Risk of introducing African horse sickness virus into the Netherlands by international equine movements

African horse sickness (AHS) is a vector-borne viral disease of equines that is transmitted by Culicoides spp. and can have severe consequences for the horse industry in affected territories. A study was performed to assess the risk of introducing AHS virus (AHSV) into the Netherlands (P_AHS) by international equine movements. The goal of this study was to provide more insight into (a) the regions and equine species that contribute most to this risk, (b) the seasonal variation in this risk, and (c) the effectiveness of measures to prevent introduction of AHSV. Countries worldwide were grouped into three risk regions: (1) high risk, i.e., those countries in which the virus is presumed to circulate, (2) low risk, i.e., those countries that have experienced outbreaks of AHS in the past and/or where the main vector of AHS, Culicoides imicola, is present, and (3) very low risk, i.e., all other countries. A risk model was constructed estimating P_AHS taking into account the probability of release of AHSV in the Netherlands and the probability that local vectors will subsequently transmit the virus to local hosts. Model calculations indicated that P_AHS is very low with a median value of 5.1×10(-4)/year. The risk is highest in July and August, while equine movements in the period October till March pose a negligible risk. High and low risk regions contribute most to P_AHS with 31% and 53%, respectively. Importations of donkeys and zebras constitute the highest risk of AHSV release from high risk regions, while international movements of competition horses constitute the highest risk of AHSV release from low and very low risk regions. Preventive measures currently applied reduce P_AHS by 46% if compared to a situation in which no preventive measures are applied. A prolonged and more effective quarantine period in high risk regions and more stringent import regulations for low risk regions could further reduce P_AHS. Large uncertainty was involved in estimating model input parameters. Sensitivity analysis indicated that uncertainty about the probability of non-notified presence of AHS in low and very low risk regions, the protective effect of quarantine and the vector-host ratio had most impact on the estimated risk. Furthermore, temperature values at the time of release of AHSV largely influenced the probability of onward spread of the virus by local vectors to local hosts.