W S p R V S W R A M A T a b c d e f g h i j k l m n o p q r a A R A A K I V W h 0 0 Vaccine 33 (2015) 4368–4382 Contents lists available at ScienceDirect Vaccine j o ur na l ho me page: www.elsev ier .com/ locate /vacc ine HO Report trengthening the influenza vaccine virus selection and development rocess eport of the 3rd WHO Informal Consultation for Improving Influenza accine Virus Selection held at WHO headquarters, Geneva, witzerland, 1–3 April 2014 illiam K. Ampofoa, Eduardo Azziz-Baumgartnerb, c b Uzma Bashir , Nancy J. Cox , odrigo d e Fasce , Maria Giovanni , Gary Grohmannf, Sue Huangg, Jackie Katzb, lla Mironenkoh, Talat Mokhtari-Azadi , Pretty Multihartina Sasonoj, ahmudur Rahmank, Pathom Sawanpanyalert l, Marilda Siqueiram, nthony L. n o Waddell , Lillian Waiboci , John Woodp, Wenqing Zhangq,∗, hedi Ziegler r, WHO Writing Group1 Noguchi Memorial Institute for Medical Research, Accra, Ghana Centers for Disease Control and Prevention (CDC), Atlanta, GA, USA National Institute of Health, Islamabad, Pakistan Public Health Institute of Chile, National Influenza Center, Chile National Institutes of Health, Bethesda, MD, USA Therapeutics Goods Administration, Symonston, Australia National Influenza Center, Upper Hutt, New Zealand Reference Influenza Laboratory, Kiev, Ukraine National Influenza Center, Tehran, Islamic Republic of Iran National Institute of Health Research and Development, Jakarta, Indonesia Institute of Epidemiology, Disease Control and Research, Dhaka, Bangladesh National Institute of Health, Bangkok, Thailand Instituto Oswaldo Cruz, Rio de Janeiro, Brazil Freelancer, Stanley, UK CDC Kenya, Nairobi, Kenya Formerly National Institute for Biological Standards and Control (NIBSC), Potters Bar, UK World Health Organization (WHO), Geneva, Switzerland National Influenza Center, Helsinki, Finland r t i c l e i n f o a b s t r a c t rticle history: Despite long-recognized challenges and constraints associated with their updating and manufacture, eceived 21 June 2015 influenza vaccines remain at the heart of public health preparedness and response efforts against both ccepted 23 June 2015 seasonal and potentially pandemic influenza viruses. vailable online 3 July 2015 Globally coordinated virological and epidemiological surveillance is the foundation of the influenzavaccine virus selection and development process. Although national influenza surveillance and reporting eywords: capabilities are being strengthened and expanded, sustaining and building upon recent gains has become nfluenza vaccine viruses a major challenge. accine virus selection Strengthening the vaccine virus selection process additionally requires the continuation of initiativesHO recommendations to improve the timeliness and representativeness of influenza viruses shared by countries for detailed analysis by the WHO Global Influenza Surveillance and Response System (GISRS). ∗ Corresponding author. E-mail address: GISRS-WHOHQ@who.int (W. Zhang). 1 For Declarations of Interest see Annex 1. ttp://dx.doi.org/10.1016/j.vaccine.2015.06.090 264-410X/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4. /). W.K. Ampofo et al. / Vaccine 33 (2015) 4368–4382 4369 Efforts are also continuing at the national, regional, and global levels to better understand the dynamics of influenza transmission in both temperate and tropical regions. Improved understanding of the degree of influenza seasonality in tropical countries of the world should allow for the strengthening of national vaccination policies and use of the most appropriate available vaccines. There remain a number of limitations and difficulties associated with the use of HAI assays for the antigenic characterization and selection of influenza vaccine viruses by WHOCCs. Current approaches to improving the situation include the more-optimal use of HAI and other assays; improved understanding of the data produced by neutralization assays; and increased standardization of serological testing methods. A number of new technologies and associated tools have the potential to revolutionize influenza surveil- lance and response activities. These include the increasingly routine use of whole genome next-generation sequencing and other high-throughput approaches. Such approaches could not only become key ele- ments in outbreak investigations but could drive a new surveillance paradigm. However, despite the advances made, significant challenges will need to be addressed before next-generation technologies become routine, particularly in low-resource settings. Emerging approaches and techniques such as synthetic genomics, systems genetics, systems biology and mathematical modelling are capable of generating potentially huge volumes of highly complex and diverse datasets. Harnessing the currently theoretical benefits of such bioinformatics (“big data”) concepts for the influenza vaccine virus selection and development process will depend upon further advances in data generation, integration, analysis and dissemination. Over the last decade, growing awareness of influenza as an important global public health issue has been coupled to ever-increasing demands from the global community for more-equitable access to effective and affordable influenza vaccines. The current influenza vaccine landscape continues to be dominated by egg-based inactivated and live attenuated vaccines, with a small number of cell-based and recombinant vaccines. Successfully completing each step in the annual influenza vaccine manufacturing cycle will continue to rely upon timely and regular communication between the WHO GISRS, manufacturers and regulatory authorities. While the pipeline of influenza vaccines appears to be moving towards a variety of niche products in the near term, it is apparent that the ultimate aim remains the development of effective “universal” influenza vaccines that offer longer-lasting immunity against a broad range of influenza A subtypes. ublis 1 R n a u t a w b e p u b o u p a l b v i p t a a t t 2 s f © 2015 The Authors. P . Introduction For over 60 years the WHO Global Influenza Surveillance and esponse System (GISRS)1 has served as the foremost global coordi- ation mechanism for monitoring and responding to the evolution nd spread of influenza viruses, and ensuring the use of the most p-to-date vaccine formulations. The GISRS vaccine virus selec- ion process involves the coordinated collection and laboratory nalysis of hundreds of thousands of clinical specimens each year, ith the goal of determining which vaccine compositions will est protect against disease during upcoming northern and south- rn hemisphere influenza seasons. Due to severe time and other roduction constraints inherent in current influenza vaccine man- facturing technologies, the vaccine virus selection process must e completed almost a year in advance of the predicted peak f influenza activity in the season in which the vaccine is to be sed. The GISRS also continually monitors and assesses the risks osed by potential pandemic viruses and provides guidance on ppropriate public health responses. In recent years, data simi- ar to that used for seasonal influenza vaccine development have een used to select viruses for use in 2009 A(H1N1) pandemic accines, and in vaccines against other influenza virus subtypes, ncluding A(H5), A(H7) and A(H9) for pandemic preparedness urposes. In 2010, the convening of the first WHO Informal Consulta- ion for Improving Influenza Vaccine Virus Selection provided unique opportunity to review in detail this highly complex nd collaborative process [1]. Building upon the outcome of his review, a second consultation was held in 2011 to discuss 1 Formerly known as the Global Influenza Surveillance Network prior to the adop- ion of the World Health Assembly Resolution WHA 64.5 on 24 May 2011. As of May 014, the GISRS consisted of 141 National Influenza Centres (NICs) in 111 countries, ix WHO Collaborating Centres (WHOCCs), 12 WHO H5 Reference Laboratories and our WHO Essential Regulatory Laboratories (ERLs).hed by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). the key principles of influenza surveillance and representative virus sharing, the virological characterization of candidate vac- cine viruses, vaccine manufacturing and regulatory requirements, and the potential application of new and emerging vaccine technologies [2]. In the intervening period between these two meetings, the Pandemic Influenza Preparedness (PIP) Frame- work for the sharing of influenza viruses and access to vaccines and other benefits was adopted. This important milestone event reflected growing recognition of the importance of the timely sharing and characterization of viruses, and of the equitable provision of effective vaccines against both seasonal and pan- demic influenza. In response, WHO has continued working to improve knowledge of the global patterns of influenza activity; support the development of informed national policies, aided by the work of its Strategic Advisory Group of Experts (SAGE) on Immunization; increase global influenza production capacity and supply as part of its Global Action Plan for Influenza Vaccines (GAP); and promote expanded access to vaccines under the PIP Framework. Since the re-emergence of human cases of H5N1 influenza (“bird flu”) in 2003 and the 2009 H1N1 pandemic, growing awareness of influenza as an important threat to public health has driven an expansion of surveillance and response capacities in many countries. Nevertheless, many countries are now facing major chal- lenges in sustaining and building upon the gains made. In light of recent national, regional and global initiatives to promote efficient surveillance and representative virus sharing, allied to ongoing advances in vaccine development and production technologies, it was felt timely to convene a third WHO informal consultation in order to:• update participants on the progress made since the previous meeting; • further discuss surveillance as the foundation of vaccine virus selection; 4 ccine • • • • e p 2 ( I a ( o s 2 r i 2 i p e a i b u s s • • • • N s t t t c t o t m v p A d d i 370 W.K. Ampofo et al. / Va discuss newly emerging insights into the circulation and viro- logical characteristics of influenza in tropical regions with the potential to strengthen vaccine composition and deployment decisions; discuss new assays, new technologies and new approaches and their potential for bringing about improvements in both vaccine effectiveness and manufacturing efficiency; discuss the regulatory and other practical issues that must be considered in relation to both existing and emerging vaccine technologies; and continue to provide a forum for stakeholders to review and evaluate potential improvements to the influenza vaccine virus selection process. Approximately 128 participants drawn from 51 countries cov- ring all six WHO regions, and representing a wide range of WHO artner organizations and other stakeholders attended (Annex ). Participants were drawn from WHO Collaborating Centres WHOCCs), WHO Essential Regulatory Laboratories (ERLs), National nfluenza Centres (NICs), WHO H5 Reference Laboratories, the cademic research community, National Regulatory Authorities NRAs), national public health agencies, veterinary institutions and rganizations, vaccine manufacturers, donor agencies and other takeholders. . Strengthening influenza surveillance and improving the epresentativeness, timeliness and availability of candidate nfluenza vaccine viruses .1. Efforts at national, regional and global levels National influenza surveillance, reporting