INTRODUCTION — There have been five influenza pandemics during the past 100 years, and each has been caused by the emergence of a novel virus. In the 1957 and 1968 pandemics, the new viruses contained components of previous human as well as avian influenza viruses. The origin of the influenza virus responsible for the 1918 pandemic, which killed more people in a single year than the bubonic plague, remains uncertain, but it appears to have been an adapted avian influenza strain. The emergence of a novel H1N1 human-swine-avian reassortant virus in 2009 in North America resulted in a new pandemic.

Sporadic transmission of avian influenza H5N1 to at least 860 humans since 2003, and more recently of avian influenza H5N6, has prompted concerns that conditions are suitable for emergence of a pandemic H5 influenza virus [1]. Two features of avian influenza H5N1 outbreaks are striking: the predominance of children and young adults, and the high mortality rate [2,3]. However, seroprevalence studies have found that some exposed individuals may have had a subclinical or mild infection, suggesting that the reported case-fatality rate may be an overestimate.

Issues related to avian influenza vaccines will be reviewed here. Treatment and prevention of avian influenza (other than vaccination) and the epidemiology, transmission, pathogenesis, clinical manifestations, and diagnosis of avian influenza are discussed separately. (See "Avian influenza A H5N1: Treatment and prevention" and "Avian influenza: Epidemiology, transmission, and pathogenesis" and "Avian influenza: Clinical manifestations and diagnosis".)

GENERAL CONCEPTS — Immunization against avian influenza is an active area of research. A non-adjuvanted subvirion H5N1 avian influenza vaccine was approved by the US Food and Drug Administration in 2007 [4]. An oil-in-water emulsion (AS03)-adjuvanted subvirion H5N1 avian influenza vaccine was approved for pre-pandemic use in the European Union in 2008 and in the United States in 2013 [5,6], and an MF59-adjuvanted monovalent H5N1 avian influenza vaccine was approved in the United States in 2020 [7]. (See 'Studies in humans' below.)

Production of H5N1 vaccines by conventional means has been hampered by technical manufacturing difficulties and modest immunogenicity in immunologically naïve humans, potentially requiring the administration of several doses of vaccine to induce seroprotection. As pandemics occur unpredictably and evolve rapidly, high-priority research goals include improving production speed and increasing the quantity of vaccine that can be produced rapidly. Areas of research include use of cell culture systems, dose-sparing approaches (eg, whole-virion formulation, intradermal administration), and use of adjuvants and live attenuated viruses to induce more robust immune responses with lower quantities of antigen [8].

In addition, genetic evolution among circulating avian H5N1 viruses has given rise to several virus clades and subclades that are antigenically diverse, posing difficulties in selecting the best matched viruses from which to prepare vaccines in advance of a pandemic. This issue and the modest immunogenicity of H5N1 vaccines have led to the concept of prime-boosting, whereby an immunologically naïve individual could be immunized or primed with H5N1 vaccine and subsequently boosted (potentially many years later) with a different heterologous H5N1 vaccine to induce rapid cross-clade seroprotection.

Technical difficulties — Production of annual vaccines against human influenza generally involves bulk growth of influenza virus in embryonated hens' eggs. Use of highly pathogenic avian influenza viruses to produce vaccines by this traditional method is not feasible because of safety and technical reasons. The highly pathogenic avian influenza viruses are lethal to eggs and cannot be grown in sufficient quantities. (See "Seasonal influenza vaccination in adults" and "Seasonal influenza in children: Prevention with vaccines".)

Work with highly pathogenic avian influenza viruses must be undertaken in biosecure facilities in order to protect workers and limit risks of environmental contamination. Manufacturing capacity to produce H5N1 vaccine is constrained by lack of large-scale biocontainment facilities and the requirement to also supply annual epidemic influenza vaccines at the same time.

Alternative methods of vaccine production have been explored. One development in vaccine research has been the use of reverse genetics to create a well-matched nonpathogenic H5N1 virus that has high capacity to grow in eggs. However, manufacturers have reported that virus growth yields in eggs are lower than expected with reverse-genetics H5N1 viruses than with seasonal influenza viruses, reducing the amount of antigen able to be used for vaccine formulation.

