the pandemic has been declared, we would like to use this vaccine to immunize our population.  What type of data would you recommend for that?

DR. GELLIN:  A pandemic vaccine in the setting of an imminent pandemic, not a pre-pandemic vaccine.

DR. BAYLOR:  Not pre-pandemic, correct.

DR. MODLIN:  I would guess that it would be very similar to the discussion we had two years ago on the Sanofi H5 vaccine for adults.  We’re pretty much in an analogous situation for children, are we not?

What would be an adequate safety database?  Jack?

DR. STAPLETON:  John, it might help the group if you kind of went through the discussions and summarize them from the Sanofi licensure.

DR. MODLIN:  I’m sorry?

DR. STAPLETON:  It might be helpful if you wanted to summarize the discussions from the Sanofi licensure --the idea that this vaccine would be used at the start of an imminent pandemic as a way to provide some immunologic memory.  Or am I oversimplifying?

DR. MODLIN:  Be my guest.

DR. STAPLETON:  The way I remember that discussion -- in my typical oversimplification way of thinking -- is that having a stockpile of a licensed vaccine that, if a pandemic was imminent, would be distributed in a systematic way to those at highest risk and on down, not with the hope that the clade 1 vaccine would protect against the pandemic strain well, but that it would provide some immunologic memory, it might provide some protection, and it might be a boost phenomenon when the pandemic vaccine came around -- maybe I’m oversimplifying.

DR. MODLIN:  So not necessarily to protect against disease, but to protect against significant morbidity and mortality.  That was the whole idea.

Really, the question is, what do we consider an adequate safety database?  I think Roland the point, which we all would agree to, that we really are in completely uncharted territory when it comes to deciding an adequate level of antibody we could use as a surrogate for protection.  We just don’t know.  We would never know until we had the opportunity to actually employ a vaccine.  So we have to rely on other data to make judgments about what we consider to be adequate immunogenicity.

How about the safety database?  Let’s focus on the safety database here.  In other words, what sorts of studies would you consider to be adequate to support licensure in the pediatric age group?  Pamela?

DR. MCINNES:  Several hundred in each age group.  I don’t think I would stretch to thousands as being necessary, but a robust hundreds number would be -- in terms of numbers.

May I retreat one step to the previous -- if the principles of what you want to see in the immune response would be possible to articulate.  Even though we don’t know an absolute number we want somebody’s titer, we would hope that there is indication of children being primed, that there is a booster response, that we have some understanding of the kinetics of the response, duration of the response; if you could think about vaccine as essentially a la challenge with virus, that they responded.  I think we would want to see indications of kinetics that might indicative of a favorable response.

DR. MODLIN:  And you would want to test enough kids that you would have confidence that the results were generalizable.  I’m not sure that we can get a whole lot more precise than that.

DR. MCINNES:  Right.

DR. MODLIN:  Bob?

DR. DAUM:  I think Pam is right.  I like to think of the safety issues as the ones that are retrievable with the hundreds-of-thousands kind of sample size and then the ones that require the millions and tens of millions to retrieve that are really quite rare.  I don’t think it’s possible, particularly in this instance, to think about the large-scale kind of detection, with 10 million or 1 million people.  I think we have to rule out common and relatively rare -- I like the way the speaker from the Netherlands spoke this morning about that -- rule out the relatively rare side effects, with a sample size of about the size that Pam is saying, and then remember that if we need to use this vaccine on a wholesale scale, we have a disease with 60 percent mortality, or maybe more, and that finding a very rare side effect might be less important than rolling it out to large numbers of people.

So as best as I can sit here and guide you, not knowing what’s going to happen when, I would say that something that passes the hundreds-to-thousands test for safety and immunogenicity and any other data we can bring to bear on it ought to be licensed and prepared -- or considered a candidate for preparation.  Let’s put it that way.

DR. MODLIN:  Whether it includes an adjuvant or not.

DR. DAUM:  Whether it includes an adjuvant or not.

DR. MODLIN:  Seth?

DR. HETHERINGTON:  We are throwing numbers out.  Let’s, again, put them in context.  If your definition of a rare event is 1 in 1,000 or less, you need a database of 3,000 patients to rule it out.  If you are willing to accept a rate of 1 percent, you need 300 patients’ worth of data.

