Estimating the probability of arbovirus outbreaks in large southern European city infested by Aedes albopictus

Background The presence of the mosquito species Aedes albopictus, competent vector of Chikungunya (CHIKV) and Dengue (DENV), and the possible arrival of infected travellers returning from endemic countries may represent a public health risk for Southern European countries. The aim of this work was to assess the weekly risk of CHIKV, DENV and ZIKA virus outbreaks in Rome tackling both the risk of infected-host introduction and patterns of local transmission. Methods The probability of infected-host introduction was estimated by a binomial process dependent from the number of infected cases in endemic country and the probability of travelling to Rome. A geometric process with means R0HV (reproductive number for host to mosquito transmission) and R0VH (mosquito to host) estimated the probability of successful transmission. Outbreak probability was estimated in 3 scenarios of different vector-host contact ratio and 2 scenarios of epidemic outbreaks in 5 different endemic countries. Weekly data of global cases and inbound and outbound Rome travellers were joined with field-derived estimates of A. albopictus abundance within the city. Results The model correctly estimated the number of DENV and CHIKV imported cases notified to the national health system. The estimated outbreak probability was ≤1% for both DENV and CHIKV under scenarios of low vector-host contacts, but the risk increased significantly (i.e. DENV outbreak risk =21%; CHIKV outbreak risk=47%; null for Zika) under a scenario of higher vector abundance, still consistent with the abundance data from infested hot spots within the urban area. Conclusion This work disentangles the role of seasonality of mosquito dynamics, traveller’s inflow and temporal pattern of infected cases in endemic countries in building up an outbreak risk model.