Another method that avoids the problem of virus lethality to eggs is to grow H5N1 virus in cell culture [9]. A whole-virion vaccine produced using this technique has proven to be well tolerated and immunogenic in a trial described below. An additional potential benefit of this approach is that the production of a vaccine would not be limited by the time constraints of sourcing sufficient quantities of embryonated eggs, which may have limited availability [10]. (See 'Whole-virion H5N1 vaccines' below.)

In the event of an emerging H5N1 pandemic, there would be enormous worldwide demand for H5N1 vaccines. It is essential that pandemic planning strategies include clinical assessment of avian influenza virus vaccine candidates with potential adjuvants and delivery systems to optimize use of available antigenic material in advance of a pandemic [11,12].

Manufacturing capacity — In 2005, the global manufacturing capacity of seasonal trivalent influenza vaccines (containing 15 mcg per strain) was approximately 400 million doses per year. Assuming that people require two doses of monovalent H5N1 pandemic vaccine, the available doses will be approximately 600 million. If 90 mcg (six times the normal dose) is needed, then the available number of doses drops to only 100 million. By using 30 mcg (twice the normal dose) in an alum-adjuvanted vaccine, the number of available dose increases to 300 million. This supply will need to be shared among a global population of 6 billion. Increasing seasonal vaccine demand and investment in production facilities have led to increased global manufacturing capacity, with estimated global annual capacity for production of 1.467 billion doses in 2015 [13]; however, significant shortfalls in manufacturing capacity remain.

ANIMAL MODELS — Animal models have been invaluable for the development of avian influenza vaccines, particularly since it is possible to test not only immunogenicity but also clinical efficacy against a challenge of avian influenza virus. Various avian influenza vaccines have shown promising results in animal models, including inactivated vaccines [14], live attenuated vaccines [15], vector-based vaccines [16], vaccines produced using reverse genetics [17-19], plasmid DNA vaccines [20], a DNA vaccine that encodes virus-like particles (VLPs) [21], and a computationally optimized hemagglutinin VLP vaccine that was designed to elicit broadly reactive immune responses [22]. In addition, strategies that include adjuvants [23], a lipid delivery system [20], or a prime-boost regimen have demonstrated efficacy in such models [18].

STUDIES IN HUMANS — Multiple H5N1 avian influenza vaccines have been evaluated, and in the United States, several have been approved by the Food and Drug Administration (FDA) [4,6,7]. Although there have been few direct comparisons among vaccine formulations, it is clear that oil-in-water emulsion-adjuvanted and whole-virion H5N1 influenza vaccines offer advantages over conventional subvirion vaccines.

A major problem with the development of an effective vaccine against avian influenza (and H5, in particular) has been poor immunogenicity in humans naïve to avian influenza antigens [24-26]. Another significant issue is that the serology tests used as measures of immunogenicity for influenza vaccines are not standardized and display marked variation among laboratories. As a result, caution must be used when comparing results from different studies [27]. International standards for detecting antibody to clade 1 H5 antigens have been established by the World Health Organization and will help standardize vaccine immunogenicity [28].

Non-adjuvanted H5N1 vaccine — In a multicenter randomized trial, the safety and efficacy of a non-adjuvanted subvirion H5N1 avian influenza vaccine was evaluated in 451 healthy adults [29]. Participants received two doses of vaccine containing 90, 45, 15, or 7.5 mcg of hemagglutinin antigen or placebo. Although the vaccine was safe, immunogenicity was poor [30]. The only group in which more than 50 percent of subjects reached the predefined threshold for immunogenicity occurred with administration of 90 mcg, a total dose nearly 12 times that of seasonal influenza vaccines. The vaccine was well tolerated; the most common side effects were pain at the injection site, headache, malaise, and muscle pain.

This vaccine was approved by the FDA in April 2007. The vaccine is intended for use in adults from 18 to 65 years of age and is given as two doses one month apart. The vaccine will not be sold commercially; instead, it will be purchased by the United States government for inclusion in the National Stockpile for distribution by public health officials as needed [31].