and response capabil- ties continue to be strengthened and expanded in many countries, articularly during and since the 2009 H1N1 pandemic. How- ver, despite increasing awareness of the incidence, transmission nd disease burden associated with both seasonal and year-round nfluenza activity, and the continuing pandemic threats posed y A(H5N1) and other zoonotic viruses, sustaining and building pon recent gains has become a major challenge. Key elements in trengthening and sustaining influenza surveillance and response ystems include: National surveillance system building; National laboratory capacity building; improved reporting and virus-sharing procedures; enhanced capacity to rapidly detect and respond to zoonotic influenza outbreaks. Recent national surveillance system building efforts in Viet am have included the establishment of hospital-based sentinel creening for severe acute respiratory infection (SARI). The prompt ransfer of clinical samples to laboratories allows for feedback to he participating sentinel sites and simultaneous onward repor- ing of results to national authorities. However, these activities urrently rely upon external funding from the United States Cen- ers for Disease Control and Prevention (CDC) and the number f sentinel sites is decreasing. The viruses and data shared may herefore not be fully representative in terms of geography, cli- ate, age groups and epidemic timing. Selecting representative iruses to send to a WHOCC is further complicated by the two eaks of influenza activity typically observed in tropical countries. s in many other settings, obtaining good-quality denominator ata in hospitals to produce meaningful estimates of the real bur- en of disease caused by influenza is also problematic. Efforts to mprove reporting and virus-sharing procedures have included the33 (2015) 4368–4382 production of a widely circulated weekly newsletter summarizing national influenza activity, weekly reporting to the WHO FluNet platform and the submission of representative circulating viruses to a WHOCC. As a result, the quality of data on the impact and seasona- lity of influenza in Viet Nam has improved, along with the ability to rapidly detect influenza outbreaks and monitor circulating viruses in the context of national prevention and control efforts. In terms of national laboratory capacity building, the challenges experienced in Pakistan illustrate the issues NICs facing in many countries as influenza activities compete for funding with other public health priorities. Foremost among these is ensuring the sus- tainability of funding to meet the running costs of laboratories, and maintain the momentum of recent capacity-building activi- ties. In addition, virus-isolation rates are low and more training in the required skills is needed for laboratory staff. Retaining suitably qualified, trained and motivated staff at all levels of any national system will be a key factor in ensuring the quality, completeness, relevance and timeliness of virus sharing and data reporting. In Madagascar, the national influenza sentinel surveillance net- work monitors and reports upon both SARI and influenza-like illness (ILI) across five different bio-climates. The network is based upon both clinical and biological surveillance activities, with spec- imens submitted weekly to the NIC for analysis. Data-collection and reporting activities include the production of a weekly report on national influenza activity; daily and event-specific sharing of epidemiological data with the Ministry of Health; regular feed- back of laboratory results to sentinel sites; and weekly reporting to the WHO FluNet platform and to the WHO Regional Office for Africa. In common with all NICs, the Madagascar NIC aims to ensure the representativeness of the virus types and subtypes submit- ted to WHOCCs, with all unsubtypable viruses promptly shipped to a WHOCC for further investigation. Future strengthening activ- ities include the provision of support to regional (sub-national) collaborating laboratories, conducting cost-effectiveness analyses to enhance the sustainability of national influenza surveillance activities and research into topics with particular relevance to Madagascar. These include understanding the aetiology of viral ILI (and factors associated with its severity) through integrated genomic, immunological and other approaches; the spatiotempo- ral dynamics of influenza virus circulation; and estimating the risk of human infections caused by swine influenza viruses by identify- ing the genetic and antigenic characteristics of viruses that infect both humans and animals. Experiences gained in China during the strengthening of national influenza surveillance capacities and capabilities – includ- ing for the detection of zoonotic influenza outbreaks – have provided valuable insights that may be highly applicable to other countries and regions. Following its rapid expansion since 2009 (Fig. 1), the Chinese National Influenza Surveillance Network (NISN) comprised 408 network laboratories and 554 sentinel hospital sites by 2013. In that same year, the NISN was able to rapidly detect and characterize A(H7N9) and A(H10N8) viruses causing human infections. In the case of H10N8, this rare infection was detected through the Pneumonia of Unknown Etiology system, highlighting the importance of having the necessary testing and reporting sys- tems in place. For H7N9, confirmed human cases were reported to WHO on 31 March 2013 with candidate vaccine viruses published on the WHO website by 10 May followed on 31 May by formal WHO vaccine virus recommendations. The WHOCC Beijing promotes a national strategy which places laboratory capacity-building at the centre of surveillance- strengthening efforts. Important principles identified in achieving the goals of national surveillance include the collection, reporting and consolidation of data, along with regular data analysis and interpretation. Continuous efforts are required to detect, evaluate and respond to any unusual patterns in the data. The quality of W.K. Ampofo et al. / Vaccine 33 (2015) 4368–4382 4371 za sur l i a c Q v o W v i h r t c i t d v i c v T n T w t t M a s l r c u a a Fig. 1. Expanded and improved influen aboratory testing and related activities is assessed through partic- pation in both national and international initiatives. These include well-established national quality-evaluation system based upon ore timeliness and performance indicators, and the WHO External uality Assessment Project (EQAP) for the detection of influenza iruses by polymerase chain reaction (PCR). Recent initiatives at the regional level have included a study f the patterns of influenza virus submission by countries in the HO European Region. Despite being the cornerstone of GISRS accine virus selection activities, and a crucial element in meet- ng the obligations of the PIP Framework, no systematic analysis ad previously been made of the temporal and epidemiological epresentativeness of the viruses shared with WHOCCs, or of the imeliness of such sharing in relation to the February WHO Vac- ine Composition Meeting (VCM). An analysis was made of the nfluenza surveillance samples submitted by NICs in the region to he WHOCC London over two seasons (2010–11 and 2011–12). The egree of completeness of data provided in conjunction with each irus sample was also evaluated. Aggregated data for both seasons ndicated that a total of 2954 viruses were shared, with 1741 (59%) ollected prior to the February VCM deadline; however, only 946 iruses (32%) were shipped in time for consideration by the VCM. his overall figure also masked clear sub-regional variations, with a umber of highly significant sub-regions being underrepresented. he average periods between specimen collection and shipment ere 90 days and 54 days for the first and second seasons respec- ively, with both the numbers of viruses submitted and delays in heir shipping varying significantly between different sub-regions. issing data precluded further analysis of the level of demographic nd epidemiological representatives achieved over the two sea- ons. Future study objectives include determining the causes of imited, late or lack of isolate sharing in some countries and sub- egions, identifying the sampling strategies and criteria used by ountries to select positive specimens for virus isolation, and eval- ating the impact of increasingly PCR-based surveillance on the vailability and representativeness of viruses reaching NICs. Such analyses of regional submission patterns could potentially id in the revision and refining of current WHO guidance for NICsveillance laboratory capacity in China. on which viruses to share with WHOCCs and by when in order to strengthen the VCM process and its outcomes [3]. At the same time, NICs need to be able to recognize both usual and unusual circulating viruses and to decide upon an optimum submission strategy in the context of their own situation, particularly as the timing of seasons varies. It also remains the case that virus sub- mission patterns may be adversely affected by external factors in many countries, for example, customs requirements and other causes of shipping delays. Although delaying the February VCM deadline by up to three weeks could potentially bring significant gains, any such shift would have to be evaluated for feasibility within the very narrow overall timeframes currently available for vaccine development and manufacture. In the United States, the WHOCC Atlanta has provided guidance to individual States on the number of samples needed for different epidemiological situations. Such an approach could help to improve the representativeness of viruses, and may also be applicable to other regions of the world. At the global level, following adoption of the PIP Framework and other international initiatives, the demands placed on the GISRS have expanded, with activities now including: • comprehensive support for high-quality influenza surveillance and virus detection, sharing and characterization; • maintaining and enhancing electronic reporting platforms such as WHO FluNet (http://www.who.int/influenza/gisrs laboratory/flunet/en/) and FluID (http://www.who.int/influenza/ surveillance monitoring/fluid/en/) to facilitate global reporting; • the development and revision of WHO guidance on epidemiolog- ical and virological surveillance; • supporting national capacity-building and sustainability improvements (including through PIP Framework mechanisms and the WHO Shipping Fund); • collaboration with vaccine manufacturers, associated laborato- ries and regulatory agencies to facilitate vaccine-production and licensing processes; • supporting research into new surveillance and vaccine technolo- gies; 4 ccine • i w t 2 r t l h g o S s t e c i o A t c t i e t c p f u c v 372 W.K. Ampofo et al. / Va strengthening collaboration with veterinary and other animal- sector agencies working at the human–animal interface. Ongoing efforts in these and other areas will be the key to meet- ng the twin demands of continually improving influenza vaccines hile ensuring their availability to an increasingly larger propor- ion of the world’s population. .2. Efforts to increase the availability of egg isolates Influenza viruses isolated in eggs are still needed to meet cur- ent regulatory requirements for vaccine manufacture. There are hree main challenges in obtaining such isolates, namely: low iso- ation rates in eggs; the occurrence of egg-adaptive changes in the aemagglutinin (HA) gene that can lead to changes in virus anti- enic