So, really, what it comes down to is, in the setting of a pandemic, what do you think your infection rate is going to be?  What do you think your case fatality rate is going to be?  Balance that against your effect of the vaccine, which includes your seroconversion rate.  What do you think the reduction in mortality is going to be or reduction in disease burden, however you want to quantitate these things?  Then it’s a simple calculation after that.  That tells you what your expected benefit is going to be.  And then how much risk are you willing to take on in order to gain that benefit?

Unfortunately, there are going to be some assumptions that you have to make along the way.  But unless you make those assumptions, there’s no way to calculate the risk/benefit ratio.

You just have to keep that in mind as you start putting numbers out there.  Do you want a database of several hundred versus a few thousand?  What does it really mean?  You just have to think about how much of a risk you are willing to take.

I would bet that if we had, tomorrow, a severe pandemic and you started having lots of kids in the hospital and you started seeing deaths, you would probably take the data you have today and say, let’s start immunizing kids.

So you have to think of it in terms of the range of possible outcomes that you are going to get with your pandemic and then, again, what you are willing to take on as a risk.

DR. MODLIN:  Good points.  Lisa?

DR. JACKSON:  I would also agree with Pamela in arguing for the several hundred.  Also keep in mind that we would want to have the ability to evaluate the possibility of adverse events that are not uncommon and also that are expected to occur relatively proximate to vaccine administration.  Some of these more hypothetical concerns regarding autoimmunity and so forth -- I don’t know that it would be possible to study those, even with much larger sample sizes.  For one thing, you would need to follow people for a long time.  For another, you would have to have a randomized placebo group of fair size, which, especially in the younger populations, could be a real issue, because then you start getting away from the possibility of direct benefit to your study participants.

So I think we ought to be as constrained as possible.

DR. MODLIN:  Thank you.  Dr. Nelson?

DR. NELSON:  I would just like to ask a follow-up question to Seth’s balancing of the calculations, which strikes me as quite useful.  Given that the studies to support the pandemic indication would be done in the absence of an existing pandemic, would you use the calculations you went through to argue that you could use a smaller study for licensure, given that the risk/benefit in the actual use condition would be different than if you were going for, say, a pre-pandemic indication?  You would limit the size of the study which is being done in the absence of a pandemic, understanding that the information you would need would be set against a much larger risk and disease burden in the event of a pandemic?

DR. HETHERINGTON:  Right.  And that’s why I brought up earlier on the question of under what circumstance you anticipate using this vaccine.  I think we have been talking mostly about a pandemic situation, not a pre-pandemic situation, in which case you would be vaccinating many to have benefit for, potentially, very few or benefit that is pushed off well into the future.  You have to discount what that benefit is relative to your immediate risks around the time of immunization.

That’s an even more difficult calculation that I can’t even begin to conceive.  But I think if you make some assumptions about what would happen in a true pandemic situation, you might be able to come up with an answer that, in fact, would lead you to accept a smaller database ‑‑ certainly a smaller database than 3,000.

DR. MODLIN:  Bruce?

DR. GELLIN:  Seth is right.  The math is easy; it’s the assumptions that are the hard part, and which assumptions you are going to use.  It’s worth remembering that part of the pandemic preparedness exercise was the development of what we refer to as the Pandemic Severity Index.  It recognizes that pandemics can roll out in different ways, and depending on the severity, you might do different things.

As a country, we have sort of taken the most severe pandemic and tried to use that as the bar.  Whether that’s the equivalent of this, that may help to drive which assumptions you use.  But the other things that we do are based on the most severe pandemic and then can back off from that, with the theory that it’s better to plan for the worst rather than hope for the best.

I think the other piece that this question -- we are discussing this question as it’s asked, but we can’t forget the second part of this.  Once a vaccine is licensed, it doesn’t mean we stop looking.  A large part of the discussion around the Sanofi vaccine initially was the systems in place that would be able to evaluate the things that you wouldn’t be able to pick up pre-licensure.

So I think we can’t forget that.  That’s not what the question is about, but it’s not as though once this thing is licensed, that’s the last time we consider it.

DR. MODLIN:  Good points.  Dr. Joffe?

