Welcome and Introductions
On behalf of the National Institute of Allergy and Infectious Diseases (NIAID), Dr. Linda Lambert, Program Officer at NIAID's Division of Microbiology and Infectious Diseases (DMID), welcomed 48 participants to the workshop on the development of a clinical trial plan for pandemic influenza vaccines. Attendees included representatives of Federal and international agencies, academic research institutions, and numerous representatives from the influenza vaccine industry. Dr. Lambert noted that NIAID supports a wide range of research on influenza, including biology, epidemiology, pathogenesis, immunology, and the development of new influenza vaccines and antiviral drugs. Other agencies that are involved in influenza research include the Center for Biologics Evaluation and Research (CBER) and the Center for Drugs Evaluation and Research (CDER) at the Food and Drug Administration (FDA), the Centers for Disease Control and Prevention (CDC), the U.S. Department of Agriculture (USDA), and the U.S. Department of Defense (DOD).
Representatives from these agencies and many individuals at the present workshop participated in a NIH sponsored workshop 8 years ago to identify gaps in the scientific understanding of pandemic influenza preparedness, with the goal of helping to define a research agenda for the world's research community. Dr. Lambert noted that impressive progress has been made in many areas of the pandemic research agenda and presented the following workshop objectives.
Workshop Objectives
The primary objectives of the present workshop were to bring U.S. and international influenza vaccine experts together to:
• Review clinical studies that have been conducted previously with vaccines containing novel hemagglutinin (HA) and neuraminidase (NA).
• Discuss manufacturing and regulatory issues related to the production and clinical evaluation of pandemic influenza vaccines.
• Identify pandemic reference strains, reagents, needed vaccines, and opportunities for international collaboration and sharing of these resources.
• Determine the types of vaccine studies that must be conducted during the interpandemic period to build a scientific base of knowledge for their development and to improve preparedness for future pandemics.
• Identify points to consider for the design of a clinical protocol to evaluate a pandemic vaccine.
Workshop Structure
The workshop was designed to maximize participant involvement and input. Moderators were assigned to each topic to facilitate discussion. The first day of the meeting included a morning session with five presentations and an afternoon session with discussions on preclinical and clinical issues. The second day of the meeting focused on the presentation and discussion of key points raised during the first day.
Top of Page
Vaccine Experiences/Lessons From the Past
Dr. Robert Couch provided a summary of the research on inactivated influenza vaccines for pandemics and pandemic threats. He reviewed findings from key studies of vaccines in 1957, (influenza A [H2N2], “Asian”), 1968 (influenza A [H3N2], “ Hong Kong ”), 1976 (influenza A [H1N1], “swine”), and 1977 (influenza [H1N1], “Russian”). After describing various comparisons, Dr. Couch presented a series of lessons learned and summarized the opinions of some experts.
Key lessons learned from past influenza pandemics and pandemic threats include the following:
• Increasing purity reduces reactogenicity.
• Increasing the dose, giving two doses, or using an adjuvant increases the antibody response.
• Priming is a major factor for determining response to one dose.
• Whole-virus vaccines may not be uniformly more reactogenic than split/subunit vaccines.
• Efficacy is variable and may not relate well to serum HAI antibody titers.
• Available data suggest that the higher the immune response, the greater the protection, although the desired level is uncertain. The current surrogate of protection is serum HAI antibody, but researchers should look for alternative surrogates with broader and more complete immune responses.
In the discussion that followed his presentation, Dr. Couch suggested that, given the lack of immunologic surrogates for protection in a pandemic situation, researchers should continue to use the current surrogate of serum HAI antibody and seek a vaccine that, through a combination of factors (e.g., manufacturing capabilities and adjuvant use), produces the highest possible titers of antibody. He also noted that the limited understanding of immune responses does not provide enough evidence to determine whether providing the same amount of HA through one dose or two doses is the best strategy. He added that the data on the persistence of HAI antibody indicates that up to 3 years is a good guideline, with dropoff of twofold to eightfold occurring during that period.
Top of Page
Recent Vaccine Experience With Novel Antigens
Dr. Karl Nicholson provided an overview of five recent vaccine studies that used novel HAs, including a recombinant H5 (rH5) vaccine; a H5N3 vaccine, with and without an MF59 adjuvant and with one or two doses; comparisons of a H2N2 split-product (SP) vaccine and a whole virion (WV) vaccine with and without Alum adjuvant; and comparisons of a H9N2 WV vaccine and surface antigen (SA) vaccine with or without Alum adjuvant. Dr. Nicholson summarized key points from the studies and commented on their relevance to protocol design for clinical trials of pandemic influenza vaccines.
Dr. Nicholson's comments on the implications for clinical trials are summarized below:
• Adverse Events— Candidate vaccines must be carefully assessed in young children for tolerability.
• Number of Doses— Clinical studies should always be designed to include two doses rather than only one dose and to assess the effect of age on antibody responsiveness.
• Dose Range —With conventionally prepared material, there seems to be little point in using “high” doses of HA because vaccines will be limited, at least during the first wave of a pandemic. Dose range must be explored with and without adjuvants to ensure that the adjuvants truly augment the immune response.
• Dose Interval — “Accelerated” two-dose regimens should be examined. The interval between doses should not exceed 21 days. A short interval between doses is more likely to achieve protection.
• Antigenicity— Because H5 clearly still poses a pandemic threat, much more information on H5 is needed.