profile; and other changes associated with the development f high-growth reassortants (hgrs) for use in vaccine production. uch difficulties continue to be compounded by the reduced provi- ion of egg isolates by NICs. Potential solutions to the problem of low isolation rates iden- ified by the WHOCC Atlanta include increasing the age at which mbryonated eggs are inoculated from 9–10 days to 13–15 days; hanging the inoculation route to the allantoic cavity; and chang- ng the egg incubation ◦ ◦ temperature from 33 C to 35 C. The use f such approaches has progressively increased the percentage of (H3N2) viruses isolated from 0.8% in 2011 to 11% in 2013, with he trend appearing to continue in 2014 (18% to date). Similar suc- ess was also reported by the WHOCC London following a switch o eggs obtained from a specific breed of hen. Modifying egg- solation parameters has thus resulted in significantly improved gg-isolation rates for A(H3N2) viruses in recent years. In relation o the issue of egg-adaptive changes to the HA gene and further hanges associated with the development of hgrs, the use of egg/cell aired viruses in routine virus characterization can reveal how dif- erences in the substrate used for virus propagation can impact pon antigenic profiles, thus aiding selection of the best A(H3N2) andidate vaccine viruses. Further studies of the egg-adaptation pathways of A(H3N2) iruses may eventually allow for the selection of viruses that are Fig. 2. Egg adaptive changes in the H33 (2015) 4368–4382 antigenically more similar to their mammalian cell propagated counterparts. In the future, such issues may be overcome alto- gether through the use of alternative or emerging technologies such as reverse genetics or the development of synthetic viruses. At present, however, significant adaptive mutations associated with egg-propagated A(H3N2) viruses continue to occur and are being monitored by WHOCCs (Fig. 2). Given the challenges inherent in the isolation and characterization of A(H3N2) egg isolates, consid- eration could be given in the short term to requesting that NICs increase the submission of matching clinical materials along with virus isolates to WHOCCs. 2.3. Efforts from vaccine manufacturers The provision of human-vaccine serum panels by vaccine man- ufacturers remains an essential element in the broad range of well-established cooperative activities between the WHO GISRS and Industry [1,2]. Such panels enable assessments to be made of the reactivity of pre- and post-vaccination sera with influenza viruses of interest (both seasonal and pre-pandemic) and thus generate extremely valuable data for the WHO VCM. At the request of WHO, manufacturers continue to provide human serum panels from different regions of the world – either through contracts or in the case of Europe as part of annual man- ufacturer clinical trials. Following the decision to phase out clinical trials from European Union annual update licensing requirements (http://www.ema.europa.eu/docs/en GB/document library/Scientific guideline/2014/07/WC500170300.pdf), alterna- tive arrangements will be needed in this region and discussions are now being held on the best way forward. Through the International Federation of Pharmaceutical Man- ufacturers & Associations (IFPMA) Influenza Vaccine Supply (IVS) Task Force, Industry also supports research and development into hgrs, which have long been used as the basis of influenza type A components in seasonal influenza vaccines. As the diffi- culties in obtaining A(H3N2) egg isolates in recent years have at least partially been addressed, the number of such isolates has greatly increased thus permitting the improved selection of optimal high-yield strains for vaccine production. In addition to A of A(H3N2) influenza viruses. ccine p t a i v v p s i a a c p r a o r i t c a c c p a f i w g a t ( o C a s p t a t a e g s 2 a i l t w n r o o a r o V l H W.K. Ampofo et al. / Va roviding isolates of antigenic significance to Industry and reassor- ant laboratories, WHOCCs also perform the subsequent antigenic nalysis and sequencing of the resulting strains, thus highlight- ng the interdependence of GISRS and Industry in producing hgr accine viruses that are antigenically closely related to wild-type iruses. Related advances such as the use of monoclonal rather than olyclonal antibodies as selection reagents has the potential both to ignificantly accelerate the speed of production of reassortants and ncrease the number of suitable candidate vaccine viruses. Recent dvances in influenza B reassortant technologies could also soon be pplied to overcome a currently rate-limiting step in influenza vac- ine production, particularly as manufacturers move towards the roduction of quadrivalent vaccines in which type B viruses will equire 50% of vaccine virus production capacity. The IVS Task Force has also collaborated with several WHOCCs nd reassortant producers in evaluating the feasibility of devel- ping virus isolates in qualified cell lines that would meet with egulatory acceptance. Preliminary data indicate that MDCK qual- fied cell lines are suitable for influenza virus isolation, with he majority of viruses produced retaining the properties of the orresponding WHO reference viruses. In conjunction with the cceptable growth rates achieved using MDCK cell isolates in both ell-culture and egg-based manufacturing processes, and the out- ome of risk assessments and supporting literature review, the osition taken by Industry is that the approach appears to be suit- ble. It is envisaged that a regulatory framework could be developed or the use of cell isolates as candidate vaccine viruses, potentially nvolving the provision of approved cell lines to WHOCCs if this ere feasible. The use of “synthetic” technologies may also bring significant ains by allowing for the accelerated production of better-matched nd high-yielding vaccine viruses. This technology was used during he 2013 H7N9 outbreak response, with the HA and neuraminidase NA) genes being synthesized and vaccine viruses rescued within ne week of the China CDC posting the genetic sequencing data. linical trials indicated that the resulting vaccine was both safe nd immunogenic. The routine early and continuous sharing of equencing data could lead to the full realization of the potential ublic health benefits to be gained using synthetic seeds, including he acceleration of pandemic response activities, the rapid avail- bility of higher-yielding and better-matched strains and the ability o generate candidate vaccine viruses in multiple locations as soon s epidemiological and virological data emerge. Other Industry fforts in this area will include collaborative efforts to identify enetic markers of yield that can be used to select for high-yielding trains. .4. Collaborative animal influenza surveillance and response ctivities The importance of sustained and coordinated inter-sectoral nfluenza surveillance at the animal–human interface, and of col- aborative assessment of the risk of a human pandemic, continues o be reflected in the FAO-OIE-WHO Tripartite Concept (http:// ww.who.int/influenza/resources/documents/tripartite concept ote hanoi 042011 en.pdf?ua=1). Initiated in January 2011 and enewed in January 2014 for a further 5 years, this concept sets ut the roles of both WHO and the OIE-FAO Network of Expertise n Animal Influenza (OFFLU) in coordinating global activities to ddress health risks at the animal–human–ecosystems interface. Recent progress has included improved sharing of viruses and eagents for their characterization, and the increased availability f appropriate OFFLU information and data at the biannual WHO CMs. Key information now routinely shared includes epidemio- ogical overviews and phylogenetic data of highly pathogenic avian 5N1 in animals; the results of antigenic testing of specified isolates33 (2015) 4368–4382 4373 using ferret-derived antisera; and information on a broad range of animal viruses considered to be of public health concern, including H9, H7 and H5 (other than H5N1) subtypes. Long-standing obstacles to sustainable influenza surveillance in animals include limited awareness of the need to manage pandemic risks; the lack of drivers for non-notifiable influenza surveillance; and associated absence of legislative frameworks. In many settings, under-resourced veterinary services face challenges in securing sustainable funding and obtaining industry involve- ment, particularly when balancing financial and other incentives against potential disincentives. As a result, advocacy efforts are still needed to further strengthen the commitment of national vet- erinary agencies to global public health objectives, and to make the case for the extra resources and increased awareness that are needed to support the strengthening and sustaining of veterinary services. Improving influenza surveillance in farmed animal populations, and clearly establishing the roles and responsibilities of all agen- cies working in this area, will become increasingly necessary, especially as surveillance capacities and technologies evolve. For example, in 2004 the very limited capacity for H5 surveillance led to the establishment of GISRS H5 Reference Laboratories and the resulting close collaboration with OFFLU in a broad range of surveillance, pandemic risk assessment and vaccine development activities. Following a subsequent dramatic increase in H5 surveil- lance, including by WHOCCs, consideration is being given to the current role of H5 reference laboratories. 3. Improving the understanding of influenza activity and addressing the complexities related to vaccines in tropical regions 3.1. The concept of influenza transmission zones WHO is working to identify epidemiological “transmission zones” in which countries with similar influenza activity and trans- mission trends are grouped together in order to better reveal global and regional patterns of influenza spread. Eighteen transmission zones were provisionally created, initially based upon existing geo- graphical regions and adjusted according to knowledge of influenza transmission trends (Fig. 3). Following analysis of FluNet data sub- mitted by countries that met specified criteria, it was concluded that the provisionally identified zones did appear to consist of countries exhibiting similarities in their patterns of influenza trans- mission. Further time-series analyses were carried out to better categorize tropical transmission patterns, again using countries with continuous FluNet data for a specified period. In line with recent published findings [4], statistical assessment of the num- ber and timing of national influenza activity peaks indicated that although influenza patterns definitely exist in tropical regions of the world, such patterns were less evident than those observed in the northern and southern hemispheres. Quality surveillance data are now being generated to allow for more-detailed trend analyses and for periodic review of transmission zone divisions. Improving the methodology used and supplementing data with literature reviews could lead to a greatly improved understanding of global influenza trends; strengthened seasonal influenza pre- paredness; and a more accurate assessment of burden of disease to guide the development and expansion of national vaccination policies. The continuous collection of both laboratory and epidemi- ological influenza surveillance data will be the key to achieving such goals. Current limitations include the reliance on national level data which may not be regionally