Seasonal influenza remains a major global health challenge, yet the factors governing its evolution in human populations are difficult to disentangle. The COVID-19 pandemic created a unique natural experiment. The widespread implementation of non-pharmaceutical interventions (NPIs)—such as masking, travel restrictions, and school closures—abruptly reconfigured transmission networks for respiratory viruses (Federici et al., 2018; GBD 2017 Influenza Collaborators, 2019; Madhusudanan et al., 2023). This perturbation offers a rare opportunity to examine how shifts in transmission intensity and host immunity shape viral evolution, particularly for influenza B virus (IBV), a significant contributor to seasonal burden and vaccine mismatch risk (Liu et al., 2022; Smith et al., 2004). With the reported functional disappearance of the Yamagata lineage, understanding the evolutionary drivers of the dominant Victoria lineage (IBV-V) has become increasingly urgent (Marchi et al., 2024).

Population-genetic theory predicts that sharp reductions in incidence and effective population size slow adaptive evolution (Coffin and Swanstrom, 2013; Lynch, 2020). For influenza, such effects should be most pronounced in hemagglutinin (HA)—the primary target of host immunity—relative to the more conserved neuraminidase (NA) and nucleoprotein (NP) (Zhao et al., 2024). Consequently, stringent NPIs were expected to decelerate HA evolution, with subsequent relaxation permitting re-acceleration. The magnitude and timing of these shifts, however, likely varied across countries with divergent policies, vaccination histories, and demographics (Aruffo et al., 2022).

National responses to COVID-19 differed markedly. China enforced sustained, centralized containment; the United States adopted a decentralized and temporally heterogeneous strategy, characterized by stringent measures in spring 2020, followed by partial and localized measures in late 2020–early 2021, with most state-level interventions lifted by mid-to-late 2021; and Australia combined strict border controls with phased domestic relaxation (Dergiades et al., 2022; Hale et al., 2021; Wei et al., 2022). These policy disparities-contextually comparable via indices such as the Oxford COVID-19 Government Response Tracker (OxCGRT) - provide a powerful comparative framework (Drummond and Hasnine, 2023; Leiva et al., 2023). While studies have reported suppressed influenza activity and altered phylogenetic patterns during the pandemic (Gupta et al., 2022; Hammond et al., 2021; Olsen et al., 2020), a systematic, cross-national analysis linking policy intensity to evolutionary metrics of IBV-V remains limited.

Here, we test the hypothesis that COVID-19 NPIs were associated with a detectable deceleration in IBV-V HA evolution, followed by re-acceleration upon relaxation, and that the magnitude of these shifts correlates with national policy stringency. We analyzed curated IBV-V HA, NA, and NP sequences from GISAID (2016–2024) for China, the United States, and Australia, stratifying pre-pandemic, intervention, and post-relaxation periods (Bernasconi et al., 2021; Shu and McCauley, 2017). To mitigate sampling bias, we applied stringent quality controls and sensitivity analyses (including year- and country-stratified random sampling) (Lindegren et al., 2012). We then estimated genetic divergence and rate of genetic distance change using standard phylogenetic workflows (e.g., multiple-sequence alignment, model selection, time-scaled phylogenies/root-to-tip regression) and assessed associations with policy intensity metrics (Abraham et al., 2018; Takeuchi and Kawashima, 2023).

Our study makes three contributions. First, by focusing on a single IBV lineage across multiple countries, we isolate lineage-specific evolutionary signals from general suppression effects. Second, by comparing HA to NA and NP, we test whether observed shifts align with theoretical predictions of antigen-specific selection pressures. Third, we provide a policy-aware molecular surveillance framework to inform vaccine-strain selection and preparedness planning as interventions evolve (Timmis et al., 2017).

In summary, the pandemic provides a unique vantage point to understand how macro-level interventions influence viral evolution. By leveraging national policy differences, our work links transmission ecology to evolutionary outcomes in IBV-V and underscores the need to integrate genetic surveillance with policy metrics to anticipate and mitigate future antigenic drift and vaccine mismatch.

To assess the impact of COVID-19 non-pharmaceutical interventions (NPIs) on IBV-Victoria evolution, we stratified the study timeline (2016–2024) into three distinct epidemiological periods: Pre-pandemic (2016–2019), Pandemic/NPI (2020–2021), and Post-relaxation (2022–2024). All subsequent analyses and visualizations are framed within this temporal framework.

