Wild birds that reside in aquatic environments are the major reservoir of avian influenza viruses (AIVs). Since this reservoir of AIVs forms a constant threat for poultry, many countries have engaged in AIV surveillance. More and more commercial enzyme-linked immunosorbent assays (ELISA) are available for serologic surveillance, but these tests are often developed and validated for use in domestic poultry. However, for a correct interpretation of ELISA test results from wild bird sera, more information is needed. In the present study, four ELISA test kits (ID-Vet IDScreen, IDEXX FlockChek AI MultiS-Screen Ab Test Kit, Synbiotics FluDETECTBE, and BioChek AIMSp) were compared for the serologic analysis of 172 serum samples from mallard, mute swan, and Canada goose. Samples were selected based on ID-Vet IDScreen results to obtain an approximately equal number of positive and negative samples. In addition, 92 serum samples from experimentally infected specific-pathogen-free (SPF) chickens and Pekin ducks were included in the tests for validation purposes. Cohen's kappa statistics and Spearman correlation coefficients were calculated for each combination of two tests and for each bird species. Test agreement for mallard sera varied from poor to moderate, while test results for Canada goose and swan sera agreed from fair to almost perfect. The best agreement was obtained with sera from experimentally infected SPF chickens and Pekin ducks. This study shows that some care must be taken before using nucleoprotein ELISAs for the testing of sera from wild birds and that more reliable validation studies should be considered before their use in the serologic surveillance of wild birds.