DR. JOFFE:  We have been talking mainly about the safety database in terms of numbers needed to have a reasonable sort of precision around safety estimates.  But I think an issue that Dr. Jackson raised is, how long of a safety follow-up do you need on these kids?  I think it’s an important issue to be reckoned with.  Are we talking about safety follow-up on the order of months?  Do we need one-year follow-up?  Do we need more than one-year follow-up?  I just think it’s worth raising this discussion, particularly because some of the things that are more theoretical, and presumably rarer, but are the things we worry most about, like autoimmunity, are things that are not likely to be detected in very short-term studies.

The second half of that is, what are the safety endpoints that really need to be collected?  Some of the short-term ones are obvious, but some of the more intermediate-term ones, I think, are less obvious, at least to me.  It would be helpful to offer some guidance on what those safety endpoints might be.

The final point -- again, to go to something Dr. Jackson brought up and something I mentioned earlier -- I feel very strongly that the appropriate controls for these studies -- that it is appropriate to have randomized saline placebo-controlled controls for these studies, because I think that’s the only way you are really going to be able to have any confidence in your safety conclusions.  I think having a placebo that has its own toxicities is going to confound the ability to evaluate safety.

There seems to be an ethical rationale for wanting to do that, in the sense of being able to offer some benefit to those who get randomized to the control group.  I would be interested to know what others think, but to me that seems like an inaccurate ethical conclusion.  There’s nothing unethical about randomizing the control group to a saline placebo, and there is no ethical imperative to randomize them to something that offers them some sort of off-target possibility of benefit.

 DR. MODLIN:  I don’t know if others want to address that.  I think the obvious reason why the placebo contains adjuvant is to understand what added toxicity there may be to the vaccine component itself, which is difficult to know if you don’t have an adjuvant-containing placebo.  I think, under ideal circumstances, you might want to do both, which would give you the maximum information about toxicity.

Norm?

DR. BAYLOR:  I want to make a comment about the ‑‑ when we compare what we did with the Sanofi vaccine, the landscape was different than it is now.  As I hear some of the discussion about size and some of the numbers that have been thrown out, we are in a difficult position.  If the pandemic doesn’t come tomorrow, theoretically I can collect more data.  The longer the pandemic is delayed, the more data I can collect.  You have to look at the timeframe when you are thinking about licensure.  We do know that as we collect data -- and, say, the pandemic did come tomorrow ‑‑ we could use an emergency-use authorization.  At least we would have some data to base it on.

But it begs the question of when you stop collecting the data.  As these trials are going on and you have an endpoint of licensure, you have to keep that in mind.  If the emergency was tomorrow, you may not be able to collect all of the data you want, but you would be able to collect some.

I guess I would like some comment on -- look at this in a timeframe.  We still have time to collect data.  Where do we draw that line and say, this is sufficient for licensure or this is sufficient for use in case we have that outbreak coming sooner than later?  I think those are different scenarios.  I think more time we have, the more time we have to collect additional data.

Some of the numbers that were thrown out, some of the ideas that were thrown out seemed a minimum.  Was that because of the urgency?  If this happens tomorrow -- or would you say, even if it didn’t, you would feel that was sufficient?  What you threw out, Pam, would you feel that was sufficient, regardless?

DR. MCINNES:  I’m still in the hundreds.  I’m thinking about not just one vaccine.  It could be a whole series of vaccines that you are collecting a body of data on.  I think you might titrate your decision, depending on what you were seeing across the spectrum of vaccines.

If I could ask for a clarification, Norm:  Are you really envisioning non-deployment until -- you want data for non-deployment of vaccine, and deployment only in the face of pandemic?  Or are we talking about a scenario where you might be trying to prime a population in pre-pandemic?

DR. BAYLOR:  We really wanted to stay away from the pre-pandemic.  I think the pre-pandemic gets a bit more complicated.

DR. MCINNES:  Right.  So we are talking about a body of data for deployment in the event of a pandemic.

DR. BAYLOR:  Yes.

DR. MODLIN:  Once you have safety data with a manufacturer’s vaccine that is made according to a certain process, that contains a stable adjuvant that you know a fair amount about, you have an adequate safety database, perhaps, with a few hundred people, particularly if you apply Seth’s math to the risk/benefit ratio -- being able to tolerate a relatively low number, but definable adverse outcomes, in the face of a pandemic.