• Formulations/Adjuvants— Large, possibly multicenter, studies incorporating materials from different vaccine manufacturers and different adjuvants are urgently needed. The relationship of MF59 with avian antigens must be explored in greater detail.
• Surrogate Vaccines— Attenuated high growth reassortants containing the wild type HA should be used whenever possible instead of a not-ideally matched reference virus.
• Antibody Testing and Criteria— New criteria for pandemic vaccines may be warranted. International collaboration among laboratories is required to assess pandemic vaccines and ensure comparison of materials from different laboratories. Several antibody tests should be incorporated, including NA antibody testing. Consider the use of alternative systems (e.g., erythrocytes) to measure antibody response.
• Antibody Persistence— Future studies should consider the role of boosters in more detail.
• Age Effect— Age effect needs to be explored in future trials. Children and the elderly must be included in future studies.
• Statistics— A consensus needs to be developed about the most appropriate size for phase I and phase II studies.
Participants discussed, but were unable to definitively answer, the question of whether poor responses to vaccines with novel HAs indicate that the HAs are truly less immunogenic in some unique way or reflect what would occur with any vaccine used in a group that was totally unprimed. The results from additional clinical trials evaluating novel HAs is critical to answer this question.
Top of Page
Potential Pandemic Viruses: Priority for Vaccine Evaluations
Dr. Klaus Stöhr introduced the topic of criteria for choosing priorities for pandemic influenza strains. He reviewed the following criteria, which were discussed earlier this year at a meeting of World Health Organization (WHO) Influenza Collaborating Centers:
• Biological—The need for sera that has a broad spectrum reactivity
• Public health—Whether the strain has caused disease in humans or has appeared recently in mammals
• Epidemiologic—How widespread the strains are in avians, particularly domestic birds
• Nonbiological—Intellectual property (IP) issues and access to strains.
Accompanying activities identified at the meeting included addressing IP issues and defining safety conditions under which companies and institutions could work with the vaccines.
Dr. Robert Webster focused on the prioritization of animal influenza viruses for prospective vaccines and gave a brief update on the human/animal interface. He noted that 15 HA and 9 NA influenza A subtypes circulate in the aquatic birds of the world and periodically cross into mammalian species. Intermediary hosts involved in interspecies transmission of avian influenza viruses include pigs, quail, and chickens. Dr. Webster described the levels of priority to influenza strains, as depicted in the following table:
Level of Priority |
Virus Subtype |
Rationale |
Highest Priority |
H1, H2, H3 |
• Have caused widespread infection in humans with pandemic outcome (H1 and H3 in annual vaccine), |
High Priority |
H5, H6, H7, H9 |
• Have caused infection of humans but failed to demonstrate significant human-to-human spread (H5, H7, H9)
• Have caused infection of domestic poultry and have established lineages in poultry (H6)
• H7and H9 subtypes have been detected in pigs |
Lower Priority |
H4, H10, H13 |
• Have caused transitory infection of mammals and domestic poultry, including pigs and chickens |
Lowest Priority |
H8, H11, H12, H14, H15 |
• Are rarely found in land-based domestic avian species (H8, H11, H12)
• Are rarely found even in wild aquatic birds (H8, H14, H15) |
Dr. Webster reviewed several candidates for highest, lower, and lowest priority influenza strains and provided rationales and specific candidates in each category (See Appendix A). NA subtypes also must be taken into account. Dr. Webster outlined the following considerations for selecting viruses in a subtype:
• Screen viruses in a subtype for antigenic diversity using hyperimmune sera and ferret sera.
• Use phylogenetic analysis of HA sequences to detect diversity (i.e., number of families of subtypes that exist).
• Most subtypes will require two or more vaccines for coverage.
In the discussion that followed, Dr. Webster suggested that working at two levels to choose priority strains should be done by consensus: (1) developing a research strategy that allows strategic selection of a couple of strains each from subtypes H5 and H9 and conducting clinical trials to test different parameters and (2) monitoring events in nature to determine what viruses are being transmitted to humans and developing reference seed strains for those viruses, even if they are not tested in the first round of clinical trials. This approach will help build a knowledge base to address new emerging threats.
Top of Page
Manufacturing Risk Assessment Update and European Regulatory Update
Dr. Stöhr related that concerns raised by health safety agencies in the United Kingdom prompted the WHO Collaborating Centers and affiliated laboratories to discuss biosafety issues in the context of the production of influenza vaccines derived from reassortant avian strains. Dr. John Wood provided details on the document titled Production of Influenza Vaccines from Reassortants Derived from Avian Influenza Viruses: An Interim Biosafety Risk Assessment . He described the following components of a risk assessment such as a biosafety risk assessment:
• Identify the hazards that have the potential to cause harm to humans, scored in a range of negligible to very severe.
• Assess risks, or the likelihood of harm occurring, scored in a range from very unlikely to very likely.
• Take measures to control the residual risk indicated by the two scores.
• For a genetically modified organism, another consideration is to assess risk for the environment.
Dr. Wood made the following additional points:
• The potential risks for human health in producing vaccine pilot lots from reassortants include the recipient PR8 virus and inserted gene products with NA and modified HA from a highly pathogenic avian virus.
• The alteration of pathogenic traits as a result of the new combination of genes poses other potential hazards to human health. The removal of HA multi basic amino acids is a good safety procedure because available evidence indicates that it reduces pathogenicity for birds and mammals. Pathogenicity then should be tested in chickens, ferrets, and possibly mice.
