Résumé
Zoonotic diseases transmitted by wildlife affect biological conservation, public and animal health, and the economy. Current research efforts are aimed at finding wildlife pathogens at a given location. However, a meta-analytical approach may reveal emerging macroecological patterns in the host–pathogen relationship at different temporal and spatial scales. West Nile virus (WNV) is a pathogen with worldwide detrimental impacts on bird populations. To understand macroecological patterns driving WNV infection, we aimed to recognize unknown competent reservoirs using three disease metrics—serological prevalence (SP), molecular prevalence (MP) and mortality (M)—and test if these metrics are correlated with the evolutionary history, geographical origin of bird species, viral strain, time–space and methodology. We performed a quantitative review of field studies on birds sampled for WNV. We obtained 4945 observations of 949 species from 39 countries. Our analysis supported the idea that MP and M are good predictors of reservoir competence, and allowed us to identify potential competent reservoirs. Furthermore, results indicated that the variability of these metrics was attributable to phylogeny, time–space and sample size. A macroecological approach is needed to recognize susceptible species and competent reservoirs, and to identify other factors driving zoonotic diseases originating from wildlife.