As time goes on, the influenza antigen is going to change.  Inevitably, the manufacturer is going to want to come back and update at least the antigen in the vaccine, to the point where I think most of the information that you are going to need is going to be on immunogenicity at that point, and probably the safety issues are going to become slightly less important with an individual vaccine that is made according to the same standard.  Is that not the case?

DR. BAYLOR:  I’m not so sure.

DR. MODLIN:  It’s basically getting at the question that was raised earlier:  How small a database can you use, given the fact that we are looking at pandemic vaccines?

José?

DR. ROMERO:  Norm said something that triggered a thought.  What is the sense of urgency that we have for this?  These cases have occurred outside of the continent of the Americas.  Would our sense of urgency change if, for example, in Latin America, in Mexico, where -- somebody said running around with chickens.  I remember running around with chickens in my grandfather’s and grandmother’s batio (phonetic).  Would this prompt us to change our view of this?  Norm’s comment about the sense of urgency -- would our approach to this change if this year they report a case of human disease in Latin America?

DR. MODLIN:  I think the answer is, without a doubt it would change.

Seth?

DR. HETHERINGTON:  I just want to make one comment also to Norm’s statement about pre-pandemic versus pandemic.  While I think the urgency for a pandemic preparedness is obviously at the forefront, I hope that the FDA and the manufacturers are thinking about a pathway to a pre-pandemic scenario and what would be acceptable for approval.  There is very good reason to think there are benefits to priming this population, assuming that you get whatever additional safety data is necessary to feel comfortable about that.  Now, that’s going to be probably an order of magnitude in terms of safety data than you would be willing to accept for a pandemic vaccine.  But the benefits are probably also an order of magnitude greater to have a primed population.

So I just hope that that’s being discussed and considered.  I’m sure the manufacturers are moving toward that direction in their own thinking.

DR. BAYLOR:  It definitely is.  In fact, we did present some scenarios of that a couple of VRBPACs ago.  So that is ongoing.

DR. STAPLETON:  Along the same line, and echoing what Pamela was, I think, implying, perhaps the numbers that you need are not that much beyond what we have, but in going forward, looking at whether clade 1 primes well for different clade 2s, what kinds of immune responses you get to clade 2, whether clade 2 is better at priming -- those types of kinetics experiments would be probably more helpful.

DR. MODLIN:  Further comments?  Norm?

DR. BAYLOR:  I’m going to back up a little bit.  I want to go back to the second part of question 1, if I may, on the adult data -- we touched on that lightly -- the adult data needed prior to going into those pediatric studies.  As you have been presented today, there are adult studies ongoing.  Are those sufficient to move into the pediatric population?  Say, for instance, we had an adjuvant that was not discussed today or a novel manufacturing process that was not discussed today.  From what you have seen today, are those the types of numbers in the adults that would be adequate to move into the pediatric population?

DR. MODLIN:  The numbers we saw were tens of thousands of adults that were being tested, so I would think, certainly, that the answer would be yes, from my standpoint.

Does anybody feel any differently?

Are you asking us how many adults would need to be tested to be confident to go into the pediatric population?

DR. BAYLOR:  Yes.

DR. MCINNES:  I would want to see more numbers in adults than for children.  I think you want to be very confident that you have characterized the safety and immunogenicity profile in adults.

DR. MODLIN:  Can I push you a little bit, Pam?  Is that for -- adequate to assess adverse events that we would consider to be relatively uncommon or adequate to assess adverse events like Guillain-Barré syndrome that are extraordinarily rare, the 1 per 100,000, the 1 per 1 million?

DR. MCINNES:  No.  I’m talking about characterizing events that are more temporally related to vaccination.

DR. MODLIN:  Norm, it’s 2:55.  Is there anything that we haven’t addressed, either for you or Dr. Nelson or Dr. Pratt?

DR. BAYLOR:  I guess not.

DR. MODLIN:  I do know people have cabs waiting and need to get off to the airport.  I do want to express my personal thanks to all the members of the committee, those that have joined us as voting members.  Certainly terrific presentations from the FDA and from industry.  Thank you, everyone.  See you at the next meeting.

(Whereupon, at 3:00 p.m., the meeting was adjourned.)