In a follow-up study, 337 participants received a third dose of vaccine containing 90, 45, 15, or 7.5 mcg of hemagglutinin antigen. Microneutralization geometric mean titers were ≥1:40 in 78, 67, 43, and 31 percent of recipients in each group, respectively, one month after the third dose [32]. Five months later, microneutralization geometric mean titers remained significantly greater than titers after the second dose, suggesting that, after priming with vaccine, antibody responses can be further enhanced by vaccine boosting.

A number of studies have evaluated prime-boosting schedules with H5 vaccines [33,34]. Thirty-seven individuals who had been vaccinated eight years earlier with a pilot recombinant hemagglutinin H5N1 vaccine received a single intramuscular injection of an inactivated subvirion H5N1 vaccine [33]. A higher percentage of those who had been vaccinated previously responded to the vaccine (73 versus 41 percent) compared with H5-naïve individuals who received two doses of the inactivated subvirion H5N1 vaccine. In a separate study, low-dose priming with a clade 1 non-adjuvanted H5N1 vaccine improved the rapidity, magnitude, and cross-reactivity of the immunologic response following a single high-dose booster of a mismatched non-adjuvanted clade 2 H5N1 vaccine [34]. These studies suggest that a prime-boost strategy could enhance the immunogenicity of an avian pandemic influenza vaccine.

There is interest in the immunogenicity of administering two doses of H5N1 vaccines with a short interval between doses since such a strategy could be beneficial following a pandemic declaration when rapid protection is required. In a study in which non-adjuvanted subvirion H5N1 vaccines containing clade 1 or clade 2 antigens were administered 7, 14, 28, or 180 days apart, the accelerated schedule of vaccines administered on days 0 and 14 was as immunogenic as a schedule of vaccines on days 0 and 28, but both schedules were less immunogenic than a schedule of vaccines on days 0 and 180 [35]. The schedule that involved doses on days 0 and 7 was the least immunogenic. The use of an accelerated schedule has also been evaluated for AS03-adjuvanted vaccines. (See 'AS03' below.)

Adjuvanted H5N1 vaccines — Adjuvantation enables lower doses of antigen to be given, which has the potential to increase production capacity of pandemic influenza vaccine [36,37]. Oil-in-emulsion adjuvanted vaccines (AS03 and MF59), but not alum-adjuvanted vaccines, are associated with better overall immune responses [24,38-42].

Some of these "antigen-sparing" approaches have used subvirion versus whole-virion vaccine design.

Subvirion H5N1 vaccines

Aluminum — Aluminum adjuvanted avian influenza vaccines have induced only modest antibody responses compared with non-adjuvanted vaccines [12,38,40,41]. As an example, in a randomized trial, 600 adults received two doses one month apart of subvirion inactivated influenza A/H5N1 vaccine containing 7.5, 15, or 45 mcg of hemagglutinin (HA) with or without 600 mcg of aluminum hydroxide (AlOH), or 3.75 mcg of HA with or without 300 mcg of AlOH [41]. After the first dose of vaccine, there were no differences between those who received adjuvanted and non-adjuvanted vaccines and there was only a modest improvement in immunogenicity after the second dose in those who received 7.5 mcg of HA with AlOH.

AS03 — Two recombinant H5N1 subvirion vaccines were administered 21 days apart to 400 healthy volunteers 18 to 60 years of age [39]. Four antigen doses (3.8, 7.5, 15, and 30 mcg hemagglutinin) with or without AS03, a novel oil-in-water emulsion adjuvant, were administered to eight groups of study participants. The following findings were noted:

●After the first adjuvanted vaccine dose, all doses of 7.5 mcg and higher led to seroprotection, as defined by standard European and United States criteria.

●After the second adjuvanted vaccine dose, all doses (including the lowest antigen dose of 3.8 mcg) met these criteria.

●The non-adjuvanted vaccines were poorly immunogenic except for the highest antigen dose (30 mcg), which led to seroprotection only after the second dose of vaccine.

Based upon these results, this antigen-sparing AS03-adjuvanted H5N1 vaccine was licensed for use in Europe [5].

