Circulation of Influenza Virus & Selection of Vaccine Strains
Initially, only one influenza virus was known. The understanding that different viral strains could be created by spontaneous modification of their protein composition changed the strategies of immunological prevention. A sudden, extensive modification of HA or NA sometimes observed (replacement with the protein from another virus) is called shift. Another milder change, the modification of the sequence of one of these proteins without a complete modification of their antigenic specificity, is called drift. The consequences of these changes are that the composition of the vaccine has to be changed to maintain the efficacy of protection. The delay in production, after the choice of vaccine strains has been made, sometimes make this difficult to achieve. After several years of satisfactory use of the newly developed U.S. Army vaccine, unexpectedly during the 1947 vaccination campaign in the USA, the vaccine was found to be less effective in correctly vaccinated subjects. It was rapidly discovered that the antigenic composition of the seasonal influenza virus had changed and that the epidemic strain was not well covered by either the immunity induced by the vaccine or that induced by a previous infection (drift). This new variant virus was initially called A prime (A′). Although not many strains were isolated during that winter due to a lack of suitable techniques, those that were isolated were all new strains. The virus had different surface antigens, but the same internal proteins and therefore were members of the A strain family; later this family became known as A(H1N1). The main new antigen was one of the glycoproteins important in immunity, that is, HA. It was during this winter that the virus was isolated for the first time in France thereby initiating research on the vaccine in the country.
The problem recurred in 1958 when another seasonal vaccination failure occurred due to the presence of a new virus, which had two surface antigens that were very different from those of the previous virus (shift) and which totally replaced it (Figure 1). At this time, the need for a more precise nomenclature for the influenza strains was recognized; a system that took into consideration the two surface antigens that contribute massively to the antigenic properties of the virus was adopted (Box 1). The new virus was named A(H2N2). The proposed bivalent vaccine thus contained A(H2N2) and B. A dramatic pandemic caused by this new agent extended rapidly to the whole world and resulted in 4–5 million victims .The strains chosen varied depending on the year or the appearance of slightly different variants within both the H2N2 and B influenza viruses subtypes (Figure 1).
It became obvious that it was necessary to modify regularly the composition of the vaccine, and that it was necessary to use vaccine strains selected according to any emerging new virus. The vaccine composition is now adapted annually to integrate viral strains as similar as possible to the epidemic strains. This choice is based on the best estimate of which strains are likely to be predominant because, even with improved surveillance systems, it is impossible to be certain that viral strains will circulate in advance.
The A(H2N2) virus family was in circulation for about 10 years, until the Hong Kong influenza pandemic caused by a new virus started in 1968 in China. As in 1958, the new virus was different from the previous one, but this time only the HA was modified while the NA remained the same. It was later understood that this new strain resulted from the replacement of the HA of the previous strain A(H2N2) by that from an avian virus (H3). The new virus was named A(H3N2). As in 1958, it immediately and totally replaced the previous virus and in 1970, the modified vaccine was still bivalent, but it contained A(H3N2) and B. As previously, the selected viruses were then adapted annually to respond to the epidemiological situation within each subtype.
A different phenomenon occurred again in 1978 when a new virus appeared, but the previous virus was not eliminated so that both subtypes began to co-circulate, either simultaneously or alternatively, depending on the seasons and the regions. The new virus, A(H1N1), that first appeared in Russia, was an analog of the virus that had disappeared in 1958. The vaccine then became trivalent with the strains A(H1N1)/A(H3N2)/B (Figure 1).
In late 2003, then, on a larger scale, in 2007 a virus of avian origin, A(H5N1), began to circulate in wild and domestic fowl and then began to cause sporadic cases in humans. Human vaccines were developed against this avian virus that was extremely virulent and pathogenic for humans, but these vaccines were not used because of the low risk associated with the circulating forms, which were not very contagious in humans and did not give rise to seasonal epidemics. In 2009, another virus, A(H1N1)pdm09, probably from a swine origin began to circulate and caused worrying epidemics in Central and South America. This pandemic influenza strain circulated throughout the world, but was ultimately found to be less dangerous than had originally been suspected when it first appeared. Monovalent pandemic vaccines were developed, although the recommendations for use and the use of these vaccines varied between regions and countries. However, this strain, A(H1N1)pdm09, has since become responsible for seasonal epidemics and has now replaced the previous A(H1N1) strain. The variant is now present in the most recent trivalent vaccines, which therefore contain A(H1N1)pdm09, A(H3N2) and B.
Although lower degrees of effectiveness have been observed in the elderly and some high-risk subjects, this annual updating of the vaccine strains has been shown to be generally effective since 1987 except when mismatching occurs. If the virus drifts in nature after the choice of strains (in February for the Northern hemisphere), then the vaccine is not as efficient because of the specificity of the immune response. This situation called mismatch cannot always been avoided even with efficient international surveillance systems; between 1987 and 1997, 23 of the 30 recommended strains were similar to the circulating virus during the following winter. In 5 of the 10 seasons, a perfect match for all three strains was observed. When a good match was not observed for one of the strains, the efficacy of the vaccine was reduced depending on the degree of mismatch.
Since 1973, WHO has issued recommendations for the composition of influenza vaccines based on the results of surveillance systems, and these recommendations are always followed by the manufacturers. Since 1999, there are two sets of recommendations every year, one for the northern hemisphere in February and the other for the southern hemisphere in September; these recommendations are issued sufficiently early to allow timely production of vaccine conform with the recommendations (Figure 2). The evaluation of risks for persistence and dissemination of new strains is based on a careful analysis of epidemiological data; antigenic identification of strains, pathogenic potential and transmissibility. The selection process is highly coordinated and involves continual year-round integration of viral data and epidemiological information from the six WHO Collaborating Centers (WHOCCs: Atlanta, Beijing, London, Melbourne, Memphis, Tokyo), through antigenic and genetic characterization of viruses as part of the process of selecting suitable candidate vaccine viruses. The preparation of suitable reassortants and corresponding reagents for vaccine standardization is the responsibility of the WHO Essential Regulatory Laboratories. The WHO recommendations for the composition of the trivalent inactivated influenza vaccine for the northern hemisphere from 2000 to 2013 are summarized in Figure 3. A mathematical model has recently been developed to map the antigenic variations and to have a better understanding of the direction of the drift.
Summary of the annual process of development, manufacturing and distribution of influenza vaccines in the northern hemisphere.
Data taken from .
Expert Rev Vaccines. 2013;12(9):1085-1094. © 2013 Expert Reviews Ltd.