• PR8 reassortants have been shown to be avirulent in man. However, the potential for transfer of sequences to related microorganisms exist and thus risk of secondary reassortment must be contained.
• To control this residual risk, biosafety level (BSL) 2+ containment is needed. Recommended alternatives for vaccine manufacturers were suggested.
Dr. Wood briefly described how influenza vaccines are currently licensed in Europe . The process includes a 73-day fast track procedure for the annual update of European influenza vaccine licenses. The procedure reviews quality data and immunogenicity data from small clinical studies. A European Vaccine Expert Group has developed a plan that would institute a new “rolling review” process to permit the ongoing review of pandemic vaccine data as they become available. Manufacturers would prepare a core market authorization (MA) dossier for a mockup vaccine during the interpandemic period. This dossier would contain quality, preclinical, and clinical data. The rolling review process could allow a pandemic MA dossier to be reviewed in a matter of days.
Top of Page
Regulatory Issues
Dr. Karen Midthun from the Food and Drug Administration (FDA) reviewed the routine licensing actions for influenza vaccines in the United States , noting that a manufacturing supplement to an existing license for annual strain changes does not require clinical data for FDA approval. During an influenza pandemic, the new pandemic strain used in a licensed manufacturing process would be reviewed as a manufacturing supplement for a strain change. Either a wild type or a reassortant virus acceptable to WHO National Influenza Centers could be used to produce the vaccine. A key question is whether one dose of a vaccine containing the 15 mcg of each HA specified for current vaccines will be optimal for a pandemic strain.
Dr. Midthun agreed that the preparations for the influenza pandemic must be made during the interpandemic period. Establishing the manufacturing base during this period could involve increasing manufacturing capacity by current licensed manufacturers in response to perceived demand and encouraging new manufacturers to apply for product approvals. Additional preparations for the pandemic that can be done now include the following:
• Investigational vaccines for influenza A subtypes with pandemic potential can be manufactured according to current licensed processes where feasible.
• Dose-ranging studies can determine the amount of HA and the number of doses needed to induce antibody response.
• The preparation of investigational vaccines with pandemic potential will require reference viruses and reagents and strain-specific antigen and antiserum to test the potency of the vaccine by Single Radio Immunodiffusion (SRID). Libraries of reference viruses and reagents also may be helpful.
Dr. Midthun suggested ways that an investigational inactivated pandemic vaccine could be developed for licensure. Any investigational vaccine made according to a new process would need a new license and would require product and preclinical data to support the initiation of clinical trials. Clinical studies would need to demonstrate the safety and efficacy of the new vaccine produced with current circulating strains. This demonstration would provide the basis for subsequently inserting different strains during the interpandemic period and then inserting the pandemic strain. For adjuvanted vaccines, additional preclinical and phase III studies will be needed. A clinical endpoint should constitute the primary efficacy endpoint for new inactivated influenza vaccines. Once a vaccine is licensed, FDA approval for strain changes requires a manufacturing supplement but not clinical data.
During the discussion that followed, Dr. Midthun and Dr. Roland Levandowski made the following points about regulatory considerations for pandemic vaccine development:
• Whole-virus vaccines for pandemic influenza should be considered.
• If a manufacturer develops an influenza vaccine with a lower dose of HA and an adjuvant to make more doses available worldwide, clinical outcome data are required to show efficacy.
• The FDA's animal efficacy rule, which accepts results from two animals models as surrogates for clinical data, pertains only to situations in which testing in humans is not ethical or feasible.
• The proposed legislation for Project BioShield includes a provision for emergency use authorization that may provide additional flexibility.
• Evaluating products for safety and efficacy is crucial to maintaining confidence in the vaccines after the pandemic is over.
• The FDA has mechanisms, such as informed consent under an investigational new drug ( IND ) application, for dealing with novel vaccines in a pandemic situation.
Representatives of manufacturing companies made the following comments on their capabilities:
• Most companies still use the egg-based process to develop strains for vaccine production, and new methods such as cell culture will not be ready for 3 to 5 years.
• It is important to consider using reverse genetics processes for pandemic vaccine production.
• Many companies do not have BSL-2, much less BSL-3, containment facilities.
• The only possible way to produce a sufficient number of influenza vaccine doses for a pandemic would be to use a monovalent, low-dose vaccine with an alum adjuvant that is filled in multidose vials.
• Data from clinical trials using currently circulating human strains would not be relevant to evaluating a pandemic vaccine designed to protect naïve subjects.
• It would be difficult to persuade company executives to invest in the evaluation of a vaccine product that might never be used.
Top of Page
Workshop Discussion: Preclinical Tools and Issues
Dr. Nancy Cox provided an overview of issues and options for the discussion about preclinical tools and issues. She discussed needs related to seed viruses for manufacturers, substrates for production, reagent requirements, safety considerations, and other regulatory and IP issues. Drs. Lambert and Levandowski moderated the discussion on the following points.
Availability of Pandemic Reference Strains
The following reference strains are currently available:
• H7N7 from the Netherlands with a cleaved HA on the PR8 (proof-of-principle done outside GMP facilities)
• H5N1 (GMP produced)
• H9N2 G9-like candidate was successfully sent to European company (classical reassortment)
The following reference strains are under development:
• H7N1 from the Italian highly pathogenic virus
• H7N7 using the apathogenic Netherlands H7 and H10N7 from the Netherlands with appropriate NA on a PR8 background (using classical reassortment techniques); also using North American H7N3 apathogenic virus on a PR8 background)
Considerations in Prioritizing Pandemic Strains
• Continuing to generate as many strains as possible to develop a library of reference strains and reagents for those strains might be useful when a pandemic threat appears.