Subsequently, two-dose regimen of an even lower dose (1.9 mcg) of oil-in-water emulsion adjuvanted H5N1 vaccine was more immunogenic (92 percent had a fourfold increase in microneutralization titers versus 42 percent) and induced a higher proportion of cross-neutralizing antibodies (39 to 65 percent versus 7 percent) compared with 7.5 mcg of non-adjuvanted vaccine [43].

In studies enrolling older individuals >61 years of age (in whom vaccine might be expected to have lower response rates), a two-dose series of 3.75 and 7.5 mcg of AS03-adjuvanted vaccine given 21 days both met the European Committee for Human Medicinal Products criteria for seroconversion and seroprotection [44,45].

AS03-adjuvanted H5N1 vaccines are not only highly immunogenic against the homologous vaccine strain but also induce cross-clade neutralizing antibodies against circulating antigenically distinct H5N1 viruses [39,43-45], demonstrating a potential pre-pandemic use. (See 'Prepandemic stockpiling and vaccination' below.)

As stated above, there is interest in the immunogenicity of administering two doses of H5N1 vaccines with a short interval between doses since such a strategy could be beneficial during an evolving pandemic (see 'Non-adjuvanted H5N1 vaccine' above). In one trial, adults aged 18 to 64 years were randomly assigned to receive an AS03 subvirion vaccine that contained 3.75 mcg of the HA antigen 21 days apart (standard schedule), 14 days apart, 7 days apart, or on the same day [46]. Although the highest antibody titers against the homologous strain of virus were observed with the standard schedule, each of the accelerated schedules met predetermined immunogenicity licensing criteria.

The most common adverse effect of AS03-adjuvanted vaccine seems to be injection site pain [6]. Other common adverse effects include myalgias, headache, fatigue, and injection site redness and swelling.

Following the 2009 H1N1 swine influenza pandemic, increases in childhood narcolepsy, a sleep disorder characterized by excessive daytime somnolence, were observed in Scandinavian countries where vaccination coverage with administration of AS03-adjuvanted vaccines was high [47]. Similar observations were described in England, France, and Canada, suggesting a causal association with administration of AS03-adjuvanted pandemic influenza vaccine [48-50]. This is discussed in greater detail separately.

MF59 — A randomized trial compared a two-dose regimen of an H5N1 vaccine in 394 healthy subjects who received either placebo or vaccine alone at 45, 30, or 15 mcg per dose; vaccine at 15 or 7.5 mcg per dose adjuvanted with MF59, an oil-in-water adjuvant containing squalene; or vaccine at 30, 15, or 7.5 mcg per dose adjuvanted with aluminum hydroxide [40]. One month after the second dose, seroprotection was achieved significantly more often with a 15 mcg dose of the MF59-adjuvanted vaccine compared with a 45 mcg dose of non-adjuvanted vaccine (63 versus 29 percent). The aluminum hydroxide adjuvanted vaccine did not provide additional benefit compared with the non-adjuvanted vaccine.

In another trial, 40 healthy adults were vaccinated with MF59-adjuvanted H5N1 subunit vaccine (at a dose of 7.5 or 15 mcg) or non-adjuvanted H5N1 subunit vaccine (at a dose of 15 mcg) administered on days 1, 22, and 202 [51]. Only the MF59-adjuvanted vaccine elicited high titers of neutralizing antibodies, a large pool of H5N1-specific memory B cells, and CD4+ T cells that were broadly reactive with drifted H5 virus. An increase in the frequency of H5-specific CD4+ T cells measured after the first dose of vaccine predicted the rise in neutralizing antibodies after booster immunization as well as their persistence six months later. Similar findings were observed in a trial of a virosomal vaccine [52] (see 'Virosomal vaccine' below). Although a larger trial is needed to confirm these findings, this trial suggests that the CD4+ T cell response serves as an early marker of immunogenicity.

Another trial evaluated the immunogenicity of an H5N1 influenza (Anhui) vaccine with or without MF59 in individuals who had previously received one or two doses of an antigenically distinct non-adjuvanted H5N1 influenza (Vietnam) vaccine one year earlier and in individuals who had not previously received an H5N1 influenza vaccine [53]. Among those who had not previously been vaccinated, the optimal vaccine dose was 7.5 mcg of antigen with MF59, resulting in seroconversion in 81 percent of participants. The most robust response to non-adjuvanted vaccine occurred at the highest dose, 90 mcg. Individuals previously vaccinated with Vietnam vaccine and then given Anhui vaccine developed cross-reacting antibodies to both antigens with MF59-adjuvant enhancing the response. In contrast, H5 vaccine-naïve individuals who received two doses of Anhui vaccine developed little or no antibody to Vietnam hemagglutinin.