representative, due for example to the widely differing sizes of different countries. In some cases, data may not even be nationally representative due to wide variations in 4374 W.K. Ampofo et al. / Vaccine 33 (2015) 4368–4382 a trans s a f a b s e h 3 t l c s i n i l c s c t s t r c t g v d p m t b f m o s Fig. 3. Influenz easonal patterns across very large countries such as Brazil, China nd India. There is thus a need to better understand the underlying actors that result in the creation of both broad transmission zones nd of regional outliers. As understanding increases, it may even ecome feasible to reduce the number of transmission zones, while till bringing about significant practical refinements and other ben- fits not available using the current three “de facto” zones (southern emisphere, northern hemisphere and the tropics). .2. Addressing the complexities related to influenza vaccines in ropical regions WHO issues its recommendations for influenza vaccine formu- ations in February and September each year in preparation for the orresponding northern and southern hemisphere influenza sea- ons (November–April and May–October respectively). However, n tropical regions of the world, such well-defined seasonality does ot always occur. For example, in Kenya and many other countries n tropical Africa, influenza viruses circulate year-round and despite imited epidemiological data, there is growing evidence of a signifi- ant burden of influenza disease in the region. Following improved urveillance and virus-sharing efforts in Kenya and other African ountries, it appears that selected priority populations would stand o benefit from receiving either the northern or southern hemi- phere vaccine formulation. Further work is needed to evaluate he optimal delivery mechanisms in a country with almost year- ound virus circulation, and to determine the added benefits and osts of strategically using both vaccine formulations as opposed o a single formulation. As influenza vaccine use in tropical Africa rows, more data will be needed in countries to inform national accination policies. In the American tropics, similar issues exist in determining the egree of influenza seasonality, predicting annual transmission atterns and selecting between southern or northern vaccine for- ulations. An analysis of data from countries in the region found hat for many countries influenza epidemics typically occurred etween May and September during the austral winter and lasted or 4–5 months, rather than year round. Although the specific for- ulation used in individual countries varies, an estimated 81% f the predominant strains in the American tropics were repre- ented in the southern rather than northern hemisphere vaccinemission zones. formulation, with the result that this was the most up-to-date com- position. The viruses that circulate in the American tropics tend to be similar throughout the region but those characterized as predominant one year in a particular sub-region tend to become dominant in other sub-regions in successive years in a clearly dis- cernible geographical order. In Americas, influenza activity in the tropics is often preceded by the release of the southern hemisphere vaccine and by Vaccination Week in April each year. Against this regional backdrop, national patterns of influenza transmission and seasonality are complex. Nevertheless, as quality surveillance data become available for more years, the optimal timing of vaccination is likely to become clearer. In Brazil, a large latitudinal range encompassing both temperate and tropical regions, with good epidemiological data available in certain of its sub-regions, provides an example of how the optimum timing of vaccination cannot be reduced to the current northern–southern hemisphere paradigm. One study comparing the historic use of southern hemisphere vaccine recommendations and schedule against a hypothetical northern hemisphere vaccine scenario concluded that a higher degree of matching between cir- culating and vaccine viruses would have been achieved had the northern hemisphere vaccine composition and vaccination sched- ules been followed [5]. Although based on relatively few virus isolates, this study highlights the complexities of influenza vacci- nation in a large tropical country. Although most tropical countries in the Asian region technically lie in the northern hemisphere, they do not exhibit the influenza seasonality seen in temperate regions, with some level of influenza virus circulation typically occurring throughout the year. Asian countries lying between the equator and approximately 30◦ N in latitude experience a peak in influenza during the monsoon sea- son (June–September). In countries closest to the equator, there is year-round circulation with no discrete peak season. In order to better understand patterns of influenza virus transmission in the region and inform national vaccination approaches, a study was conducted which aimed to characterize influenza seasonality in tropical and subtropical countries of southern and south-eastern Asia; identify latitude gradients associated with discrete seasona- lity; and determine the best time of the year for national influenza vaccination campaigns [6]. Weekly surveillance data from 10 Asian countries over the period 2006–2011 clearly indicated peak ccine 33 (2015) 4368–4382 4375 p d Box 1: CONSISE—a global standardization and w information-sharing platform of influenza seroepi- s demiology c Early in the course of the 2009 H1N1 pandemic it was realized that there was a need for timely and standardized seroepide- p miological data to better estimate disease severity and attack r rates during non-seasonal events in order to inform policy decisions. Following its establishment in 2011, the Consor- c tium for the Standardization of Influenza Seroepidemiology r (CONSISE) has worked to standardize the seroepidemiology s of influenza and other respiratory pathogens, and to develop d comprehensive investigation protocols for use in responding a to both seasonal and potentially pandemic influenza viruses, c and other respiratory pathogens. CONSISE now has more than E 100 members in over 40 countries who openly and freely share study and laboratory assay protocols and other information on t the internet (https://consise.tghn.org/). t A number of CONSISE evaluation and standardization studies a were conducted to assess and improve the standardization of a antibody assays worldwide. Despite the publication of WHO u protocols for both HAI and MN assays [9] significant variations t were found between laboratories in terms of assay protocols, o and in the determination and expression of endpoint titres. New consensus protocols for the HAI assay, 2-day enzyme- linked immunosorbent assay (ELISA) MN assay and 3-day HA 4 MN assay were therefore developed and published based upon i “required” or “recommended” parameters. During the 2013 H7N9 event, CONSISE in its capacity as a unique international forum for seroepidemiology laboratories r organized a teleconference and promptly published web-based i u i z c a f n p d d b t w n G a a a c a p S u l w t o s a l r i u s ( W.K. Ampofo et al. / Va eriods of influenza activity in seven of the countries, with no istinct seasonality in the remaining three. In many cases, it as apparent that the current vaccination schedules used were uboptimal and that most tropical countries in Asia might benefi- ially consider conducting vaccination in April–June each year, i.e., rior to influenza peak circulation, using the most recent WHO- ecommended vaccine formulation. Compiling and sharing data on influenza from individual ountries in a given region appears to potentially allow for a egional consensus to be reached on the circulation patterns and easonality of influenza, which could also incorporate latitudinal ifferences. At present, the effects of climatic factors in particular re poorly understood especially in remote areas of large tropical ountries and in countries where the required data remain sparse. ven in those tropical countries where there are likely to be at least wo peaks of influenza activity each year, improved knowledge of he main peak has the potential to result in improved vaccination pproaches. Across all the tropical regions of the world there now ppears to be a paradigm shift occurring based upon improved nderstanding of the extent to which influenza is seasonal in order o strengthen national vaccination policies and select vaccines of ptimal composition. . Improving the characterization and selection of nfluenza vaccine viruses Maintaining good levels of influenza vaccine effectiveness [7] equires regular vaccine composition updating and annual admin- stration. Until radically different approaches such as the use of niversal influenza vaccines become feasible, the updating process s likely to remain based primarily upon the antigenic characteri- ation and selection of egg-isolated (and potentially cell-isolated) andidate vaccine viruses for the production of both inactivated nd live attenuated vaccines. WHOCCs combine the data obtained rom the antigenic characterization of viruses using HAI and virus eutralization assays and the serological reactivity of pre- and ost-vaccination human sera with extensive genetic sequencing ata and epidemiological and clinical information. The resulting atasets form the scientific basis for expert consideration at the iannual WHO VCMs. Despite being the traditional assay of choice since the 1940s, here remain a number of limitations and difficulties associated ith the use of HAI assays. As a result, a range of corrective tech- iques and complementary assays are used by WHOCCs and other ISRS laboratories [2]. Efforts are also currently under way to ddress issues such as the differential reactivities of egg-derived nd cell-derived viruses with ferret sera, and the complications rising from the binding of the virus NA surface protein to red blood ells. As a result, new approaches continue to be needed to improve ssay sensitivity and accuracy, streamline the throughput of sam- les and improve the reproducibility of data between laboratories. uch approaches would also need to be sufficiently flexible to be sed in the analysis of antibody responses to emerging viruses. Until significant advances are made in the development of new aboratory assay platforms and/or vaccine technologies WHOCCs ill continue to face acute time pressures, particularly around the ime of the biannual VCMs. Any requirement for the introduction f further assays, or for the greatly expanded use of existing but electively applied approaches such as microneutralization (MN) ssays, are likely to prove problematic, particularly given the rate- imiting step of growing viruses and developing the reassortants equired for some assays. Realistic aims at present would appear to nclude the more-optimal use of HAI and other assays; improved nderstanding of the data produced by neutralization assays; and trengthening of national and global initiatives such as CONSISE Box 1) for the standardization of serological testing methods.protocols for the detection of antibodies to the emerging virus. Attempts to overcome the inherent limitations of the traditional HAI assay through the development of assays based upon synthetic beads or solid matrices have had only limited success. Studies into the development of synthetic red blood cells consisting of beads coated in either purified natural glycans or synthetic sialyl-glycans have highlighted that, despite evidence of high reproducibility, a range of complex