Seasonal influenza vaccination and the 2009 pandemic.
G. Mercer,H. Kelly
DOI: https://doi.org/10.1093/cid/cir035
IF: 20.999
2011-03-15
Clinical Infectious Diseases
Abstract:To the Editor—We were pleased to see the editorial by Glezen [1], which outlined a hypothesis to explain the observation made by Cowling and colleagues [2] in the same issue of the Journal that receipt of the seasonal trivalent influenza vaccine (TIV) appeared to increase the risk of infection with pandemic influenza 2009 (pH1N1). This observation has been made previously, but the effect size has varied. Four studies from Canada reported odds ratios (OR) for pH1N1 infection following receipt of seasonal TIV in the range 1.4–2.5 (confidence interval [CI] 5 1.03–2.74) [3], whereas the Hong Kong investigators reported a nonsignificant increased risk, with an OR 5 1.11 (.54–2.26) [2]. Investigators in Australia reported an OR 5 .97 (.60–1.56) [4]. The timing of the pandemic compared to the previous influenza season provides a possible explanation of these apparently discrepant findings. Although not yet a fully accepted phenomenon, there is a growing body of evidence that infection with one strain of influenza gives a ‘‘short-lived strain transcending immunity’’ [5]. This concept of heterosubtypic temporary immunity is a necessary inclusion for models to reproduce the observed influenza phylogenetic structures [5, 6] and has recently been used by us to model and explain the apparent increased risk of infection with pH1N1 following receipt of seasonal TIV [7, 8]. There is evidence for temporary immunity from human studies during the 1957 H2N2 pandemic [9] and the reemergence of H1N1 in 1977–1978 [10]. Temporary heterosubtypic immunity is thought to last 3–6 months, but the evidence for its duration is derived from the modeling literature [5] and needs additional investigation. During the first wave of the pandemic (April–August 2009, before pandemic vaccine was available), an individual with a recent seasonal influenza infection would have acquired some degree of temporary immunity to pH1N1 and would therefore have had a lower chance of pH1N1 infection. Individuals who had received the 2009 seasonal TIV had a lower probability of infection with the seasonal strain [3] and hence had an apparent higher risk of pH1N1 infection, having foregone the protection against pH1N1 infection resulting from seasonal influenza infection. This risk varied with seasonal vaccination coverage in the population. In both Canada and Hong Kong, the previous seasonal influenza epidemic occurred between January and April 2009. Circulation of pH1N1 occurred weeks to months after circulation of seasonal influenza, allowing for temporary immunity to pH1N1 to have developed, and persisted, in persons infected with seasonal influenza. In contrast, in the Southern hemisphere, the previous influenza season was between June and October 2008, so that any temporary immunity from seasonal influenza infection had waned by the time pH1N1 was first detected in April 2009. Cowling showed that seasonal vaccine did not protect against pH1N1 infection [2], and, in the absence of temporary immunity, this was what was found in Australia [4]. In conclusion, the timing of the first appearance of the pandemic strain relative to the previous seasonal influenza epidemic, combined with temporary heterosubtypic immunity, give a plausible explanation for the findings of various estimates of the risk of pH1N1 infection following receipt of seasonal TIV around the world. Of interest is whether similar effects were seen in other countries, such as the United Kingdom, that had a sizable first pandemic wave.