A trial was performed to assess whether the concomitant use of an H5N1 influenza vaccine and a seasonal influenza vaccine would be immunogenic; 401 healthy adults were randomly assigned to receive MF59-adjuvanted H5N1 vaccine as a mixed injection with a non-adjuvanted inactivated seasonal influenza vaccine, simultaneous separate injections of the H5N1 and seasonal influenza vaccines, or a single injection of either the H5N1 or the seasonal influenza vaccine [54]. When administered as either a mixed injection or as simultaneous separate injections, both the H5N1 and the seasonal influenza vaccines were as immunogenic and safe as either vaccine given alone. An MF59-adjuvanted tetravalent vaccine containing A/H5N1 and seasonal A/H3N1, A/H1N1, and B components has also been tested in 601 healthy adults [55]. The administration of the tetravalent vaccine induced similar antibody responses and reactogenicity compared to the administration of separate A/H5N1 and seasonal vaccines, suggesting that pre-pandemic priming could be incorporated into seasonal influenza immunization schedules.

Whole-virion H5N1 vaccines

●A trial evaluated the safety and immunogenicity of four different formulations of an alum-adjuvanted whole-virion H5N1 vaccine in 120 volunteers in China [56]. The formulation with the best immunogenicity was the 10 mcg group after two vaccine doses. These data suggest that use of whole-virion vaccines may allow a dose-sparing approach while achieving high levels of protective efficacy [57].

●Another trial in China evaluated an inactivated alum-adjuvanted whole-virion H5N1 vaccine [58]. Two doses of vaccine containing 5, 10, or 15 mcg of hemagglutinin were administered to 301 volunteers 28 days apart; two doses of 10 mcg of vaccine were given to 101 volunteers 14 days apart. The highest antibody responses were seen in the 15 mcg vaccine group, with 90 and 100 percent of subjects having a titer ≥1:40 by hemagglutination inhibition and microneutralization assays, respectively. Higher antibody responses were elicited in those who were vaccinated 28 days apart than in those who were vaccinated 14 days apart including against two heterologous H5N1 virus strains.

●A Taiwanese phase I clinical trial evaluated an inactivated alum-adjuvanted whole-virion H5N1 vaccine [59]. Two doses of vaccine containing 7.5, 15, or 30 mcg of hemagglutinin were administered to 36 adults 21 days apart. The highest antibody responses were seen in those who had received the highest dose, which resulted in seroconversion in 91 percent of recipients.

●A randomized, dose-escalation, phase I and II trial evaluated the safety and immunogenicity of a Vero cell–derived whole-virion H5N1 vaccine in 275 volunteers [9]. Subjects received two doses of vaccine 21 days apart that each contained 3.75 mcg, 7.5 mcg, 15 mcg, or 30 mcg of HA antigen with alum-adjuvant or 7.5 or 15 mcg of HA antigen without adjuvant. The vaccine induced neutralizing antibodies against the clade 1 virus strain used in the vaccine and also against clade 2 and 3 strains.

The maximum responses occurred in those who received the formulations containing 7.5 mcg or 15 mcg of the HA antigen without adjuvant, suggesting that a non-adjuvanted whole-virion vaccine is a promising candidate H5N1 vaccine candidate. Furthermore, the mode of production (in Vero cell cultures) has potential advantages over egg grown vaccines. (See 'Technical difficulties' above.)

●In a follow-up study, a subset of the individuals received a booster containing 7.5 mcg of HA antigen without adjuvant from an antigenically distinct clade 2.1 strain 12 to 17 months after the priming regimen [60]. The prime-boost regimen resulted in higher cross-reacting antibody responses against clades 1, 2.1, 2.2, and 2.3 viruses than those induced after the priming regimen.

