Age-Specific Mortality During the 1918–19 Influenza Pandemic and Possible Relationship to the 1889–92 Influenza Pandemic

To theEditor—The uniquely increased mortality rate among adults aged 20–40 years during the 1918–1919 influenza pandemic has never been adequately explained. Viboud et al present extensive data from Kentucky death registrations indicating that peak mortality occurred between 24 and 26 years of age and concluded that a combination of risk factors determined mortality [1]. Subsequent work by Gagnon et al showed distinct modal mortality at 28 years of age from 8 Canadian and 3 US civilian data sources [2]. The actual age of peak mortality may be important as some have suggested that infection while a young child during the previous influenza pandemic of 1889–1892 (thought to have been due to influenza A[H3Nx]) may have predisposed these individuals to a more severe outcome approximately 28 years later during the subsequent 1918–1919 influenza A(H1N1) pandemic [2, 3]. We sought similar mortality information from civilian and military populations in New Zealand and Australia to determine whether the pattern seen was the same as that observed in the United States and Canada. Death registration information for 4813 people who died during the 1918–1919 pandemic was available in New Zealand from an earlier historical study [4]. Data from Auckland City have subsequently been lost, but data from all other parts of New Zealand were included. It is also likely that deaths among the indigenous Māori had been underreported. Denominator data were calculated from the 1916 National Census with Statistics New Zealand estimates for the total population in 1918 [5]. We also considered influenza mortality data for 1239 soldiers in the Australian Army (with deaths mainly in Europe and the Middle East), aspects of which have previously been reported [6]. Age-specific mortality per 10 000 population by single year due to pneumonia and influenza during the 1918–1919 pandemic among civilians in New Zealand, stratified by sex (A), and among Australian soldiers (>99% male) predominately while in Europe and the Middle East (B). Multiple risk factors contributed to mortality in 1918, including age, sex, indigenous group, urban/rural status, time since recruitment into the military, and exposure to bacterial pathogens, as most deaths were caused by secondary bacterial pneumonia (due to pathogens such as Streptococcus pneumoniae and Staphylococcus aureus) [6, 8–10]. Since there is no intrinsic reason to think that adults separated by few years in the range from 20 to 40 years of age would react differently to a severe viral respiratory infection, it seems possible that the otherwise inexplicable peak age of mortality near 28 years of age might be evidence of dysfunctional immunity to a previous influenza virus infection. Potential mechanisms to explain enhanced mortality caused by previous influenza virus infection include nonneutralizing antibodies and/or dysfunctional CD8+ lymphocyte responses triggering excessive cytokine production [2, 3]. Persons aged 28 years in 1918 would have been members of the 1890 birth cohort, whose first influenza virus infection would have almost certainly been by the 1889–1892 pandemic virus. Since not all children are infected with influenza virus in their first year of age, some broadening of the peak mortality rate would not invalidate the hypothesis that infection during one pandemic predisposed to severe outcomes during a subsequent influenza pandemic. Besides the unique age-associated mortality pattern, the 1918–1919 pandemic caused very heterogeneous effects in what otherwise would have been expected to be identical populations. Although some populations experienced catastrophic mortality, others had a 10–50 times lower case-fatality proportion while experiencing similar infection rates. There is no simple explanation for the observed epidemiology as the reconstructed 1918 virus has no unique genomic features and is similar to the pandemic virus of 2009 [11]. Mortality risk is a composite of pathogen virulence, host susceptibility, and environmental factors. The peak age-specific mortality rate near 28 years of age in both the United States and Canada, as well as in 2 additional populations (from New Zealand and Australia), seems to provide some evidence for the dominance of host factors and possibly reflects a common exposure, such as early childhood infection during the 1889–1892 influenza pandemic. Disclaimer. The opinions expressed are those of the authors and do not necessarily reflect those of the Australian Defence Force. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.