issues (including the prohibitive cost of synthet- ically produced glycans) would need to be addressed before the approach could become feasible. Efforts to harness a number of non-bead technologies based upon the use of sialyl-glycans or red blood cell membrane fragments is ongoing. Despite the adoption of multiple strategies for developing HAI replacement assays, no viable alternative has yet emerged. The development of such assays would be enhanced by improved understanding of the glycans rec- ognized by different influenza types/subtypes. As previously reported [2], the NA surface protein plays a key role in the life cycle of the influenza virus and its transmission. Despite ongoing high levels of interest in the potential role that antibodies to NA could play in the development of more-effective vaccines, there is presently no regulatory requirement for the pre- cise determination and standardization of vaccine NA content. Nevertheless, different NA subtypes are known to be genetically distinct, exhibit discontinuous antigenic drift and give rise to anti- bodies associated with protection against homologous, and to a lesser extent heterologous, influenza viruses. Efforts are therefore continuing to improve understanding of the patterns of antigenic drift in the NA of seasonal influenza viruses. The application of “antigenic landscape” modelling approaches may also provide enhanced understanding of the quality and breadth of human antibody responses elicited against HA (and potentially against NA) following infection or vaccination, and of the influence of prior immunity on vaccination responses. If corre- sponding advances in predicting the course of viral evolution prove to be feasible then such approaches could be used to inform and 4 ccine e t a h [ e V e v p c a n b p t d t e o a s i r a 5 v p r t h N a o c a t b ( g m r e h c a n s p a m g w o i D S j D t t c 376 W.K. Ampofo et al. / Va valuate vaccination strategies. Antigenic mapping (cartography) echniques have been studied for about ten years and despite vari- ble red blood cell binding effects, and other practical issues, can elp to visualize the antigenic evolution of predominant viruses 8]. Antigenic landscaping approaches have now been used to xamine historic HAI datasets from a household cohort study in iet Nam in order to recreate the recent course of influenza virus volution and explore pre- and post-infection responses. Obser- ations of increasing immunity over time clearly highlighted a ersistent “back-boost” effect in which vaccination with antigeni- ally advanced vaccines appeared to stimulate a recall of previous ntibody responses. In terms of vaccination strategies such a phe- omenon may imply that the optimum vaccine strain to use might e ahead of the centre of the current cluster of evolving strains, thus roviding both recall benefits as well optimal de novo responses o new epitopes. By combining such insights with improved pre- ictions of the course of future virus evolution it may be possible o improve the vaccine virus selection process and increase the ffectiveness of influenza vaccines. At present, there are a number f practical and theoretical assumptions underlying the approach nd further studies are required. These could include prospective tudies of vaccination with an antigenically advanced virus; stud- es across different age groups; and studies based upon data from a ange of alternative laboratory assays (such as NA assays, MN and ssays that detect stalk-reactive antibodies). . New technologies and tools for improving influenza accine virus selection A number of emerging new technologies and tools have the otential to revolutionize influenza virological surveillance and esponse activities. In the context of national surveillance activities, he increasingly routine use of whole genome sequencing and other igh-throughput approaches in large and technologically capable ICs is providing significant insights in areas such as virulence ssessment, phylogenetic and transmission studies and evaluation f antiviral susceptibility. As such high-throughput methodologies hange the focus of surveillance away from single HA and NA char- cterization and towards whole genome sequence determinations hey will become not only a key element in outbreak investigations ut will also drive a shift towards a new surveillance paradigm Fig. 4). Further development and increased availability of next- eneration sequencing and associated technologies and equip- ent, together with reduced operating costs and data-analysis equirements, could also allow for the examination of global gene xpression at the level of pathogen and host. Such approaches ave the potential not only to greatly improve understanding of irculating viruses and their evolution, but to reveal the nature nd extent of intrahost viral diversity and degree of fitness; sig- al the potential emergence of drug resistance; aid in vaccine virus election and allow for the accurate assessment of risk, including andemic risk. Datasets and pipelines could also be rapidly gener- ted for examining genetic variation in populations of viruses that ay be associated with a particular phenotype. The use of synthetic enomics technologies based on the sequencing data obtained ould then allow for the rapid synthesis for future use of “libraries” f numerous HA, NA and other gene segments. Efforts under way n this area include the National Institute of Allergy and Infectious iseases (NIAID), National Institutes of Health (NIH)-supported ynfluenza project at the J. Craig Venter Institute (http://gsc. cvi.org/projects/gsc/Synfluenza/index.php) and the United States epartment of Health and Human Services (HHS)-supported syn- hesis of whole viruses for vaccine production (as occurred during he production of inactivated H7N9 vaccines in 2013) and for asso- iated challenge, transmission and pathogenesis studies.33 (2015) 4368–4382 As evidenced in 2013, assessing and responding to the risk to human health posed by zoonotic influenza infections is currently highly problematic for national surveillance systems. Substantial viral diversity exists within animal populations, even within the same subtype, and only limited genetic sequencing and antigenic data are typically available. Responding to outbreaks of zoonotic influenza viruses is also compounded by pronounced variability in their growth properties. Efforts undertaken to address this sit- uation include virus-challenge studies conducted at the WHOCC Memphis, which have highlighted a number of potentially attain- able benefits of genomic and bioinformatics approaches to avian influenza surveillance, while providing insights into the method- ological and other challenges that will need to be overcome. In addition, a case study conducted by the United States CDC, and based upon clinical samples taken during the H7N9 event in China, indicated that despite multiple challenges, properly applied next-generation sequencing technologies to detect and assess the properties of emerging intrahost genetic variants can lead to improved risk assessment. Studies have also been conducted on the potential of use of reconstructed ancestral A(H5N1) influenza viruses to develop cross-clade protective vaccines [10], and on the utility of reverse genetics techniques to improve H5N1 vaccine virus growth rates [11]. Long-term goals in efforts to detect and respond to zoonotic influenza outbreaks include the routine use of surveillance and risk-assignment activities based upon the use of genetic sequencing data, followed by the use of reverse genetics approaches to rapidly generate suitable candidate vaccine viruses. Despite the advances made to date, significant challenges will need to be addressed before next-generation technologies become a routine part of national surveillance and response paradigms, par- ticularly in low-resource settings. These include the need to fully understand the utility of more-comprehensive genomic datasets in vaccine virus selection and development. In addition, the use of current technology platforms requires considerable expertise, with advanced research and development efforts now under way to refine and harmonize the pipelines and bioinformatics tools required at different stages of the process. Bioinformatics demands in particular remain high given the ever-evolving nature of instru- mentation and protocols, and the costs associated with the analysis of sequencing data. Interest also remains high in the development and potential applicability of mathematical modelling approaches for predict- ing the course of influenza virus evolution, and thus potentially accelerating the selection of new vaccine viruses. Recent work on integrating data on the antigenic and genetic evolution of influenza viruses indicates that such a combined approach can provide potentially valuable new insights into the factors that determine observed patterns of antigenic drift [8]. By pinpointing the precise mechanisms by which changes in viral genes result in antigenic change it may soon be feasible to predict which of the viruses circu- lating at the start of the influenza season will come to predominate in terms of number of new infections. The related concept of viral “fitness” provides a further potential means of evaluating and pre- dicting patterns of virus evolution. This approach is based upon the integration of publicly available HAI, sequence and biophysical data plus regional information to develop a joint epitope/non- epitope fitness model. Preliminary retrospective studies indicate that such an approach can successfully predict the evolution of HA sequence clades year on year. The predictive potential of such a model at the phenotypic level remains to be demonstrated. Poten- tial further improvements and refinements include more-rigorous data-quality control, use of a wider range of input-data categories and improved integration of selected datasets. Approaches based on systems genetics and systems biology con- cepts also have the potential to provide valuable new insights. Systems genetics approaches can capture human genetic diversity, W.K. Ampofo et al. / Vaccine 33 (2015) 4368–4382 4377 he new surveillance pyramid. a s i g d u g c b f i a t o a a h a u a g g g n e v U i u i v r r m t u ety (“metadata”). A collaborative United States Genome Sequence u Centers-BRC metadata working group has now been established in t an attempt to support the capture of relevant, standardized and i b i d Fig. 4. Paradigm shift—t nd allow, for example, the identification of the specific host- usceptibility genes that regulate disease outcomes following viral nfections. Systems biology approaches provide an opportunity to enerate new types of datasets with the potential to identify both iagnostic and prognostic markers; understand pathogenic and vir- lence mechanisms; evaluate vaccine performance and responses; enerate improved cell lines for virus cultivation; and identify orrelates of protection. Despite being at an early stage, systems iology approaches allied to current statistical capabilities could easibly be used to elucidate some of the molecular correlates of mmune responsiveness and related immunogenicity issues. Such ssociative studies may soon be feasible in the context of clinical rials to determine their potential application. However, in terms f understanding causality much remains hypothetical at this stage nd further work is required to identify the key pathways of interest nd develop effective therapeutic approaches. As viruses rely upon ost factors to replicate, and often hijack the cellular processes initi- ted in response to infection, such approaches could even be based pon the suppression of host responses. Human influenza vaccine responses and course of infection have lso been evaluated during a longitudinal study which combined enetic, transcriptional and immunological data [12]. Such an inte- rative genomics approach allows for the identification of host enetic factors that contribute to variations in vaccine responsive- ess, and may uncover important mechanisms affecting vaccine fficacy. The results indicate that genetic variations between indi- iduals are important determinants of vaccine immunogenicity. nderstanding the complex mechanisms that underlie variations n vaccine responses may allow for the identification of individ- als who do not develop a protective antibody response following nfluenza vaccination. Appropriate modifications to the dose or accine type given to such individuals could potentially lead to a eduction in the proportion of the population who would otherwise emain unprotected. The approach may also be useful in guiding the odification of factors, such as adjuvant use, intended to enhance he immune response to influenza vaccines in all recipients. Taken together, the range of new technologies and tools now nder development are capable of generating potentially huge vol- mes of highly complex and diverse datasets. Recent trends in he acceleration of data volumes, velocity and variety are driv- ng a newly emerging concept of “big data” (Fig. 5). Efforts will e required to ensure the availability, accessibility and qual- ty of the data generated, including through the validation of atasets; to enable the meaningful integration of often highlyFig. 