The hemagglutination inhibition (HI) assay is commonly used to assess the humoral immune response against influenza A viruses (IAV). However, the microneutralization (MN) assay has been reported to have higher sensitivity when testing sera from humans and other species. Our objective was to determine the agreement between MN and HI assays and compare the proportion of positive samples detected by both methods in sera of mallards primary infected with the A/mallard/MN/Sg-000169/ 2007 (H3N8) virus and subsequently inoculated with homosubtypic or heterosubtypic IAV. Overall, we found poor to fair agreement (prevalence-adjusted bias-adjusted kappa [PABAK], 0.03–0.35) between MN and HI assays in serum samples collected 2 weeks after H3N8 inoculation; the observed agreement increased to moderate or substantial in samples collected 4 to 5 weeks postinoculation (WPI) (PABAK, 0.52–0.75). The MN assay detected a higher proportion of positive samples compared with HI assays in serum samples collected 2 WPI (P = 0.01). This difference was not observed in samples collected 4 WPI. Also, a boosting effect in MN and HI titers was observed when birds were subsequently inoculated with IAV within the same H3 clade. This effect was not observed when birds were challenged with viruses that belong to a different HA clade.

In summary, the agreement between assays varies depending on the postinfection sample collection time point and the similarity between the antigens used for the assays. Additionally, subsequent exposure of ducks to homosubtypic or heterosubtypic strains might affect the observed agreement.