Virosomal vaccine — In a phase I trial of 60 adults who received two doses of a virosomal H5N1 vaccine adjuvanted with the immunostimulating complex Matrix M, the vaccine effectively induced CD4+ Th1 cell responses [52]. In addition, the frequency of influenza-specific Th1 cells after the first dose of vaccine predicted subsequent seroprotective antibody responses (measured by hemagglutinin inhibition, microneutralization, and single radial hemolysis assays) after the second dose of vaccine. These results suggest that Th1 cell responses may serve as a predictive biomarker for an effective humoral immune response. Similar findings were observed in a trial of an MF59-adjuvanted H5N1 vaccine [51]. (See 'MF59' above.)

Modified vaccinia virus vector vaccine — An H5N1 avian influenza modified vaccinia virus Ankara vaccine (brand name Foclivia) was approved in Canada in 2021. In a phase 1/2a study, 80 adults (age 18 to 28 years) received 1 or 2 doses of modified vaccinia virus Ankara vaccine encoding the hemagglutinin gene of influenza A/Vietnam/1194/2005 at 10⁷ or 10⁸ plaque forming doses or empty vector vaccine [61]. Most recipients had local and systemic reactions. The higher vaccine dose induced higher hemagglutination inhibition antibody responses at 4 and 8 weeks, with higher responses after the 12-month booster. Subsequent analysis of samples against a panel of antigenically distinct H5 viruses (including newly emerged H5N6, H5N8, and different clades of H5N1 viruses) found potent cross-reactive antibody and T cell responses after booster immunization at 12 months [62]. (See 'Prepandemic stockpiling and vaccination' below.)

Adjuvanted H5N3 candidate vaccines — The ability to induce and boost antibody responses that are cross-reactive to antigenically diverse H5 strains would be beneficial in a strategy in which at-risk individuals are vaccinated prior to a pandemic. Before the adoption of reverse genetics to attenuate the virulence of H5N1 viruses, an approach used was to develop adjuvanted candidate vaccines based upon a nonpathogenic influenza variant, H5N3 influenza A [24,25,36,63].

A clinical trial of MF59 or non-adjuvanted H5N3 vaccine (7.5, 15, and 30 mcg hemagglutinin) found that two doses of the adjuvanted vaccine induced neutralizing antibody titers and seroconversions, whereas the non-adjuvanted vaccine resulted in poor antibody responses [36]. A third dose of the same vaccine formulation 16 months later induced broadly cross-reactive immune responses to a range of wild-type H5N1 virus variants. Again, the non-adjuvanted vaccine was associated with poor antibody responses.

In an open-label study, 24 subjects who had been primed with non-adjuvanted or MF59-adjuvanted H5N3 vaccines in clinical trials were boosted up to seven years later with an antigenically distinct single low dose (7.5 mcg) of MF59-adjuvanted H5N1 vaccine [63]. Comparator subjects were those who had not previously received H5 vaccine. Sera were taken at days 7, 14, 21, and 42 after boosting, and antibody titers to all known human H5 virus variants at the time were measured. Among primed subjects, there was a higher response rate to all strains than in the unprimed subjects. Subjects primed with MF59-adjuvanted vaccine had higher response rates than those primed with non-adjuvanted vaccine. At day 7 after boosting, over 80 percent of MF59-primed subjects had protective neutralizing titers to all strains tested. In addition, significant numbers of cross-reactive H5-specific memory B cells were induced by the booster dose, suggesting that pre-pandemic H5 priming would induce long-lasting memory responses [64].

H9N2 candidate vaccines — H9N2 vaccine candidates have been assessed in several studies.

Non-adjuvanted H9N2 vaccine — In the United Kingdom, healthy adults aged 18 to 60 years were randomly assigned two doses of inactivated whole virus or subunit A/Hong Kong/1073/99 (H9N2) vaccine [26]. Surprisingly, more than 40 percent of prevaccination serum samples had detectable anti-H9N2 reactivity, suggesting cross-reactivity from exposure to earlier circulation hemagglutinins. Consequently, subjects could be divided into those who were immunologically naïve or primed.