• It must be decided early which one or two particular avian strains, preferably those that already have been made, should be selected as candidate pandemic-like vaccines that could go into clinical trials.
• Choosing avian strains that are potential candidates for experimental vaccine trials requires a more organized or standardized approach to assessing the antigenicity and antigenic variation in avian strains.
• Efforts to choose candidate strains are hampered by the continual appearance of new strains.
• It may not be important to select specific strains after subtypes have been chosen because there is almost no chance of correctly predicting which strain will cause the pandemic.
• Achieving consensus on the selection of strains could facilitate matches between preclinical work (where animal models can help establish correlates of protection) and clinical trials.
• It is useful to understand the antigenic characterization of the variety of viruses that are circulating in nature.
• Many manufacturers are limited in their choice of strains by local regulatory authorities and quarantine authorities.
Clinical Trials Issues
• Vaccine clinical trials must focus on building a logistical framework by testing the process of producing a pandemic vaccine and should address issues such as production scale-up, adequate inactivation and filling capacity, and approval of reverse genetics processes.
• It should be decided whether the clinical trials should be conducted for licensing purposes, to generate additional information on the immunogenicity of strains, or both.
• Two possible approaches include (1) working with as many potential pandemic vaccine candidate strains as possible to determine how they can be grown in eggs or cell-culture systems, inactivated, and made into preclinical vaccine preparations or (2) concentrating on two or at most three pandemic vaccine candidates, making GMP material for clinical trials, and looking at vaccine dose, boosters, and immune correlates.
Genetically Modified Organism (GMO) Issues
• The only way to produce a pandemic vaccine from an avian strain is to use a seed derived from reverse genetics, which would be considered a GMO for European authorities.
• A major goal of clinical trials in the interpandemic period should be to identify potential problems associated with regulatory concerns about GMOs and to work out how manufacturers can overcome these problems before a pandemic occurs.
• To test regulatory reactions before a pandemic in Europe , a vaccine manufacturer would need to begin by drawing up a proposal based on the WHO model risk assessment, the virus it intends to use, and its facilities.
Intellectual Property Issues
• Dr. Kathy Coelingh from Medimmune Vaccine, Inc., explained that the company intends to make its IP freely available to support developmental and clinical studies conducted to determine the correct antigen content, adjuvants (if any), doses, and dosing interval for pandemic vaccines. If a company commercializes a vaccine, Medimmune intends to negotiate in good faith for a reasonable price that would be a percentage of sales, if the vaccine were sold. This position allows for a company under Government contract to produce an investigative lot of a pandemic vaccine candidate with a seed created by reverse genetics.
• Manufacturers would be more comfortable knowing the company's requirements before they invest in clinical trials and product development.
• Intellectual Property Rights: Innovation in Public Health , prepared for the World Health Assembly (WHA) in May 2003, provides a discussion of IP issues.
• Key international and national public health agencies could provide nonindustry leadership and allow a forum for companies to discuss this problem collectively without raising antitrust issues.
• FDA has not viewed the use of a reverse genetics virus for an influenza vaccine as a regulatory hurdle if the seed is developed from an acceptable source and the FDA is satisfied about safety issues.
• The National Vaccine Program Office should undertake more indepth discussions about IP issues regarding reverse genetics; an international process involving all companies that have or claim to have IP rights in this area should be initiated.
Select Agent Rule
• As of February 2003, highly pathogenic H5 and H7 are considered select agents in the United States .
• It is important to determine whether a H5N1 that is no longer highly pathogenic would still be considered a select agent.
Availability of Reagents
• Reagents for various subtypes of influenza are available through NIAID's repository and are being shared with WHO.
• NIAID currently supports reagent preparation of baculovirus-expressed HA and NA proteins, including antigens for the preparation of monospecific antiserum, antigens for ELISA screening, and antigens to internal proteins. However, these preparations are likely to produce serum that is more specific than those made with HA isolated from the virion. It is important to determine whether human sera have the same degree of specificity when given those antigens.
• Identification of priority pandemic strains will help pharmaceutical companies determine their needs for reagents.
Clinical Trials: Interpandemic and Pandemic
Dr. John Treanor introduced the session on a clinical trial plan for pandemic influenza virus candidates by raising several questions for discussion. Input from participants on these questions is summarized below.
Problems That Need To Be Addressed Before Vaccine Evaluation Can Begin
• Before it can develop a pandemic vaccine, industry must receive the seed and be told how to formulate a vaccine that will provide protection.
• Companies may be willing to take initial developmental trials to the stage where they can identify how to make a vaccine, but then they may not be willing to do efficacy or large phase III trials unless regulatory arrangements are more flexible.
• A regulatory requirement that influenza vaccines with adjuvants must undergo phase III efficacy trials for licensure will discourage companies from manufacturing anything but monovalent vaccines, which will not meet needs worldwide. Regulatory concerns must be balanced with the public health imperative to vaccinate and protect as many people in the world as quickly as possible during a pandemic.
Scope and Purpose of Clinical Trials
• Consider developing clinical trial protocols to look at a range of dose concentrations (based on HA content) to inform a decision on the optimal dose of antigen to use in a vaccine that is highly likely to be in short supply.
