Childhood flu infection leaves lasting immune imprint

Why the flu you caught as a child could determine your risk decades later. and what it means for future outbreaks and vaccination strategies.

Study: Childhood immune imprinting shapes cohort and period influenza mortality. Image credit: PeopleImages/Shutterstock.com

A study in Science Advances suggests that population-level patterns consistent with strain-specific immunologic imprinting by childhood influenza infection are associated with differences in lifelong mortality risk from influenza.

Influenza virus antigens

The influenza A virus (IAV) has two surface antigens, hemagglutinin (HA) and neuraminidase (NA). These are the dominant antigens targeted by host antibodies and are key determinants of individual susceptibility to the virus.

Antigenic drift refers to the change in antigenicity caused by accumulated mutations in these antigens, which contributes to immune evasion and recurrent infections. Antigenic shift, in contrast, refers to the reassortment of these antigens to create novel combinations that lead to the emergence of a new IAV subtype, causing pandemics.

The existing literature shows that childhood influenza infection shapes the immune response to subsequent influenza infections. The “original antigenic sin” evokes the highest antibody titer. Moreover, these individuals tend to exhibit protection patterns consistent with reduced risk of infection by seasonal influenza or novel avian influenza viruses with the same HA phylogeny as the first childhood virus strain.

They also exhibit antigenic seniority, persistently producing higher antibody levels against the first childhood strain than against later strains.

Historical trends

The first H1N1 IAV pandemic of 1918-1920, the so-called “Spanish flu”, was followed by three phases of antigenic evolution, ending with the “A-Prime” variant circulating during 1946-1947.

In 1957, H2N2 underwent a major antigenic shift, displacing most variants and triggering another pandemic. It was displaced in 1968 by H3N2, marking the beginning of the group 2 IAV virus pandemics. H3N2 viruses have shown the most rapid antigenic drift, causing large seasonal outbreaks and higher mortality. H1N1 was reintroduced into human populations in 1977 and remains in circulation alongside H3N2.

In 2009, H1N1pdm09 replaced the seasonal circulating H1N1 strains globally. It shares HA epitopes most closely with the earliest 1918 H1N1 strains.

Overall influenza mortality

The current study explored mortality rates by age and period in US birth cohorts between 1860 and 2020, by single-year age and by single-season, covering seasons from 1968-1969 to 2020-2021. The aim was to estimate model-based differences in seasonal influenza mortality risk in cohorts likely imprinted in childhood, based on circulating strains.

According to the analysis, influenza mortality decreased during 1968-2009 (with the emergence of H3N2). It then increased, suggesting higher mortality during H1N1pdm09 seasons than earlier H1N1 outbreaks. After 2010, overall influenza mortality rose again. In 2020-2021, mortality dropped steeply due to COVID-19-related public health measures.

The increased mortality post-2010 might reflect greater severity, improved diagnosis, changes in death certificate coding, or the aging of the population. During this period, almost half the deaths were in older adults (85 and over).

Single-year age analyses

Age-stratified analyses revealed two patterns. As expected, younger cohorts had lower mortality across all influenza subtypes.

The results suggest that H1N1 seasons prior to 2009 were associated with approximately 97 % lower mortality rates compared to H3N2 seasons, based on model-adjusted relative risk estimates. From 2009 onwards, H1N1pdm09 seasons had higher mortality rates than H1N1 seasons but not H3N2 seasons. Prior to 2009, mortality during H1N1 seasons was similar between H1N1 and H3N2 cohorts. During the H1N1pdm09 seasons, H1N1 cohorts had lower mortality rates than H3N2-imprinted cohorts.

During the H1N1pdm09 seasons, mortality was lower than expected in the 1940-1944 cohorts, who were likely imprinted by antigenically similar early H1N1 strains, consistent with enhanced protection. A possible trade-off is suggested by higher mortality rates, even after age adjustment, in older H1N1-imprinted cohorts compared with younger cohorts during H3N2 seasons.

Cohorts imprinted by later H1N1 variants had higher mortality rates, suggesting a progressive weakening of the protective effects of imprinting as antigenic changes persisted.

The protective effect of H1N1 imprinting appears stronger and more consistent than that of other strains, while protection against other subtypes is more limited or variable. H2N2-imprinted cohorts had higher-than-expected mortality for their age, suggesting weaker protection at the HA level, though this remains an interpretation rather than a definitive mechanism.

The mortality among H3N2-imprinted cohorts in H3N2 seasons was surpassed only by pre-1918 cohorts (who were not imprinted by the 1918-2020 H1N1 strains). Several H1N1 and H2N2 cohorts had approximately 20 % lower mortality from H3N2 during H3N2 seasons, as reflected in regression-adjusted estimates rather than uniform effects across all cohorts.

The authors offer possible explanations, including a weaker antibody response to group 2 HA stalk antigens (such as H3N2); fewer unique H3N2 exposures in childhood as H1N1 began to cocirculate within a decade, leading to weaker imprinting; and more rapid antigenic evolution.

As a negative control, the authors used all-cause and pancreatic cancer mortality, confirming that the observed trajectories were specific to influenza mortality.

Future influenza mortality projections

The authors project a higher mortality risk for aging H3N2- and H2N2-imprinted cohorts in future H1N1pdm09 seasons, that is, for much of their life. This is likely to persist as long as H1N1pdm09 is in circulation. Moreover, the authors suggest that if avian H5N1 viruses begin to circulate among humans, H3N2 cohorts may suffer greater mortality than the older H1N1 cohorts.

This strain-dependent protection emphasizes the importance of seasonal influenza vaccination, “which can provide protection against strains that are mismatched to strains that individuals are imprinted with in childhood.” Future studies are needed to identify differences between vaccine-induced and infection-related imprinting in infants, who are often vaccinated before their first infection.

If they are found to be equivalent, “our results suggest that vaccinating naïve children in such a way that ensures strong H1N1 response, while maintaining protection against other seasonal strains, may provide broader protection throughout their lives.” Meanwhile, a universal influenza vaccine remains sought after.

Limitations

The study has several important limitations. Imprinting was inferred from circulating strains and birth year rather than directly measured, and the analysis assumes a constant risk of influenza exposure across seasons. The identification of dominant strains relied on viral specimen testing, which may underrepresent less severe variants, while the effects of neuraminidase (NA) and hemagglutinin (HA) imprinting could not be separated.

In addition, using mortality data captures only the most severe cases and excludes the milder infections that account for the majority of the influenza burden.

The findings may also be influenced by potential misclassification or variation in death certificate coding, and the analysis does not account for individual-level factors such as comorbidities or healthcare-seeking behavior.

Download your PDF copy by clicking here.