For primed individuals, one dose of either vaccine was sufficient to boost anti-H9 responses. However, in naïve subjects, one dose of either vaccine was poorly immunogenic. Although whole-virus vaccine was more immunogenic than subunit vaccine, the majority of subjects failed to achieve seroprotective responses even after two doses.

Adjuvanted H9N2 vaccine — As noted for the H5N1 vaccines, the use of adjuvanted H9N2 vaccine is associated with greater immunogenicity and allows for a lower dose of antigen to be used in vaccine design.

●In a phase I double-blind randomized trial, the safety and immunogenicity of inactivated H9N2 vaccine was evaluated with and without MF59 adjuvant [65]. Geometric mean antibody assays demonstrated that one dose of adjuvanted vaccine produced similar titers as two doses of non-adjuvanted vaccine.

●Another study compared the safety and immunogenicity of AS03-adjuvanted inactivated H9N2 vaccine containing 1.9 or 3.75 mcg hemagglutinin with non-adjuvanted vaccine containing 15 mcg hemagglutinin administered 21 days apart [66]. After a single dose, all adjuvanted formulations achieved seroprotection and seroconversion rates of 95 and 90 percent respectively. The adjuvanted vaccines were thus dose sparing (1.90 to 3.75 versus 15 mcg).

●In another study, alum-adjuvanted whole virus A/Hong Kong/1073/99 (H9N2) vaccine was compared with subunit vaccine containing either 7.5, 3.8, or 1.9 mcg H9 hemagglutinin and plain whole-virus 15 mcg vaccine [37]. The use of alum in vaccine preparation allowed a reduction in H9 content to 1.9 mcg per dose while maintaining immunogenicity.

Live attenuated H9N2 vaccine — In an open-label study of a live attenuated H9N2 vaccine, 92 percent of H9-seronegative individuals had ≥4-fold increases in hemagglutination inhibition antibody, and 79 percent had ≥4-fold increases in neutralizing antibody after two doses [67]. The vaccine was well tolerated and vaccine shedding was minimal.

VACCINE EFFICACY — It is unknown if the antibodies induced by these vaccines will afford sufficient protection in the event of a pandemic virus. There remains little data on vaccine efficacy in young, older adult, or immunocompromised patients.

NOVEL VACCINES — Clinical studies assessing novel adjuvants and formulations of pandemic influenza vaccines are underway or have been completed [68]:

●Multiple clinical trials assessing pandemic vaccine candidates have been conducted or are underway, including inactivated egg-grown subunit and whole-virion vaccine, cell culture–derived vaccine, adjuvanted (aluminum hydroxide, aluminum phosphate, polyoxidonium, and MF59), and live attenuated virus formulations [12].

●Assessments of subunit, adjuvanted, virosomal, and whole-virus H7 and H9 vaccine candidates are being performed [69].

●The successful deployment of adenoviral vector based vaccines against severe acute respiratory syndrome coronavirus 2 has demonstrated the potential of an emerging technology that could be designed to protect against avian influenza viruses [70].

PREPANDEMIC STOCKPILING AND VACCINATION — There are several options available for the use of pandemic H5N1 avian influenza vaccines [71-74]. Because circulating H5N1 viruses are evolving genetically, and thus future pandemic strains are unknown, a well-matched and so-called "pandemic-specific" vaccine cannot be produced until the pandemic virus has emerged and been characterized. Limitations in vaccine production capacity as well as the logistics of distribution and two-dose administration schedules are likely to delay widespread availability of pandemic-specific vaccine for several months following the onset of a pandemic.

Pre-pandemic vaccine production and stockpiling may, in part, offset vaccine availability limitations. Some governments, including that of the United States, have stockpiled H5N1 vaccine as part of their preparedness planning. However, antigenic diversity of circulating H5N1 viruses and uncertainty of the next pandemic virus subtype mean that vaccine prepared in advance of the pandemic may not be well matched to the eventual outbreak strain. Furthermore, even if stockpiles of pre-prepared influenza antigen are held, this may not address limitations in adjuvant availability or secondary manufacturing issues such as syringe filling and vaccine formulation. Defining the highest risk groups for vaccination campaigns during a pandemic will depend upon several factors, including which age groups are most affected [72]. The shelf life of stored bulk antigen is uncertain, but an open-label study of AS03A-adjuvanted H5N1 vaccine formulated with bulk antigen that had been stored for four years found that immune responses were similar to those elicited by newly manufactured vaccine [75]. In one randomized, double-blinded trial, influenza H5N1 antigen was administered with and without MF59 adjuvant using material from the United States National Pre-Pandemic Influenza Vaccine Stockpile [76]. At the time of immunization, the oldest stockpiled influenza H5N1 antigen was 12 years old, and the oldest stockpiled MF59 adjuvant was 7 years old. The tolerability and immunogenicity were comparable with historic clinical trial data, despite the extended storage years of antigen or adjuvant.