• Trials are needed to determine whether different strains have different antigenicity.
• Clinical trials should focus on current deficiencies in information, such as how to improve immunogenicity for H5 vaccines (e.g., whether an adjuvant is required, whether drifted strains behave differently). At least two antigens should be tested at the same time to see whether they behave the same way.
• Studies of H2N2 are likely to give different kinds of answers because the virus has been in the human population.
• Priority should be given to working with different subtypes that have not yet been done (i.e., H7 or H6), with H7 being more important because it has been associated with human disease.
• Clinical trials with currently licensed processes need to look at issues such as ways to measure antibody response and show immunogenicity. Clinical trials for vaccines using new manufacturing processes or vaccine approaches (e.g., adding an adjuvant) must gather information to support the process, demonstrate that adjuvants add significantly to the protective capability for primed and unprimed populations, and address safety issues associated with the new process or approach. Information on dose ranges and schedules also must be gathered before a pandemic.
• Three variables must be narrowed down: (1) virus subtypes, (2) use of egg-based or cell-based substrates, and (3) adjuvants.
• For each process, whether new or already licensed, examine the following: (1) robustness with several subtypes, (2) choice of strains that every manufacturer could use to enhance comparability of data, (3) the ability to bridge to already available data, and (4) continued evaluation and optimization of adjuvants.
• Explore tissue culture grown inactivated vaccines. It might be valuable to compare cell-culture-derived seed viruses with egg-derived seed viruses in Vero, MDCK, and egg systems.
• Allow each company to follow the vaccine approach it thinks is best, rather than comparing different approaches in one clinical trial (e.g., egg vs. cell culture, split virus vs. purified surface antigen, or conventional reassortant vs. reverse genetics).
• Conduct fairly simple, large international trials, with sera exchanged, to find statistically significant differences for dose range and dosing intervals and among different populations. Look at current methodologies, choose one or two subtypes, and expand the number of immunogenicity studies. Conduct trials on a smaller scale for experimental approaches (e.g., particular adjuvants, cell-based vaccines).
• Conducting immunogenicity and safety assessments in one laboratory would make the resulting data more meaningful. It might be easier to compare vaccines produced by different methods after determining how certain serologic markers correlate with protection in vaccines produced for interpandemic strains.
Outcomes
• Nonclinical (i.e., serologic) outcomes will be necessary particularly for viruses classified as select agents. Consider accepting tests that indicate a likely correlate to immunity (e.g., producing better antibodies indicates better protection) in lieu of phase III efficacy trials.
• A better understanding of the immune response to avian viruses, including how to interpret both protection and population-based susceptibility, is critical to determining what trial endpoints should be used in lieu of clinical efficacy studies.
Immunologic Assays
• Work on avian strains that have gone into human vaccines indicated that using a single measure of antibody response (i.e., HI response) is not enough. The neutralizing antibody response is potentially useful but requires more experience to clearly understand the problems in standardization.
• The Single Radial Haemolysis (SRH) test was valuable in helping to determine protection against H5 but is technically demanding and must be controlled carefully.
• Other immunologic measurements that should be used in vaccine clinical trials include NA antibody, secretion antibodies, and cell-mediated immunity.
• Assays for cell-mediated immunity could be useful, and newer immunologic assays (e.g., intracellular cytokines) are available. However, some of these assays are difficult to conduct and may not be sufficiently sensitive for an inactivated vaccine.
• Measures of mucosal immunity would be worth studying. Nasal secretion antibodies currently are the only surrogate for what is happening at other sites and should be measured in the clinical trials. This measure will be particularly useful when testing adjuvants.
Dose Ranges
• Dose range should be explored as much as possible, depending on what companies can produce realistically.
• One consideration is the lack of a one-to-one correspondence between increases in HA and increases in HA titers. Studies of dose response curves in the 1970s showed that HA had to be increased eightfold to tenfold to get a twofold increase in HA titers.
• Not many dose range options are available besides 7.5 and 15 µg per dose in one or two administrations.
Study Populations
• When dealing with an entirely novel strain, studies in adults can provide most of the needed information.
• Extensive safety information should be collected in adults before any testing on children.
• Studies in children are needed for reactogenicity data.
• The elderly population is growing worldwide and should be included in the trials. The inclusion of frail elderly populations requires further consideration.
• Immunocompromised individuals should be included later in the trials.
Draft Plan for Clinical Trials
Drs. Lambert and Couch summarized key points raised in the previous day's discussion and presented recommendations based on the consensus of the meeting's moderators. Although recommendations are not complete in some areas, the plan is to develop a roadmap to guide efforts so that no time is lost when a new pandemic strain appears. Participants were asked for their reactions and additions to the recommendations. The following summaries present key points as articulated by the moderators, recommendations as revised through group deliberation, and highlights of the group discussion.
Preclinical Tools
Participants considered recommendations about the reagents and references strains that would be used in the clinical trials.
Reagents
Key Points
• Current methods for production of reagents (i.e., baculovirus-derived HA, goat and sheep antisera supported by NIAID) is likely to be acceptable if the primary interest is standardizing vaccine potency.
• If the interest focuses on surveillance and the ability to detect several different strains, additional studies are needed to determine whether baculovirus-derived HA provides more broadly reactive antisera.
Recommendation
• As investigational lots of vaccines are produced for clinical trials, manufacturers should consider providing small amounts of purified HA for production of goat/sheep antisera.