5. The three “V”s of big data.2 disparate datasets; and to ensure the broadest possible relevance and application of the resulting findings. As part of its activities in these areas, the NIAID/DMID Genomics Program has invested in a number of Bioinformatics Research Centers (BRCs), including the Influenza Research Database (www.fludb.org). This free-to-use comprehensive collection of influenza-related data and analysis tools is intended to support a process of data standardization and integration encompassing a range of sequence, surveillance and immunological data categories. In addition to expanded data volumes and the accelerated velocity of data generation, current efforts to map the organizing principles of big data and its application increasingly highlight the need for parallel advances in the underlying knowledge base and interpretive context in which to set the data generated. Discerning the key patterns in data relies crucially on the capture of data vari-2 Source: Journal of PayDeg. Issue 03. August 2012 (http://paydeg.com/ CloudComputing3.html, accessed 20 April 2015). 4 ccine c n d p a d 6 v i i i d s I I i i c o i o t a S N o r f H p a A a o w i n m i v c u m ( v a t r o v r V e o t n r n e a t s 378 W.K. Ampofo et al. / Va onsistently reproducible metadata by its member projects. Har- essing the currently theoretical benefits of concepts such as big ata for the influenza vaccine virus selection and development rocess will depend upon consolidating and building upon these nd further advances in the generation, integration, analysis and issemination of potentially huge and complex datasets. . Manufacturing and regulatory aspects of improved accine virus selection and development Over the last decade, growing awareness of influenza as an mportant global public health issue has been coupled to increas- ng demands for more-equitable access to effective and affordable nfluenza vaccines. However, efforts to increase global vaccine pro- uction by establishing manufacturing and regulatory capacity in ome developing countries have met with a number of challenges. n Brazil, for example, influenza vaccine production at the Butantan nstitute was initiated through a technology-transfer programme nvolving Sanofi Pasteur that started in 1999 and was completed n 2010 as part of the WHO Global Action Plan for Influenza Vac- ines (GAP). The transferred technology enabled the manufacture f egg-based split virus vaccine with an annual production capac- ty of 20 million doses, based primarily upon the estimated size f the elderly population targeted in national policy for vaccina- ion in 1999. This unprecedented initiative proved to be a long nd involved process for all stakeholders, including government, anofi Pasteur and the federal Brazilian regulatory agency Agência acional de Vigilância Sanitária (ANVISA). However, despite the bstacles encountered, the technology-transfer process not only esulted in seasonal vaccine production in Brazil, but also allowed or the development and production of H5N1 vaccines, pandemic 1N1 vaccine lots for clinical studies and H7N9 vaccine lots for re-clinical studies. Current challenges include the need to acceler- te vaccine production and to conduct clinical trials as required by NVISA, and overcoming the distribution logistics associated with vast land area. It is intended that improvements in the acquisition f eggs, reassortants and reagents; strengthened communications ith federal customs and ANVISA; improved collaboration with nfluenza reference laboratories; and enhanced production plan- ing will result in an increased production capacity sufficient to eet the needs of a range of newly identified groups targeted for nclusion in expanded national influenza vaccination policies. Successfully completing each step in the annual influenza accine manufacturing cycle relies upon timely and regular ommunication between the WHO GISRS, manufacturers and reg- latory authorities. For northern hemisphere vaccines, production ust begin almost a year in advance of their eventual deployment Fig. 6). The seasonally shifted schedule for southern hemisphere accines involves a shorter lead time between the WHO VCM nnouncement in September and final formulation and distribu- ion. Despite early access by manufacturers to epidemiological data, egular teleconferences prior to the VCM and prompt information n available reassortants and reagents, at least one component irus of seasonal influenza vaccines must be manufactured “at isk” and up to two working seeds prepared prior to the WHO CM (Fig. 6). Subsequent steps in the manufacturing cycle are qually time critical and include the optimization and validation f the manufacturing process; the supply of calibrated reagents for he single radial immunodiffusion (SRID) assay by ERLs; and the eed to annually update the product licences in an often complex egulatory and manufacturing environment. These timelines have ow come under further pressure with the introduction of sev- ral quadrivalent vaccines. In view of such time constraints, there ppears to be no obvious means of accommodating any proposal o improve the representativeness of viruses sent to WHOCCs by trategically delaying the timing of the WHO VCM.33 (2015) 4368–4382 Eliminating avoidable delays in the vaccine production cycle caused by the use of suboptimal PR8 master donor viruses remains a key aim, including during the time-critical development of can- didate pandemic vaccine viruses. The United States Biomedical Advanced Research and Development Authority (BARDA) is work- ing to develop panels of optimized viral “backbones” that can be selectively used in the accelerated production of inactivated influenza vaccines. Three representative PR8 donor viruses from GISRS partner laboratories were compared in terms of their impact on HA yield across a broad range of different influenza virus sub- types and lineages. In some subtypes (for example, the H5N1 strain A/Hubei/1/2010) optimizing the PR8 internal genes was associated with an up to eight-fold increase in HA yield compared to subopti- mal donor use. Although the switch to a 5:3 genotype (containing wild type PB1 as well as HA and NA genes) generally had detrimen- tal effects on the yields achieved with the classical 6:2 genotypes, this did not apply to certain low-yielding 6:2 viruses in which yield could be doubled. In terms of vaccine potency, SRID tests remain the established gold standard for determining the antigen content of influenza vaccines, with studies indicating acceptably low inter-laboratory variability. However, the 2009 H1N1 pandemic placed increased pressures on vaccine production and release timelines, with clin- ical trials being either required or desirable. The availability and early use of other types of vaccine potency assays prior to the production of SRID reagents early in the pandemic resulted in renewed interest in the development of alternative approaches. The leading candidates are either physicochemical assays – such as high-performance liquid chromatography (HPLC); sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE); and mass spectrometry – or biological assays (“bioassays”) such as ELISA/enzyme immunoassay (EIA); surface plasma resonance (SPR); and immunocapture isotope dilution mass spectrometry (IC- IDMS). Physicochemical assays potentially have the advantages of being immediately available, rapid, reproducible and well suited for use in automated high-throughput approaches. Disadvantages include the inability to take into account the conformation, anti- genicity or immunogenicity of the measured HA; the potential need for reference reagents; and the expense and technical difficulty of some methods. Bioassays measure vaccine biological activity or reactivity and are usually specific for the antigen being assayed. Their main advantage is their potential for measuring biologi- cally active HA and for evaluating its stability. Moreover, some assay formats, such as ELISA, are already well established and easily implementable. Disadvantages include the need for specific reagents and the expense and technical complexity of some assays. Despite this, bioassays are preferred over physicochemical assays with the combination IC-IDMS assay also offering some potential. Workshops have now been held and studies initiated to compare and evaluate a range of physicochemical and biological assays as potential alternatives to SRID. In the United States, regulatory experience at the Food and Drug Administration (FDA) Center for Biologics Evaluation and Research suggests that the current influenza vaccine virus selection process generally works well, and does not in itself present significant reg- ulatory issues for vaccine manufacturing. However, the required analysis of manufacturer’s seed viruses and the preparation and calibration of potency reagents are highly resource intensive. This situation is now being exacerbated by an increase in the num- ber of available reference viruses and the licensure of new vaccine manufacturers. In addition, recent licensure processes for quadri- valent, cell-based, recombinant-protein and adjuvanted vaccines have highlighted a number of regulatory challenges associated with new influenza vaccine types. For quadrivalent vaccines, licensure is based upon safety and immunogenicity studies and the demonstration of non-inferiority W.K. Ampofo et al. / Vaccine 33 (2015) 4368–4382 4379 ines fo t f f c a e s i o S p g i e c ( t a w c d o o o a d r t Fig. 6. The manufacturing cycle for influenza vacc o licenced trivalent vaccines. Related key issues include the need or four sets of specific reagents for potency testing; the potential or both wild-type and reassortant viruses to be used as vac- ine viruses; and cross-reactivity and other difficulties in SRID nd identity testing arising from the relatedness