One approach that has been proposed is pre-pandemic vaccine administration to specific segments of the population. Administration of pre-pandemic H5 vaccine might induce long-acting memory responses that could be boosted by single revaccination in the future or even by exposure to the emergent virus itself. Priming subjects with MF59-adjuvanted H5N3 vaccine and boosting many years later with antigenically distinct heterologous H5N1 vaccine provide evidence that this is a viable approach [63,64]. Similarly, a heterologous prime-boost regimen of a whole-virion H5N1 vaccine with a booster given 12 to 17 months after the priming regimen resulted in higher antibody responses than the priming regimen alone, as well as cross-reactive antibodies to related clades [60]. (See 'Adjuvanted H5N3 candidate vaccines' above and 'Whole-virion H5N1 vaccines' above and 'Modified vaccinia virus vector vaccine' above.)

The development of tetravalent vaccines combining H5N1 with seasonal influenza antigens [54,55] indicate that pre-pandemic priming could be incorporated into routine seasonal influenza immunization programs. However, the widespread use of pre-pandemic vaccine among the population must be balanced with the risk of unexpected adverse reactions. This approach should be considered first in those at higher risk of exposure to pandemic influenza, such as healthcare workers [73].

PUBLIC HEALTH PERSPECTIVE — Public health issues related to avian influenza are discussed separately. (See "Avian influenza A H5N1: Treatment and prevention", section on 'Public health perspective'.)

SUMMARY

●Sporadic transmissions of avian influenza H5N1 to more than 860 humans since 2003 have prompted concerns that conditions are suitable for emergence of a pandemic H5 influenza virus. (See 'Introduction' above.)

●Production of H5N1 vaccines by conventional means has been hampered by technical manufacturing difficulties and modest immunogenicity in immunologically naïve humans, potentially requiring the administration of several doses of vaccine to induce seroprotection. As pandemics occur unpredictably and evolve rapidly, high-priority research goals include improving production speed and increasing quantity of vaccine that can be rapidly produced. Areas of research include use of cell culture systems, dose-sparing approaches (eg, whole-virion formulation, intradermal administration), and use of adjuvants and live attenuated viruses to induce more robust immune responses with lower quantities of antigen. (See 'General concepts' above.)

●A non-adjuvanted subvirion H5N1 avian influenza vaccine was approved by the US Food and Drug Administration in April 2007. An oil-in-water emulsion (AS03)-adjuvanted subvirion H5N1 avian influenza vaccine was approved in the European Union in 2008 and in the United States in 2013. An MF59-adjuvanted monovalent H5N1 avian influenza vaccine was approved in the United States in 2020. (See 'General concepts' above.)

●Although there have been few direct comparisons among vaccine formulations, it is clear that oil-in-water emulsion-adjuvanted and whole-virus H5N1 avian influenza vaccines offer advantages over conventional subvirion vaccines. (See 'Studies in humans' above.)

●Some governments, including that of the United States, have stockpiled H5N1 avian influenza vaccine as part of their pandemic preparedness planning. However, antigenic diversity of circulating H5N1 viruses and uncertainty of the next pandemic virus subtype mean that vaccine prepared in advance of the pandemic may not be well matched to the eventual outbreak strain. Another approach that has been proposed is pre-pandemic vaccine administration to specific segments of the population. Administration of pre-pandemic H5 vaccine might induce long-acting memory responses that could be boosted by single revaccination in the future or even by exposure to the emergent virus itself. (See 'Prepandemic stockpiling and vaccination' above.)