The COVID-19 pandemic and associated non-pharmaceutical interventions (NPIs) created a global natural experiment for respiratory virus evolution. Our analysis of IBV-Victoria sequences from 2016 to 2024 across three countries reveals a gene-specific modulation of evolutionary dynamics. HA genetic divergence showed a pattern of deceleration during the peak intervention period (2020–2021) followed by an increase thereafter, whereas NA and NP exhibited comparatively smaller and less consistent changes. Pairwise between-period contrasts further indicate that HA exhibits the largest Pre-NPI and Post-NPI differences within countries, with NA/NP differences modest.

These observations are consistent with population-genetic expectations. The sharp decline in HA divergence is consistent with a scenario of reduced effective population size under stringent NPIs, which attenuates positive selection and amplifies genetic drift. As the primary antigenic target, HA is under intense immune pressure during normal transmission; NPIs may have relaxed this pressure through the suppression of transmission chains. The subsequent increase post-2022 coincides with the restoration of transmission opportunities and immune-driven selection. The stark contrast with the much more subdued temporal profiles of NA and NP-genes subject to stronger structural constraints and slower antigenic drift-underscores that the pandemic’s impact was not a genome-wide effect, but a targeted modulation of the most dynamically evolving antigenic component.

While the direction of HA’s response-downturn followed by rebound-was consistent across China, the United States, and Australia, the magnitude and timing of these shifts offer nuanced insights. The pronounced and synchronized “V-shaped” trajectory in China and Australia is broadly consistent with the implementation of more uniform and stringent NPIs, which created strong, population-wide transmission suppression. In contrast, the earlier and more volatile pattern in the United States is consistent with its decentralized and heterogeneous policy landscape, resulting in a “patchwork” of transmission dynamics that afforded the virus more opportunities to evolve. Although our ecological design precludes definitive causal attribution, these cross-country patterns are consistent with the hypothesis that the intensity and consistency of transmission suppression, rather than the mere presence of interventions may be important factors associated with the tempo of viral antigenic evolution.

Our period-based analysis simplifies complex national timelines. The inclusion of 2022 in the post-relaxation phase captures early recovery while acknowledging its transitional nature. Other factors, such as pandemic-related shifts in influenza surveillance intensity or post-pandemic immunity landscapes, could have contributed to the observed genetic patterns. While our three-gene framework helps isolate antigen-specific signals, future work integrating epidemiological and mobility data is needed to separate these concurrent drivers.

Our findings have practical implications for public health. First, policy-aware genomic surveillance-interpreting HA divergence in the context of intervention timing-may improve the anticipation of antigenic drift during future public-health transitions. Second, the post-2022 rebound in HA underscores a potential increase in antigenic-drift and vaccine-mismatch risk during reopening, arguing for enhanced monitoring and timely candidate vaccine virus evaluation. Third, routine gene-resolved analytics that benchmark HA against NA/NP can help separate antigen-specific selection from background genomic drift, refining early-warning signals.

This study has several limitations. Despite stratified sampling, heterogeneity in sequence submissions may introduce bias. We note that some country-year strata contain fewer sequences due to limited availability in GISAID (e.g., Australia-2020), which may increase sampling variability; therefore, year-specific estimates from sparse strata should be interpreted cautiously, with emphasis on multi-year and period-level patterns. Our use of linear regression on per-sequence genetic distance values provides a descriptive metric of temporal trend but differs from model-based substitution-rate estimates. Furthermore, we did not jointly analyze incidence, NPI stringency, or mobility data; thus, policy-evolution links remain associative. Additionally, the per-sequence genetic distance values used in our statistical comparisons are not fully independent, as individual sequences contribute to multiple pairwise calculations. This non-independence may affect the precision of p-values. However, our primary conclusions—the distinct response pattern of HA compared to NA/NP, and its consistency across countries—are based on large and consistent effect sizes, which remain robust to this limitation. Moreover, our definition of epidemiological periods, while aligned with major policy transitions, represents a simplification of complex, country-specific timelines. Finally, gene-level analyses cannot localize selection to specific epitopes, warranting future site-level studies for mechanistic insight.

In conclusion, by comparing three countries with divergent intervention histories, we provide evidence consistent with the idea that macro-level transmission shifts during the COVID-19 era may have left a detectable, gene-specific molecular signature on IBV-Victoria. These results underscore the value of integrated genetic-epidemiological surveillance to anticipate influenza evolution and strengthen preparedness for future epidemics.