of the two B lin- ages. Influenza B strains are also typically the slowest growing trains and the lack of ideal high-growth reassortants may result n manufacturing and lot-release delays. Currently, there is only ne mammalian cell based influenza vaccine licenced in the United tates, with studies indicating that initial concerns around the otential tumorigenicity of MDCK cell lines were unfounded. A sin- le recombinant protein influenza vaccine has also been licenced n the United States and is based upon an insect virus (baculovirus) xpression system and recombinant DNA technology. Although andidate vaccine viruses are not needed to produce such a vaccine only sequence information for the recommended HA is needed) here are likely to be issues in relation to the determination of ppropriate potency assays and current lack of established path- ays for product improvement. Any development and licensure pathway for adjuvanted vac- ines will require careful attention to preclinical testing, study esign, dosing decisions and safety monitoring. The complex nature f adjuvants and adjuvanted vaccines necessitate the collation f extensive safety data, including clear indications of duration f follow-up, monitoring of adverse events of special interest nd a focus on the potential for autoimmune/auto-inflammatory isease development. The only adjuvanted influenza vaccine cur- ently licenced in the United States is an AS03-adjuvanted H5N1 3 Source: International Federation of Pharmaceutical Manufacturers & Associa- ions (IFPMA) Influenza Vaccine Supply (IVS) Task Force.r use in the northern hemisphere winter season.3 monovalent vaccine produced under government contract as part of national pandemic preparedness. A detected association between narcolepsy and AS03 following the use of 2009 H1N1 pandemic vaccines in a small number of Scandinavian countries highlights the case-by-case approach required while emphasizing the need for appropriate safety data for all adjuvants. The emergence of potentially pandemic influenza viruses cre- ates specific regulatory and associated challenges in terms of clinical study design, timeline and interim data analysis. Over- coming such challenges requires collaboration and cooperation between multiple agencies and manufacturers. Since 2009, NIAID clinical vaccine trials have been conducted to determine the appro- priate response to emerging influenza viruses in both pandemic and non-pandemic scenarios. The ongoing emergence of multi- ple potentially pandemic influenza viruses (such as A(H3N2)v, A(H5N1) and A(H7N9) viruses) will continue to create challenges for vaccine development in terms of clinical trial study design and resource requirements. During the 2009 H1N1 pandemic, 12 clini- cal trials were undertaken by the NIAID in the context of the United States national emergency preparedness framework. These studies were complementary to planned Industry trials, and were primar- ily intended to generate data on safety and vaccine use in selected populations. Despite the challenges and constraints of current vaccine- production technologies, the influenza vaccine landscape contin- ues to be dominated by egg-based inactivated and live attenuated vaccines, with a small number of cell-based and recombinant vac- cines. During a pandemic all these platforms would potentially necessitate antigen-sparing strategies and possibly gains in manu- facturing efficiencies to meet demand. While the field of influenza vaccines thus appears to be moving towards a variety of niche products in the near term, it is apparent that the ultimate goal 4 ccine r i v a p t 7 o c n e m i s s t r t T m i g r s a g t c t o a r h A w a p c s a r G i p c O s n s t m t s l a c p r w 380 W.K. Ampofo et al. / Va emains the development of “universal” influenza vaccines offer- ng longer-lasting immunity against a broad range of influenza A irus subtypes. Such vaccines would likely remove the need for nnual vaccine virus selection, reduce production costs, eliminate otential vaccine mismatches and shortages, and greatly increase he global supply of both pandemic and seasonal vaccines. . Conclusions and future directions Virological and epidemiological surveillance is the foundation f the influenza vaccine virus selection and development pro- ess, and national, regional and global efforts will continue to be eeded to improve monitoring and reporting activities, and to nsure the timeliness and representativeness of virus sharing. In any countries, much remains to be done in establishing, expand- ng and sustaining ILI and SARI sentinel surveillance, motivating taff, improving laboratory capacity and providing training for ub-national and regional laboratory personnel. In some parts of he world, severe economic and/or logistical hurdles continue to estrict the timeliness and temporal and geographical represen- ativeness of the viruses sent to WHOCCs for detailed analyses. o help countries make the best use of available resources while eeting their obligations under the PIP Framework, WHO and ts partners will continue to work to develop regional and global uidance on optimal approaches for ensuring the timeliness and epresentativeness of influenza isolates and clinical specimens hared with the WHO GISRS. Improved knowledge of influenza transmission “zones” could llow for refined and more-targeted surveillance activities, and reater understanding of virus circulation and transmission pat- erns, including in tropical regions. Further refinement of the urrently provisional basis used to determine functional influenza ransmission zones should be considered. Even without the devel- pment of specific recommendations for tropical areas, such an pproach could bring about much-needed improvements in cur- ent guidance on where and when to use northern or southern emisphere influenza vaccine formulations in tropical countries. s part of this, consideration could also be given to promoting the ider use and potential refinement of reporting platforms such s WHO FluNet and FluID, for example to accommodate multi- le seasonality and/or wide geographical variety within a single ountry. The need for sustained and coordinated inter-sectoral influenza urveillance at the animal–human interface, and collaborative ssessment of the risk of a human pandemic, continues to be eflected in the FAO-OIE-WHO Tripartite Concept (“One Health”). iven the long-standing obstacles in implementing sustainable nfluenza surveillance in animals, and in building upon recent rogress in integrating the outcome of this into the public health ontext, improved collaboration between the WHO GISRS and FFLU is needed. Opportunities to improve the antigenic characterization and election of influenza vaccine viruses include the strategic use of eutralization assays to support HAI assay data, and the further trengthening of standardization and information-sharing initia- ives. Efforts to develop alternative simplified and high-throughput ethods and quantitative assays will also continue, along with fur- her research into the utility of approaches based upon the use of ynthetic red blood cells, characterization of virus NA and antigenic andscape technologies. Realistically, such aims must be weighed gainst the finite capacity of WHOCCs to antigenically characterize andidate viruses in typically limited timeframes.A number of emerging new technologies and tools have the otential to revolutionize influenza virological surveillance and esponse activities. These include the increasingly routine use of hole genome sequencing and other high-throughput approaches,33 (2015) 4368–4382 and the potential application of synthetic genomics, systems genet- ics, systems biology, mathematical modelling and bioinformatics approaches. However, despite the advances made to date, signifi- cant challenges will need to be addressed before such technologies become a routine part of national surveillance and response paradigms, particularly in low-resource settings. Although vaccine technology is evolving, and despite long- recognized challenges and constraints associated with current vaccine-production technologies, the influenza vaccine landscape continues to be dominated by egg-based inactivated and live atten- uated vaccines, with a small number of cell-based and recombinant vaccines. The research and development of hgrs will for the foresee- able future continue to be a major focus of Industry and government agency efforts as such viruses form the basis of the influenza type A components of current seasonal influenza vaccines. While the pipeline of influenza vaccines appears to be moving towards a variety of niche products in the near term, it is apparent that the ultimate aim remains the development of effective “univer- sal” influenza vaccines that offer longer-lasting immunity against a broad range of influenza A subtypes. In this and other related areas of research and knowledge development, WHO and its part- ners will work to ensure that the advances made are translated into improved influenza preparedness and response activities and into far more equitable global public health outcomes. Appendix A. Declarations of interest The 3rd WHO Informal Consultation for Improving Influenza Vaccine Virus Selection, 1–3 April 2014, was attended by experts from the WHO Global Influenza Surveillance and Response System (GISRS), national epidemiological institutions, national regulatory authorities, research and academic laboratories, institutions and organizations, veterinary institutions and organizations, human influenza vaccine manufacturers, and donor agencies and other stakeholders. In accordance with WHO policy, the meeting Chair, co-Chairs and joint Rapporteurs were required to complete the WHO form for the Declaration of Interests for WHO Experts prior to the consul- tation. During the opening session, the interests declared by these experts were disclosed to all participants. No current or recent (within the last 4 years) personal finan- cial or other interests relevant to the subject of the consultation were declared by the Chair (Dr. N Cox), co-Chairs (Dr. M Siqueira, Dr. M Giovanni, Dr. G Grohmann, Dr. J Katz and Professor M Rah- man) or joint Rapporteur Dr. A Waddell. A single financial interest declared by the joint Rapporteur Dr. J Wood was reviewed by the WHO Secretariat. As the joint Rapporteurs were to work collabo- ratively with the formal writing group shown above, and under the direct guidance of the Chair and co-Chairs, it was decided that the interest disclosed did not present a conflict of interest in relation to full meeting participation or the preparation of this report. Appendix B. List of participants Tarek Al Sanouri, Ministry of Health, Amman, Jordan Burmaa Alexander, National Centre for Communicable Dis- eases (NCCD), Ulaanbaatar, Mongolia Bandar Abdulrahman Alhammad, Saudi Food and Drug Authority, Riyadh, Kingdom of Saudi Arabia Wladimir Alonso, National Institutes of Health, Florianopolis, Brazil William K. Ampofo, University of Ghana, Accra, Ghana Eduardo Azziz-Baumgartner, Centers for Disease Control and Prevention, Atlanta, USA Amal Barakat, Ministère de la Santé, Rabat, Morocco ccine H A A t P H C U G T o A U d A M o S B U N J W.K. Ampofo et al. / Va Ralph Baric, University of North Carolina at Chapel Hill, Chapel ill, USA John R. Barnes, Centers for Disease Control and Prevention, tlanta, USA Ian Barr, VIDRL, Melbourne, Australia Uzma Bashir, National Institute of Health, Islamabad, Pakistan Elsa Baumeister, INEI-ANLIS “Carlos G. Malbrán”, Buenos Aires, rgentina Trevor Bedford, Fred