When a new influenza variant emerges with significant mutations to its surface proteins, the default concern is that existing immunity — whether from prior infection or vaccination — will fail to protect. The 2025 emergence of H3N2 subclade K (J.2.4.1) triggered exactly that concern, yet the actual 2025–26 season unfolded without unusual severity. Understanding why matters for how we interpret antigenic drift data and calibrate public health responses.
The study enrolled 243 individuals from the UK and Norway spanning an extraordinary age range — 1 to 105 years — and tested sera across four distinct cohorts using two established assays: microneutralisation (MN) and haemagglutination inhibition (HAI). Antigenic cartography using ferret antisera did confirm meaningful antigenic drift between subclade K viruses and the J.2-based vaccine strain used in the 2025/26 Northern Hemisphere egg-derived trivalent inactivated vaccine. Despite this, vaccination significantly boosted neutralising and inhibitory antibody titres against subclade K viruses (p < 0.001). Critically, serological profiling stratified by age — including children under five, adolescents, young adults, and adults over 60 — revealed pre-existing cross-reactive antibody responses against the emergent subclade even before vaccination.
This finding illuminates a recurring tension in influenza surveillance: ferret-based antigenic cartography, the standard tool for predicting immune escape, may overestimate the functional significance of drift when diverse human immune histories are accounted for. Decades of overlapping H3N2 exposures appear to deposit a residual layer of cross-reactive immunity that animal models cannot replicate. That said, this study is observational and serological in design — it does not directly measure clinical protection or vaccine effectiveness against symptomatic disease. The cohort, while age-diverse, is limited to two Northern European populations, and cross-reactive titres may not translate uniformly across globally distinct exposure histories. The findings are nonetheless meaningful: they suggest antigenic cartography alone is insufficient to forecast real-world season severity, and that population-level immunological breadth deserves more weight in influenza risk modeling.