Insights into the spatiotemporal dynamics of West Nile virus transmission in emerging scenarios

The incidence of West Nile fever (WNF) is highly variable in emerging areas, making it difficult to identify risk periods. Using clinical case records has important biases in understanding the transmission dynamics of West Nile virus (WNV) because asymptomatic infections are frequent. However, estimating virus exposure in sentinel species could help achieve this goal at varying spatiotemporal scales. To identify the determinants of inter-annual variation in WNV transmission rates, we designed a 15-year longitudinal seroepidemiological study (2005–2020) in five environmentally diverse areas of southwestern Spain. We modeled individual annual area-dependent exposure risk based on potential environmental and host predictors using generalized linear mixed models. Further, we analyzed the weight of predictors on exposure probability by variance partitioning of the model components. The analysis of 2418 wild ungulate sera (1168 red deer - Cervus elaphus - and 1250 Eurasian wild boar - Sus scrofa) with a highly sensitive commercial blocking ELISA identified an average seroprevalence of 24.9% (95% confidence interval (CI): 23.2–26.7%). Antibody prevalence was slightly higher in wild boar (27.5%; CI: 25.1–30.1%) than in deer (22.2%; CI: 19.8–24.7%). We observed a spatial trend in exposure, with higher frequency in the southernmost areas and a slight, although area-dependent, increasing temporal trend. Host-related predictors were important drivers of exposure risk. The environmental predictor with the highest weight was annual cumulative precipitation, while temperature variations were also relevant but with less weight. We observed a coincidence of spatiotemporal changes in exposure with the notification of WNF outbreaks in horses and humans. That indicates the usefulness of wild ungulates as sentinels for WNV transmission and as models to understand its spatiotemporal dynamics. These results will allow the development of more accurate predictive models of spatiotemporal variations in transmission risk that can inform health authorities to take appropriate action.