Discussion
• Individuals with an interest in the production of a particular reagent should contact Dr. Lambert.
• Instead of producing reagents for vaccine standardization in just one laboratory, use a small collaborative network of laboratories to share the generation and distribution of reagents for novel antigens. Confirming the calibration of new reagents would help quantitate the amount of antigen in a vaccine.
• It would be very useful for expanded clinical trials if manufacturers could produce larger quantities of antigen.
• Vaccine manufacturers should identify the scale of clinical trials early and notify Dr. Lambert, Dr. Levandowski, or Dr. Wood to ensure a supply of reagents adequate to meet demand.
• A comparison of isolates derived from eggs and those derived from mammalian cells should be undertaken in the future.
Reference Strain Production
Recommendations
• Continue to make high growth reference strains as potential vaccine candidates in response to the identification of “new” avian influenza viruses with pandemic potential.
• Produce reference strains to the wild type isolate and suitable nonpathogenic viruses that could be used as a reference virus, if possible.
• Produce reference strains by both classical reassortment and reverse genetics, recognizing that candidates with deleted cleavage sites must be made by reverse genetics.
• Coordinate the preparation and characterization of reference strains that could be available to manufacturers.
Discussion
• Reference strains based on the live-attenuated, cold-adapted background should be made and investigated for its potential.
• Coordination of the preparation of reference strains should entail not only a clear inventory of the availability of strains but also a preliminary demonstration that they grow well in eggs and the provision of clear directions for which strains to use.
• For vaccine reference strains, information on how well the virus grows in eggs and the overall vaccine yield can only be obtained once manufacturers start working with the virus and learn how well the strain(s) grows in their respective processes.
Plan for Manufacture of Pandemic Vaccines
The group discussed several aspects of how materials for pandemic influenza vaccines would be manufactured. Recommendations focused on the prioritization of subtypes and parameters for production.
Prioritization of Subtypes
Key Point
• The likelihood of predicting the influenza subtype or the strain of the subtype that will cause the next pandemic is questionable.
Recommendations
• Two subtypes should be selected, and they should proceed as close in tandem as possible through vaccine manufacture and clinical evaluation.
• The highest priority subtype is H5N1. [This priority was determined by general consensus.]
• The second highest priority is H2 or H7. [Group preference was nearly evenly divided between H2 and H7. The meeting moderators had initially proposed H7, H9, or H6.]
• If a problem arises bringing either of the two strains forward, a third, alternative strain (preferably H9 or H6) should be added.
• Interpandemic vaccine preparations and trials should move forward with the recognition that IP issues leading to commercialization still must be resolved.
Discussion
• H2 should be considered as a fairly high priority. The advantages of H2 include experience working with the strain, the availability of both primed and unprimed populations, the fact that H2 is a mammalian strain, and the current circulation of H2 in the poultry in China .
• A combination of H5 and H2 is preferred from a development and production perspective because manufacturers would like to work with as heterogeneous a combination as possible for clinical trials.
• Clinical trials should include work on vaccines that pose significant regulatory and scientific challenges. Although working with these viruses would require more attention to regulatory issues, they would provide a good model for developing effective ways to address these concerns during a pandemic.
• A second strain of a priority subtype could be evaluated if the immunogenicity of the initial strain is poor.
• Manufacturers who want to make a vaccine using a subtype not recommended by consensus as a priority can do so and have it considered by the clinical trial apparatus being discussed. However, most manufacturers prefer guidelines about which vaccine they should make and are unlikely to select a subtype that is not recommended as a priority.
• Companies are not likely to embark on producing a pandemic strain unless the seed and reagents are available. Consequently, an inventory of reference strains and reagents that are available to support the recommended priorities is critical.
• Companies also must ensure that the reference strains they receive were derived from sources that meet regulatory requirements and are qualified to support advanced product development. This concern can be addressed up front through a consultation among all global partners involved in vaccine preparation.
• The decision to select two, rather than three, priority candidates was based on the perceived feasibility for clinical trials because the number of different variants and products that could be studied for each subtype increases with each additional candidate.
• It was noted that only one industry representative voted for H7 in the nearly evenly divided poll on preferences for H2 or H7 as the second priority subtype.
• The selection of two candidates represents only a prioritization. All subtypes must eventually be evaluated.
Parameters for Production
Recommendations
• At least two different reference viruses per subtype should be evaluated clinically.
• The highest priority in the United States is the production and evaluation of investigational lots of vaccine using conventional/licensed processes. Whole virus vaccines should be considered.
• Companies making investigational lots of vaccines for clinical trials should first use the manufacturing processes that they will propose to use for commercial production in response to a pandemic. In the United States , this use entails conventional/licensed methodologies.
• Companies should plan to use the same methodology to produce investigational lots of vaccines against the two different subtypes.
• One subtype should be H5N1, and the other subtype will be determined.
• Companies are encouraged to work with the reference strain made from the pathogenic virus or, if not possible, with the nonpathogenic surrogate.
• Companies are encouraged to produce lots of investigational vaccines that are “unlicensed” but may provide conservation of antigen or other advantages (e.g., with cell culture base, adjuvants, or novel delivery systems). If possible, these products should be evaluated in parallel with vaccines made by licensed methods.
Note: No recommendation was made to directly compare vaccines produced by eggs with those produced by cell culture, or candidates made by reverse genetics with those made by traditional reassortment.
