Résumé
Abstract Mosquito-borne West Nile virus (WNV) infection is a growing global health problem. About 0.5% of infected individuals develop encephalitis. We previously showed that 40% of patients in six cohorts had WNV encephalitis because of circulating auto-antibodies (auto-Abs) neutralizing type I IFNs. In seven new cohorts, we found that the prevalence of auto-Abs was highest (40% [17-44%]) in patients with encephalitis, and very low in a small sample of individuals with asymptomatic or mild infection. In the 13 European, Middle-Eastern and American cohorts available, odds ratios for WNV encephalitis in individuals with these auto-Abs relative to those without them in a large sample of the general population untested for WNV infection range from ∼20 (OR=17.7; 95% CI: 13.8-22.8, p <10 −16 ) for auto-Abs neutralizing only 100 pg/mL IFN-α2 and/or IFN-ω to >2000 (OR=2218.4; 95% CI: 125.1-39337.7, p <10 −16 ) for auto-Abs neutralizing high concentrations of IFN-α2 and high or low concentrations of IFN-ω. Pre-existing autoantibodies neutralizing type I IFNs are therefore causal for WNV encephalitis in about 40% of patients. Summary In 13 cohorts of individuals with WNV infection, the risk of WNV encephalitis is increased 20 to >2,000 times by circulating auto-Abs neutralizing type I IFNs, depending on the concentration and combination of type I IFNs neutralized and patient age.