Searching for Improved Flu Vaccines—The Time Is Now

(See the Major Article by Flannery et al, on pages 8–15.) Influenza A and B viruses cause seasonal epidemics between late fall and early spring annually in the United States. Disease impact between the 2010–2011 and 2017–2018 seasons has been estimated to range from 9.2 to 47.9 million illnesses, 1.2 to 22.1 million outpatient medical visits, 139 to 953 thousand hospitalizations, and 12 to 79.4 thousand deaths each year [1, 2]. The Advisory Committee on Immunization Practices (ACIP) recommends annual immunization of all persons >6 months of age who do not have contraindications to receipt of influenza vaccine. Immunization is the primary means of influenza prevention, and it may provide both direct (in the vaccine recipient) and indirect (herd immunity) protection [3–5]. The Centers for Disease Control and Prevention maintains surveillance networks through which influenza vaccine effectiveness (VE) can be estimated, and over the past 10 years, adjusted VE (aVE) against all influenza medically attended illness has ranged from 19% in the 2014–2015 season to 60% in the 2010–2011 season [6]. Despite the relatively low protection afforded by vaccination in some seasons, millions of cases of medically attended illness, tens of thousands of hospitalizations, and thousands of deaths have been averted annually [1, 2]. In the current issue of the Journal of Infectious Diseases, Flannery et al [7] in the US Flu VE Surveillance Network report aVE estimates among ambulatory persons who presented with an acute respiratory illness during the 2018–2019 influenza season. Influenza B viruses represented <3% of isolates identified in the study. Influenza A (H1N1)pdm09 viruses predominated in the first part of the influenza season, and aVE against these viruses for all age groups was 44%. During this time, there was also a low level of circulation of H3N2 viruses similar to the strain contained in the vaccine (3C.2a1 genetic group), and aVE against these viruses was estimated to be 46% (95% confidence interval [CI], 11%–68%). However, in the latter part of the season, a drifted H3N2 variant emerged (3C.3a genetic group), and vaccination afforded no significant measurable protection (aVE, 5%; 95% CI, −10% to 19%) overall or in any age group. Overall, aVE against all H3N2 virus infections also was not significant (aVE, 9%; 95% CI, −4% to 20%) because the 3C.3a H3N2 viruses were identified in more than 90% of patients infected with H3N2 strains [7]. The low aVE observed against H3N2 strains compared with H1N1pdm09 or B strains this past season is not a new phenomenon. A meta-analysis of test-negative design VE studies reported between 2004 and 2015 found that pooled VE against H3N2 virus infections was 33% compared with 54% for influenza B and 61% for H1N1pdm09 virus infections [4]. Belongia and McLean [8] recently reviewed possible factors contributing to the lower VE associated with H3N2 virus infection; these are summarized in Table 1. The factors can be divided into 3 areas: virus, host, and vaccine. Of these, the only modifiable group of factors is the vaccine. Factors Potentially Associated With Decrease VE Against Influenza A(H3N2) Viruses and Possible Solutionsa Abbreviations: VE, vaccine effectiveness. aAdapted from Belongia and McLean [8]. Factors Potentially Associated With Decrease VE Against Influenza A(H3N2) Viruses and Possible Solutionsa Abbreviations: VE, vaccine effectiveness. aAdapted from Belongia and McLean [8]. Virus evolution is driven by population immunity [9], and H3N2 viruses have circulated since 1968. Over the past 5 decades, many different antigenic variants have arisen. Viruses belonging to different subclades currently circulate in human populations during any given season, and which of these will emerge as a major epidemic strain is difficult to predict. Although influenza surveillance programs have expanded greatly around the world over the past decade, and much more information is available to aid in vaccine strain selection through the characterization of virus isolates, strain selection is complicated by the complex and imperfectly understood interactions between virus and host that affect virus emergence. Thus, even with expanded information, as Flannery et al [7] note, it remains “a challenging race against the clock” to select strains for vaccine production in time for the next influenza season because of the reliance on egg-based technology for the production of the majority of influenza vaccines used in this country. What can be done to address this dilemma? The simple answer is to make better influenza vaccines. However, this is not an easy task, and the development of improved influenza vaccines has been a goal for decades. In 2018, the National Institute of Allergy and Infectious Diseases announced a strategic plan to develop a universal influenza vaccine (ie, one that protects against all types/subtypes of influenza) through improving our understanding of the natural history and pathogenesis of influenza infection, more precisely characterizing influenza immunity and correlates of protection, supporting research into the rational design of influenza vaccines, and providing research infrastructure and resources that can be used to develop and assess vaccine candidates [10]. It is likely that improvements will be incremental, and over the past decade we already have seen promising developments. We have evidence of how to improve vaccine immunogenicity and effectiveness, at least for some populations, compared with the immunogenicity and effectiveness of standard dose (15 µg of hemagglutinin per antigen) inactivated influenza vaccines (IIVs). High-dose (60 µg of hemagglutinin per antigen) and MF59-adjuvanted IIV enhance immunogenicity compared with standard-dose vaccine [11] in person ≥65 years of age, and the improved immunogenicity has translated into improved efficacy and effectiveness, including against H3N2 viruses [12, 13]. Recombinant hemagglutinin (45 µg per antigen) also has significantly improved immunogenicity and efficacy compared with standard-dose vaccine in persons ≥50 years of age, with the comparative efficacy having been measured in a year when the H3N2 vaccine strain did not match the circulating virus [14]. The use of higher hemagglutinin dosages has been associated with the generation of more cross-reactive antibodies for H1N1 viruses [15]. The potential value of higher dosages of vaccine or adjuvant use in other age groups or special populations (eg, human immunodeficiency virus-infected and immunocompromised patients) remains to be determined, although enhanced immunogenicity also has been reported [16–19]. A number of other strategies to increase the breadth of vaccine protection are under investigation [20]. Attempts to induce broadly cross-reactive antibodies that recognize well conserved epitopes, such as those in the hemagglutinin stem or the ectodomain of the M2 protein (M2e), are underway. Conserved peptides from internal viral proteins (eg, nucleoprotein) also are being evaluated as universal vaccine candidates that stimulate cellular immunity to influenza. The viral neuraminidase is another potential target for vaccination [21], because higher levels of antibodies to this protein are associated with a lower risk of disease [22]. These strategies are not specific for H3N2 viral infections. A strategy that is worthy of consideration is the addition of a second H3N2 strain to the vaccine, much as is done with some veterinary influenza vaccines and as was done for influenza B in human IIVs several years ago [23]. We need to recognize that the successful development of an effective influenza vaccine does not necessarily mean continued success. Studies of live-attenuated influenza vaccine during development and after licensure consistently showed improved effectiveness over inactivated vaccines in children 2 to 8 years of age, leading the ACIP to recommend its preferential use in this age group [24]. However, subsequent lack of effectiveness in 3 consecutive seasons led to removal of its recommendation for use. A nonpreferential recommendation was reinstated by the ACIP in 2018 after modifications to the vaccine [25]. The use of egg-based technology for much of vaccine production requires adaptation of the virus for growth and can result in the introduction (or loss) of glycosylation sites on the hemagglutinin and adversely affect antibody responses and VE against circulating strains, although the significance of these observations remains to be established [8, 26]. Vaccine production through recombinant expression systems or with virus propagation in mammalian cells instead of eggs may circumvent this issue [13, 27]. Influenza VE surveillance networks help identify when problems arise. As we pursue the development of improved influenza vaccines, studies such as that by Flannery et al [7] serve this purpose and remind us that the clock is ticking. Acknowledgments. We thank Dr. Hana El Sahly for review and helpful comments on the manuscript. Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.