Interactions between seasonal temperature variation and temporal synchrony drive increased arbovirus co-infection incidence

Abstract

ABSTRACTThough instances of arthropod-borne (arbo)virus co-infection have been documented clinically, the overall incidence of arbovirus co-infection and its drivers are not well understood. Now that dengue, Zika, and chikungunya viruses are all in circulation across tropical and subtropical regions of the Americas, it is important to understand the environmental and biological conditions that make co-infections more likely to occur. To understand this, we developed a mathematical model of cocirculation of two arboviruses, with transmission parameters approximating dengue, Zika, and/or chikungunya viruses and co-infection possible in both humans and mosquitoes. We examined the influence of seasonal timing of arbovirus cocirculation on the extent of co-infection. By undertaking a sensitivity analysis of this model, we examined how biological factors interact with seasonality to determine arbovirus co-infection transmission and prevalence. We found that temporal synchrony of the co-infecting viruses and average temperature were the most influential drivers of co-infection incidence. For seasonal patterns typical of a tropical region, we observed non-negligible incidence irrespective of arrival time when two arboviruses arrived simultaneously. Under our default parameter settings, this corresponded to a maximum co-infection cumulative incidence of 83 per 1,000 individuals and a minimum cumulative incidence of 32 per 1,000 individuals in the year following arrival. For seasonal patterns typical of a more temperate region, co-infections only occurred if arrivals took place near the seasonal peak, and even then, did not reach 0.01 co-infections per 1,000 individuals. Our model highlights the synergistic effect of co-transmission from mosquitoes, which leads to more than double the number of co-infections than would be expected in a scenario without co-transmission. Our results show that arbovirus co-infections are unlikely to occur in appreciable numbers unless epidemics overlap in space and time and in a tropical region.