Discussion
• Companies developing an investigational product (e.g., vaccine lots for investigative use with a new pandemic antigen) can submit an IND application, which requires less information than a new license application. Companies that manufacture the material according to a licensed process will not need to provide as much documentation as they would for an unlicensed process.
• The FDA will use flexibility in deciding whether it views changes in facilities (e.g., use of high containment laboratories) as new processes.
• It is not practical for manufacturers to change their facilities to suit the safety issues of a new seed or process. Consequently, the seed virus provided to industry must be prepared so that its safety closely matches the capabilities of present facilities.
• Questions about the respective roles of various entities in the clinical trials (e.g., whether manufacturers will supply material to a central body for use in clinical trials or undertake their own clinical trials) have yet to be answered.
• NIAID coordinates grant support for the preparation of pandemic vaccines in the United States with an interagency working group and plans to test vaccines made by industry in its Vaccine Treatment and Evaluation Units (VTEUs).
• Because pandemic influenza is a global issue, international coordination of clinical trials is important. Agencies such as the European Medical Evaluation Agency, and the National Institute for Biological and Serological Standards are expected to play a role (e.g., in providing reagents), and resulting products should be available to investigators outside the United States .
• The FDA encourages European companies that plan to pursue licensure in the United States to do their investigational development under an IND application. However, an IND is not required to use European studies and study results in the licensure process.
• Dr. Lambert envisioned two processes for supporting the pandemic influenza clinical trials process. One avenue is based on the strong interest of the National Institutes of Health (NIH) in supporting the clinical development of new and improved vaccines to control influenza. NIAID largely supports partnerships with industry to evaluate new vaccines and antiviral drugs through clinical contract sites (VTEUs) that conduct Phase I/II clinical trials. A second avenue for support from NIAID has been instituted through funding opportunities for biodefense. Under this mechanism, a recent request for applications (RFA) specifically targets manufacturers willing to undertake a pandemic influenza vaccine development program on their own. Although collaboration with one or more other entities is optional, the RFA does not include the large comparative trial that had been discussed during the current workshop. However, the NIH would consider a joint proposal by companies interested in undertaking such a trial.
• The Influenza Vaccine Supply Task Force (IVS Task Force) should determine what it can do alone and what roles other public health agencies must play. The task force should communicate with the NIH and the WHO Global Influenza Programme about these decisions.
• Egg-based vaccines versus tissue culture-based vaccines should be explored, but head-to-head comparisons would be best done during the interpandemic stage.
Clinical Trial Design
Key Point
• The goal of the clinical trial is to identify the appropriate dose and dosing schedule that is most effective for a particular subtype.
Recommendations
The clinical trial design should include the following elements:
• Two-dose regimen, given approximately 1 month apart
• Collection of data on antibody persistence for at least 6 months
• A sample size of at least 25 subjects per group, preferably 50 or more per variable (To be evaluated to obtain appropriate statistical input for the study question)
• Dose-ranging studies, which will include the following:
• Evaluation (for conventional products) of threefold increases in dose (e.g., 5 mcg, 15 mcg, 45 mcg)
• Products that contain adjuvants can extend the dose response curve lower than 5 mcg . Head-to-head comparisons could be made at all dose levels with products that do and do not have adjuvants—if appropriate. Alternatively, consider evaluating the highest dose without adjuvant and the lowest dose with adjuvant. Comparing the lowest formulated dose with and without an adjuvant can provide needed information as to whether a particular adjuvant improves immunogenicity a different at very low antigen dosage.
• Considerations of age groups include the following:
• Use healthy young adults for all primary evaluations of crucial data.
• Consider elderly and pediatric subjects if follow-on studies are needed.
• Immunocompromised individuals and those with other high-risk disorders do not need to be tested.
• Opportunities for priming should be assessed among different age groups.
Discussion
• For trials of products with adjuvants, consider evaluating the highest dose without adjuvant, the lowest dose with adjuvant, and no lowest dose without adjuvant. However, comparing the lowest dose with and without adjuvant may provide important information currently lacking on whether adjuvants actually make a difference at low doses.
• The recommended sample size is per variable, which includes each company's product. The stated sample sizes are for the initial stage of the trial and will be larger for the more definitive follow-on stage.
• Studies of elderly and pediatric individuals may have fewer variables (e.g., only two different doses) because it may be difficult to achieve large enough sample sizes in these populations. Studies in healthy young adults will guide what questions need to be asked in children and the elderly.
• Small phase I studies might be conducted in children before deciding whether to include them in expanded studies.
• If an H2N2 vaccine is evaluated, the population of young adults must be defined at a breakpoint for past exposure to H2.
• No genetic homology data exist that would suggest that priming with H1, H2, or H3 would influence results with any subtype other than H9. Cross-protective immunity has not been well studied, but some recent data suggest that the source of the antigen (i.e., eggs or cell culture) may be important. Joint analysis of the data regarding priming is needed.
Immunogenicity Evaluations
Recommendations
• The primary immune outcome variable should be serum HAI antibody.
• Use additional assays, including HI, microneutralization test, anti-NA antibody, mucosal antibody, and possibly cellular immunity.
• Consider standardizing immunoassay results across laboratories by exchanging sera.
• Specimens developed in the clinical trial will be shared on an international basis with appropriately interested persons, within the limits of availability. Clinicians will decide how much blood can be drawn and made available.
