West Nile virus surveillance using sentinel birds: results of eleven years of testing in corvids in a region of northern Italy

The natural transmission cycle of West Nile virus (WNV) involves birds as primary hosts and mosquitoes as vectors, but this virus can spread to mammals, human beings included. Asymptomatic infected donors pose a risk to the safety of blood transfusions and organ transplants, as WNV can be transmitted through these medical procedures. Since 2009, the region of Emilia-Romagna in northern Italy has been implementing an integrated surveillance system in order to detect WNV circulation in the environment at an early stage. Here we report the results of the two components of the surveillance system, the active testing of corvids and humans, and demonstrate that bird surveillance alone improves a surveillance system based solely on human case detection. As WNV risk reduction measures are applied on a provincial basis, we assessed the ability of this surveillance system component to detect virus circulation prior to the notification of the first human case for each province. Overall, 99 epidemic seasons were evaluated as a result of 11 years (2013–2023) of surveillance in the nine provinces of the region. In this period, 22,314 corvids were tested for WNV and 642 (2.9%) were found to be infected. WNV was generally first detected in birds in July, with sample prevalence peaks occurring between August and September. During the same period, 469 autochthonous human cases were notified, about 60% of which were reported in August. WNV was detected 79 times out of the 99 seasons considered. The virus was notified in birds 73 times (92.4%) and 60 times (75.9%) in humans. WNV was first or only notified in birds in 57 seasons (72.1%), while it was first or only notified in humans in 22 seasons (27.8%). Active surveillance in corvids generally allows the detection of WNV before the onset of human cases. Failure of virus detection occurred mainly in seasons where the number of birds tested was low. Our results show that active testing of a minimum of 3.8 corvids per 100 km2 provides a satisfactory timeliness in the virus detection, but for early detection of WNV it is crucial to test birds between mid-June and mid-August.