Evaluation of avian influenza serologic and virologic diagnostic methods in wild Anseriformes and Charadriiformes

Evaluation of avian influenza virus (AIV) diagnostic methods, including a nucleoprotein (NP) competitive enzyme-linked immunosorbent assay (c-ELISA), hemagglutination inhibition (HI) test, type A real-time reverse transcription polymerase chain reaction (RRT-PCR), and embryonating chicken egg (ECE) virus isolation (VI), suggested validity of these tests in wild birds comparable to that reported in poultry. This was determined by analyzing the results from experimental inoculation of three species of wild birds with a low-pathogenicity AIV and from field surveillance data. The NP c-ELISA in a high-AIV prevalence setting had 100% diagnostic sensitivity (Se; 95% confidence interval [CI]: 81.5%-100%) and 91% diagnostic specificity (Sp; 95% CI: 70.8%-98.9%) in negative controls compared with the RRT-PCR. In low-AIV prevalence flocks using a > 60% inhibition positivity threshold, relative to the HI test, c-ELISA performed with 90.5% Se (95% CI: 86.2%-93.8%) and 41.2% Sp (95% CI: 38.1%-44.5%). Assessment of HI suggests a titer > or = 8 is a positive test result in wild-bird sera, and using this titer had 83.3% Se (95% CI: 58.6%-96.4%) in experimentally infected birds. The RRT-PCR diagnostic performance compared with VI in cloacal swabs varied over 2-6 days postinoculation, having high Se (83.3%-100%) and Sp (94.1%-100%) with substantial agreement (kappa = 0.8). The cycle thresholds (C(t)) for the RRT-PCR of C(t) < 37 for positivity and C(t) = 37-40 as indeterminate were found to be valid for the species included in this study. In view of the interpretative diagnostic difficulties in heterogeneous populations of wild birds, this evaluation in three species of wild birds and in surveillance data should provide greater confidence in the application of these methods routinely used in poultry.