Comparison of Optimized Serological Assays for Detecting and Quantifying Chicken Antibodies to Avian Influenza Virus

Surveillance programs for low pathogenic avian influenza virus frequently rely on serological tests to determine whether poultry flocks have been infected with the virus. Poultry sera that are positive by type A influenza screening tests are frequently tested at reference laboratories by an hemagglutination inhibition (HI) assay as a subtype-specific confirmatory test. Because of the potential to become highly pathogenic, H5 and H7 are the most common targets. An alternative to HI assay is the microneutralization (MN) assay, which is widely used for humans, wildlife, and mammalian research animals because of its sensitivity. Subtype-specific assays are also used to monitor response to vaccines. Here the observable endpoint of the HI assay was compared with the MN assay using two methods to read the MN assay results, hemagglutination assay (MN-HA), and observation of cytopathic effects (MN-CPEs). Rates of positivity for individual samples by HI assay, MN-HA, and MN-CPEs were 66.7% (32/48), 89.6% (43/48), and 87.5% (42/ 48), respectively. Agreement was low with a Cohen's kappa statistic (κ) of 0.377 for HI vs. MN-HA and 0.333 for HI vs. MN-CPEs, where discrepancies were almost exclusively where samples were negative by HI assay and positive by MN-HA or MN-CPEs (κ = 0.897 for MN-HA vs. MN-CPEs). Titer values were also significantly higher with both MN-HA and MN-CPEs than with the HI assay (P < 0.0001 and P = 0.150, respectively). Although not significant, there was a trend for MN-HA to produce higher titers and detect more positive samples than MN-CPEs.