Hutchinson Cancer Research Center, Seat- le, USA John W. Belmont, Baylor College of Medicine, Houston, USA Markus Blum, Swissmedic, Bern, Switzerland Peter Bogner, GISAID Initiative, Munich, Germany Nicola Boschetti, Crucell Switzerland, Bern, Switzerland Helen Bright, AstraZeneca Global Operations, Liverpool, UK Alfredo Bruno, Instituto Nacional de Investigación en Salud ública, Guayaquil, Ecuador Doris Bucher, New York Medical College, Valhalla, USA Mandeep Chadha, National Institute of Virology, Pune, India Yu Chen, 1st Affiliated Hospital of Zhejiang University, angzhou, China Ze Chen, Shanghai Institute of Biological Products, Shanghai, hina Nancy Cox, Centers for Disease Control and Prevention, Atlanta, SA Peter Daniels, CSIRO–Australian Animal Health Laboratory, eelong, Australia Ricardo das Neves Oliviera, Instituto Butantan, São Paulo, Brazil Valentina Di Francesco, NIAID/NIH/DHHS, Bethesda, USA Rassoul Dinarvand, Ministry of Health and Medical Education, ehran, the Islamic Republic of Iran Armen Donabedian, Biomedical Advanced Research and Devel- pment Authority, Washington, USA Ruben Donis, Centers for Disease Control and Prevention, tlanta, USA Gael Dos Santos, GSK Vaccines, Philadelphia, USA Maryna Eichelberger, Food & Drug Administration, Bethesda, SA Othmar G. Engelhardt, National Institute for Biological Stan- ards and Control, Potters Bar, UK Vincent Enouf, Institut Pasteur, Paris, France Vasily Evseenko, FORT LLC, Moscow, Russia Rodrigo Fasce, Public Health Institute of Chile, Santiago, Chile Louis F. Fries, Novavax Inc, Gaithersburg, USA Rebecca Garten, Centers for Disease Control and Prevention, tlanta, USA Lionel Gerentes, Sanofi Pasteur, Marcy l’Etoile, France Maria Giovanni, National Institutes of Health, Bethesda, USA Gary Grohmann, Department of Health, Symonston, Australia Yi Guan, Queen Mary Hospital, Hong Kong, China Julia Guillebaud, Institut Pasteur de Madagascar, Antananarivo, adagascar Vladimir Guriev, National Center for Public Health of Republic f Moldova, Chisinau, Republic of Moldova Rene Gysin, Swissmedic, Bern, Switzerland Alan Hay, National Institute for Medical Research, London, UK Siddhivinayak Hirve, Independent Consultant, Mont-sur-Rolle, witzerland Yu Hongje, Chinese Center for Disease Control & Prevention, eijing, China Sue Huang, Institute of Environmental Science and Research, pper Hutt, New Zealand Olav Hungnes, Norwegian Institute of Public Health, Oslo, orway Sandra Jackson, University of the West Indies, Kingston, amaica33 (2015) 4368–4382 4381 Daniel Jernigan, Centers for Disease Control and Prevention, Atlanta, USA Chun Kang, National Institute of Health, Seoul, Republic of Korea Jacqueline Katz, Centers for Disease Control and Prevention, Atlanta, USA Mirdad Kazanji, Institut Pasteur de Bangui, Bangui, Central African Republic Nikolai Khramstov, Protein Sciences Corporation, Meriden, USA Renu Lal, Centers for Disease Control and Prevention, Atlanta, USA Michael Lassig, University of Cologne, Cologne, Germany Karen Laurie, VIDRL, Melbourne, Australia Quynh Mai Le, National Institute of Hygiene and Epidemiology, Hanoi, Viet Nam Van Phung Le, National Institute for Control of Vaccine and Biologicals, Hanoi, Viet Nam Yan Li, Canadian Science Centre for Human and Animal Health, Winnipeg, Canada Ming-Tsan Liu, Independent Consultant, Taiwan, China Irma Lopez Martinez, Instituto de diagnóstico y Referencia Epi- demiologicos, Mexico D.F., Mexico Marta Luksza, Columbia University, New York, USA Niang Mbayame, Institut Pasteur de Dakar, Dakar, Senegal John McCauley, National Institute for Medical Research, Lon- don, UK Alla Mironenko, Institute of Epidemiology and Infectious Dis- eases, Kiev, Ukraine Cosue Miyaki, Instituto Butantan, São Paulo, Brazil Talat Mokhtari-Azad, Tehran University of Medical Sciences, Tehran, Islamic Republic of Iran Els Mol, Abbott Biologicals BV, Weesp, Netherlands Elisabeth Neumeier, GSK Biologicals, Dresden, Germany Richard Njoum, National Influenza Centre, Yaounde, Cameroon John Paget, Netherlands Institute for Health Services Research, Utrecht, Netherlands Pasi Penttinen, European Centre for Disease Prevention and Control, Stockholm, Sweden Alexander Precioso, Instituto Butantan, São Paulo, Brazil Katarina Prosenc, Institute of Public Health of the Republic of Slovenia, Ljubljana, Slovenia Mahmudur Rahman, Institute of Epidemiology, Disease Control and Research, Dhaka, Bangladesh Andrew Rambaut, Edinburgh University, Edinburgh, UK Isaias Raw, Instituto Butantan, São Paulo, Brazil Kjersti Margrethe Rydland, Norwegian Institute of Public Health, Oslo, Norway Pretty Multihartina Sasono, National Institute of Health Research and Development, Jakarta, Indonesia Pathom Sawanpanyalert, Ministry of Public Health, Non- thaburi, Thailand Richard Scheuermann, J. Craig Venter Institute, La Jolla, USA Brunhilde Schweiger, Robert Koch Institut, Berlin, Germany Vivi Setiawaty, Center for Biomedical and Pharmaceutical Research and Development, Jakarta, Indonesia Vivek Shinde, GSK Vaccines, Philadelphia, USA Yuelong Shu, National Institute for Viral Disease Control and Prevention, Beijing, China Marilda Siqueira, Instituto Oswaldo Cruz, Rio de Janeiro, Brazil Derek Smith, University of Cambridge, Cambridge, UK Igor Spinu, National Center for Public Health of Republic of Moldova, Chisinau, Republic of Moldova David Spiro, National Institutes of Health, Bethesda, USA James Stevens, Centers for Disease Control and Prevention, Atlanta, USA Ih-Jen Su, Independent Consultant, Taiwan, China 4 ccine U J e A U e s P U a U L T B B [ [ 382 W.K. Ampofo et al. / Va David E. Swayne, USDA/Agricultural Research Service, Athens, SA Masato Tashiro, National Institute of Infectious Diseases, Tokyo, apan Beverly Taylor, Novartis, Liverpool, UK Catherine Thompson, Public Health England, London, UK Florette Treurnicht, National Institute for Communicable Dis- ases, Johannesburg, South Africa Julie Villanueva, Centers for Disease Control and Prevention, tlanta, USA Anthony L. Waddell, Freelancer, Stanley, UK Lillian Waiboci, CDC Kenya, Nairobi, Kenya Niteen Wairagkar, Bill & Melinda Gates Foundation, Seattle, SA Sibongile Walaza, National Institute for Communicable Dis- ases, Johannesburg, South Africa Xiu-Feng (Henry) Wan, Mississippi State University, Missis- ippi State, USA Dayan Wang, National Institute for Viral Disease Control and revention, Beijing, China Richard Webby, St Jude Children’s Research Hospital, Memphis, SA Jerry Weir, Food & Drug Administration, Bethesda, USA David Wentworth, J. Craig Venter Institute, Rockville, USA John Wood, Formerly National Institute for Biological Standards nd Control, Potters Bar, UK Oliver Wildner, Swissmedic, Bern, Switzerland Xu Xiyan, Centers for Disease Control and Prevention, Atlanta, SA Zhang Xuemei, Changchun Institute of Biological Products Co., td., Sinopharm, Changchun, China Makoto Yamashita, University of Tokyo, Tokyo, Japan Thitipong Yingyong, Ministry of Public Health, Nonthaburi, hailand Maria Zambon, Health Protection Agency, London, UK Ye Zhiping, Center for Biologics Evaluation and Research, ethesda, USA Thedi Ziegler, Independent Consultant, Turku, Finland .1. WHO secretariat Claudia Alfonso, Essential Medicine and Health Products Terry Besselaar, Global Influenza Programme Sylvie Briand, Pandemic and Epidemic Diseases Caroline Brown, WHO Regional Office for Europe Hien Doan, Global Influenza Programme Erica Dueger, WHO Regional Office for the Western Pacific Julia Fitzner, Global Influenza Programme Keiji Fukuda, Health Security Christian Fuster, Global Influenza Programme [ Anne Huvos, Global Influenza Programme Marie-Paule Kieny, Health Systems and Innovations33 (2015) 4368–4382 Maja Lievre, Global Influenza Programme Jairo Mendez, WHO Regional Office for the Americas Dhamari Naidoo, Global Influenza Programme Catherine Oswald, Global Influenza Programme Dmitriy Pereyaslov, WHO Regional Office for Europe Raphael Slattery, Pandemic and Epidemic Diseases Katelijn Vandemaele, Global Influenza Programme Wenqing Zhang, Global Influenza Programme References [1] WHO Writing Group. Improving influenza vaccine virus selection. Report of a WHO informal consultation. Influ Other Respir Viruses 2012;6(2): 142–52 [http://onlinelibrary.wiley.com/doi/10.1111/j.1750-2659.2011.00277. x/abstract, accessed 26.05.14, WHO headquarters, Geneva, Switzerland, 14–16 June 2010]. [2] WHO Writing Group. Strengthening the influenza vaccine virus selection and development process. Outcome of the 2nd WHO Informal Consultation for Improving Influenza Vaccine Virus Selection. Vaccine 2013;31(32):3209–21 [http://www.sciencedirect.com/science/article/pii/S0264410X13006373, accessed 26.05.14, Centre International de Conférences (CICG), Geneva, Switzerland, 7–9 December 2011]. [3] Selection of clinical specimens for virus isolation and of viruses for ship- ment from National Influenza Centres to WHO Collaborating Centres. World Health Organization, 〈http://www.who.int/influenza/gisrs laboratory/ national influenza centres/20101206 specimens selected for virus isolation and shipment.pdf?ua=1〉; 2010 [accessed 29.05.14]. [4] Azziz Baumgartner E, Dao CN, Nasreen S, Bhuiyan MU, Mah-E-Muneer S, Al Mamun A, et al. Seasonality, timing, and climate drivers of influenza activ- ity worldwide. J Infect Dis 2012;206(6):838–46 [http://www.ncbi.nlm.nih.gov/ pubmed/22829641, accessed 01.11.14]. [5] De Mello WA, De Paiva TM, Ishida MA, Benega MA, Dos Santos MC, Viboud C, et al. The dilemma of influenza vaccine recommendations when applied to the tropics: the Brazilian case examined under alternative scenarios. PLoS One 2009;4(4):e5095 [http://www.plosone.org/article/info%3Adoi%2F10.1371%2 Fjournal.pone.0005095;jsessionid=4FD9FB6C0971EA38375DDE5AA427D08E, accessed 01.06.14]. [6] Saha S, Chadha M, Al Mamun A, Rahman M, Sturm-Ramirez K, Chittaganpitch M, et al. Influenza seasonality and vaccination timing in tropical and sub- tropical areas of southern and south-eastern Asia. Bull World Health Organ 2014;92(5):318–30 [http://www.who.int/bulletin/volumes/92/5/13-124412. pdf, accessed 01.06.14]. [7] Legrand J, Vergu E, Flahault A. Real time monitoring of the influenza vaccine field effectiveness. Vaccine 2006;24(44–46):6605–11 [http://www.ncbi.nlm. nih.gov/pubmed/16806607, accessed 01.06.14]. [8] Smith DR, Lapedes AS, de Jong JC, Bestebroer TM, Rimmelzwaan GF, Osterhaus AD, et al. Mapping the antigenic and genetic evolution of influenza virus. Sci- ence 2004;305(5682):371–6 [http://www.sciencemag.org/content/305/5682/ 371, accessed 01.06.14]. [9] WHO Manual for the Laboratory Diagnosis and Virological Surveillance of Influenza. Geneva: World Health Organization, 〈http://whqlibdoc.who.int/ publications/2011/9789241548090 eng.pdf〉; 2011 [accessed 26.05.14]. 10] Ducatez MF, Bahl J, Griffin Y, Stigger-Rosser E, Franks J, Barman S, et al. Feasibility of reconstructed ancestral H5N1 influenza viruses for cross-clade protective vaccine development. Proc Natl Acad Sci 2011;108(1):349–54 [http://www.pnas.org/content/108/1/349, accessed 21.06.14]. 11] Pan W, Dong Z, Meng W, Zhang W, Li T, Li C, et al. Improvement of influenza vaccine strain A/Vietnam/1194/2004 (H5N1) growth with the neuraminidase packaging sequence from A/Puerto Rico/8/34. Hum Vaccines Immunother 2012;8(2):252–9 [http://www.ncbi.nlm.nih.gov/pubmed/22426370, accessed 21.06.2014]. 12] Franco LM, Bucasas KL, Wells JM, Niño D, Wang X, Zapata GE, et al. Integrative genomic analysis of the human immune response to influenza vac- cination. elife 2013;2(June):e00299. http://elifesciences.org/content/2/e00299 [accessed 26.05.14].