Antigenic drift and immunity gap explain reduction in protective responses against influenza A(H1N1)pdm09 and A(H3N2) viruses during the COVID-19 pandemic: a cross-sectional study of human sera collected in 2019, 2021, 2022, and 2023

Abstract

Background: Non-pharmaceutical interventions implemented during the COVID-19 pandemic resulted in a marked reduction in influenza infections globally. The absence of influenza has raised concerns of waning immunity, and potentially more severe influenza seasons after the pandemic.

Methods: To evaluate immunity towards influenza post-COVID-19 pandemic we have assessed influenza A epidemics in Norway from October 2016 to June 2023 and measured antibodies against circulating strains of influenza A(H1N1)pdm09 and A(H3N2) in different age groups by hemagglutination inhibition (HAI) assays in a total of 3364 serum samples collected in 2019, 2021, 2022 and 2023.

Results: Influenza epidemics in Norway from October 2016 until June 2023 were predominately influenza As, with a mixture of A(H1N1)pdm09 and A(H3N2) subtype predominance. We did not observe higher numbers of infections during the influenza epidemics following the COVID-19 pandemic than in pre-COVID-19 seasons. Frequencies of protective HAI titers against A(H1N1)pdm09 and A(H3N2) viruses were reduced in sera collected in 2021 and 2022, compared to sera collected in 2019. The reduction could, however, largely be explained by antigenic drift of new virus strains, as protective HAI titers remained stable against the same strain from one season to the next. However, we observed the development of an immunity gap in the youngest children during the pandemic which resulted in a prominent reduction in HAI titers against A(H1N1)pdm09 in 2021 and 2022. The immunity gap was partially closed in sera collected in 2023 following the A(H1N1)pdm09-dominated influenza seasons of 2022/2023. During the 2022/2023 epidemic, drift variants of A(H1N1)pdm09 belonging to the 5a.2a.1 clade emerged, and pre-season HAI titers were significantly lower against this clade compared to the ancestral 5a.2 clade.

Conclusion: The observed reduction in protective antibodies against A(H1N1)pdm09 and A(H3N2) viruses post COVID-19 is best explained by antigenic drift of emerging viruses, and not waning of antibody responses in the general population. However, the absence of influenza during the pandemic resulted in an immunity gap in the youngest children. While this immunity gap was partially closed following the 2022/2023 influenza season, children with elevated risk of severe infection should be prioritized for vaccination.

Keywords: A(H1N1)pdm09; A(H3N2); Antibody; Antigenic drift; Immunity gap; Influenza; Serology.

Fig. 1

Influenza epidemics in Norway from October 2016 to May 2023. A The number of laboratory confirmed influenza A and B infections were extracted from the Norwegian influenza surveillance data generated by NIPH and the regional laboratories. B The proportions of A(H1N1)pdm09 and A(H3N2) among influenza A infections were estimated from patient samples tested against both subtypes, and these weekly frequencies extrapolated to the total number of detected influenza A infections. Each influenza season is indicated with a start in week 40 and end in week 25 the following year

Fig. 2

Protective antibodies against influenza A in sera from 2019, 2021, 2022 and 2023. Residual sera collected in August 2019 (n = 1054), 2021 (n = 657), 2022 (n = 1197) and 2023 (n = 456) were evaluated in HAI assay against A H1N1pdm09 strains A/Michigan/45/2015, A/Brisbane/2/2018 or A/Victoria/2570/2019 or B H3N2 strains A/Singapore/19/2016, A/Kansas/14/2017, A/Cambodia/e0826360/2020 or A/Darwin/9/2021. Sera were considered protective if HAI titers were ≥ 40, and the percentage of protective-titre sera plotted in different age groups

Fig. 3

HAI titers against H1N1pdm09 A/Victoria/2570/2019 in sera from 2021 and 2022 in different age groups. A Reverse cumulative plots were generated from the HAI titers against A/Victoria/2570/19 from 2021 and 2022 for the age groups 0–4 years, 5–14 years, 15–24 years, 25–59 years and 60 + years. The dotted line indicates 50% protective HAI titer of 40. B Number of detected influenza A infections per 100.000 individuals in the different age groups for the period week 40 2022 to week 22 2023. C Vaccine coverage in the general population obtained from the Norwegian Immunization Registry SYSVAK for the different age groups. D Reverse cumulative plots of HAI titers in a panel of 119 sera collected in 2021 and first tested against A/Victoria/2570/2019 in 2021 and repeat tested in 2022 to verify reproducibility. A Significant differences were calculated by a two-tailed Mann–Whitney test. **p < 0.01, ***p < 0.001 and****p < 0.0001, ns = no significant difference

Fig. 4

HAI titers against A/Darwin/9/2021(H3N2) in sera from 2021 and 2022 in different age groups. A Reverse cumulative plots were generated from the HAI titers against A/Darwin/9/2021 from 2021 and 2022 for the age groups 0–4 years, 5–14 years, 15–24 years, 25–59 years and 60 + years. The dotted line indicates 50% protective HAI titer of 40. B Number of detected influenza A infections per 100.000 individuals in the different age groups were extracted from the Norwegian Laboratory Database for the period week 1 2022 to week 26 2022. C Reverse cumulative plots of HAI titers in 119 sera collected in 2021 from one reference lab tested against A/Darwin/9/2021 in both 2021 and 2022 to verify reproducibility. Significant differences were calculated by a two-tailed Mann–Whitney test. ***p < 0.001 and ns = no significant difference

Fig. 5

Reduced HAI titers against A(H1N1)pdm09 clade 6B.1A.5a.2a.1 in 2022. A Maximum parsimony tree of HA sequences of Norwegian A(H1N1)pdm09 strains from the 2022/2023 influenza season, including reference strains and vaccine strains for the southern and northern hemisphere. B Residual serum samples from August 2022 with HAI titer of ≥ 160 against A/Victoria/2570/2019 (clade 6B.1A.5a.2) were evaluated in an HAI assay against the A(H1N1)pdm09 clade 6B.1A.5a.2a.1 strain A/Norway/25089/2022. B Data presented is geometric mean with error bars representing 95% confidence interval. Significance was determined using a Wilcoxon matched-paired signed rank test. **p < 0.01, ***p < 0.001 and ****p < 0.0001