Predicting West Nile virus circulation: a 20-year spatiotemporal study in humans and animals in Spain, 2003 to 2022

BACKGROUND While West Nile virus (WNV), a mosquito-borne zoonotic pathogen, is detected every year in animals in Spain, clinical human cases occur more sporadically. Most explanatory and predictive models for WNV circulation focus on single components and aggregate multiple year data into a single dataset. AIM We sought WNV circulation environmental drivers across different ecological components (vectors, reservoirs, and dead-end hosts), by analysing their spatial and temporal dynamics. METHODS We used active and passive surveillance data collected in Spain between 2003 and 2022, encompassing mosquitoes, 120 bird species, 115 mammal species, and humans. To understand WNV circulation, mosquito spatial and host spatiotemporal models were developed, incorporating current and lagged environmental variables. Our One Health approach integrated the different models to determine WNV exposure risk, including 1 year in advance. RESULTS Over 20 years, WNV exposure risk in Spain rose by 19% in birds, 17% in non-human mammals, and 38% in humans. In birds and non-human mammals, exposure more likely occurred in areas experiencing mean respective annual temperatures > 5 °C and > 8 °C in the previous year. In humans, increased exposure risk concurred with mild winters (> 5.3 °C). Integrating mosquito and host models found the country’s southern half and mediterranean coast most suited for WNV. Predictive models solely using prior-year variables yielded comparable results to contemporaneous ones. CONCLUSION The models suggest that annual human WNV transmission may occur more regularly than evidenced by surveillance, possibly due to asymptomatic or misdiagnosed cases. Our framework could serve as an early warning tool, enhancing outbreak preparedness up to 1 year ahead.