Serological Surveillance of Wild Waterfowl in Northern Australia for Avian Influenza Virus Shows Variations in Prevalence and a Cyclical Periodicity of Infection

The virological surveillance of 3582 wild waterfowl in northern Australia from 2004 to 2009 for avian influenza virus (AIV) found an apparent prevalence (AP) of 1% (31 of 2989 cloacal swabs; 95% CI: 0.71%-1.47%) using a Taqman Type A real-time reverse transcription polymerase chain reaction test and no viral isolations from 593 swabs tested by the embryonating chicken egg culture method. From serological testing using a nucleoprotein competitive enzyme-linked immunosorbent assay for AIV antibody, 1131 of 3645 sera had >= 40% inhibition, indicating an apparent seroprevalence of 31% (95% CI: 29.5%-32.6%). This value suggests that the low AP from virological testing does not reflect the dynamics of AIV infection in these populations. Spatiotemporal and species variations in seroprevalence were found at wetland sampling sites, with consistently higher values at Kununurra in Western Australia (AP 5 39%, 95% CI: 36.9%-41.4%) compared to other locations. At Kununurra, seroprevalence values had a two-year cyclical periodicity and suggest this location is a hotspot of AIV activity. From hemagglutination inhibition (HI) testing using multiple subtype antigens, the highest AP of HI reactions were to H6 and H5 subtypes. The phenomenon of cyclic periodicity in NP seroprevalence at Kununurra is hypothesized as being related to the prevalent H6 subtype that may have either become predominant or cycled back into a mostly AIV naive flock. The inclusion of serological testing provided insight into the dynamics of AIV infection in wild birds such as species risk profiles and spatiotemporal patterns, important epidemiological information for a risk-based approach to surveillance.