Note: It may be worthwhile to develop a set of standardized antisera for each subtype that can be used for comparing results between laboratories; standardized assays for some of these will be most appropriately done in single sites.
Discussion
• Multiple measurements and assays are encouraged to assess the broadening of the immune response and to provide a frame of reference for comparing different vaccines of different subtypes.
• For a set of trials of different vaccines from different companies, an absolute, standardized way of measuring immune responses is critical for comparability. Conducting all tests at one central laboratory is a good way to standardize. The establishment of reference sera and sharing them between laboratories is a good compromise.
• Consider building in a plan to assess response not only to the vaccine strain but also to a range of different strains within the subtype.
• Consider gathering information on the kinetics of the immune response, although the amount of blood that can be draw from study subjects may be a limitation, particularly in children. Another consideration is the need to have enough material to share with many laboratories.
• Consider including a trivalent inactivated vaccine contemporary comparison as one arm of the larger trial.
• As part of NIAID's pandemic preparedness efforts, Dr. Treanor has developed an outline of a protocol for the generic evaluation of inactivated vaccines for pandemic influenza that can be used and quickly modified for specific products. NIAID could make the document available to manufacturing companies for use as a model after it has undergone some additional review.
Discussion of Other Issues
Participants made the following additional comments on safety and regulatory issues:
• The risk assessment described by Dr. Wood was designed for a reverse-genetics virus derived from a highly pathogenic strain. However, the level of containment in that assessment could be used for other candidates not made with reverse genetics.
• Manufacturer liability issues are difficult to separate from safety issues and must be addressed in advance of any pandemic threat.
• Consider imposing safety restrictions for laboratories working with H2N2 because young adult employees working in these facilities could be susceptible to the virus.
• During the interpandemic period, new vaccines will require clinical development discussions with the FDA, particularly regarding efficacy studies with clinical endpoints.
• Given the unique, unprecedented product that the influenza pandemic vaccine represents, the FDA must consider avenues for modifying the phase III licensure requirements.
• The experience with the licensure of a cold-adapted influenza vaccine as well as other licensed products (e.g., the hepatitis B vaccine) could be used for an understanding of what it would take to obtain licensure of an inactivated vaccine with a new adjuvant.
• The view that an adjuvanted influenza vaccine is a new vaccine that requires clinical efficacy data is a huge issue for companies and must be discussed further.
Acknowledgements:
The NIAID Influenza Program wishes to thank all of the participants for their very thoughtful and informative contributions during the meeting to develop a clinical trial plan for evaluating pandemic influenza vaccines. We believe that this document reflects discussion points and recommendations from a wide variety of participants. Additionally, Linda Lambert would like to acknowledge Sonnie Kim, Program Officer, DMID for all of her work organizing this meeting. It would not have been possible without her efforts.
Top of Page
Appendix A: Vaccine Candidates With Rationales
Subtype |
Rationales |
Candidates |
Highest Priority |
H2 |
• Proven ability to cause pandemics
• Susceptible population younger than 35 years of age
• Continues to circulate in wild aquatic birds |
• A/Singapore/1/57 (H2N2) Ferret serum available
• A/Japan/305/57 (H2N2) Ferret serum available
• Easy, but not put in place |
H5 |
• Proven ability to infect humans
• Potential for high pathogenicity in humans and birds
• Still circulating and causing disease in Southeast Asian poultry and wild birds |
• A/Hong Kong/156/97 (H5N1) Ferret Serum available
• A/Hong Kong/213/03 (H5N1) Ferret Serum available
|
H7 |
• Proven ability to infect humans, horses, pigs, monkeys, and seals
• Potential for high pathogenicity in birds and humans
• Has caused lethal outbreaks in poultry |
• A/Netherlands/03 (H7N7)
• A/Ck/Italy/apathogenic (H7N1) ferret serum available
• Recent viruses from H7 outbreaks in Virginia/USA
• A/TK/Oregon/1/71 (H7N3) |
H9 |
• Proven ability to infect humans
• Recent strains possess a2,3 and a2,6 receptor specificity
• Genetically related to H5N1 virus (G1 lineage) [internal gene segments]
• Widespread circulation in Eurasia |
• G1 lineage A/HK/1073/97 (H9N2) ferret serum available
• A/Quail/HK/G1/97 (H9N2) ferret serum available
• G9 lineage A/Ck/HK/G9/97 (H9N2) [HGR with PR8 background exists] ferret serum available
• A/Turkey/Wisconsin/66 (H9N2) |
H6 |
• H6N1 virus genetically related to H5N1 and H9N2 viruses
• Widespread circulation in Southeast Asian birds and recent increasingly widespread circulation in domestic poultry in North America |
• A/Teal/HK/97 (H6N1)
• A/CK/CA/465/2000 (H6N2)
• A/Turkey/Massachusetts/65 (H9N2)
• A/CK/Hong Kong/S40/99 (H6N1) chicken serum available |
Lower Priority |
H4 |
• Isolated from pigs and seals |
• A/Swine/Ontario/1911-1/99 (H4N6) |
H10 |
• Isolated from mink |
• A/Mink/Sweden/84 (H10N4) |
H13 |
• Isolated from whales |
• A/Whale/Maine/1/84 (H13N2) reference strain available to H4, H10, H13 |
Lowest Priority |
H8, H11, H12, H14, H15 |
• Rarely ever isolated even from wild aquatic birds. However, reference reagents are considered important in the repository. |
[none identified] |
