Arbovirus coinfection and co-transmission: A neglected public health concern?

Epidemiological synergy between outbreaks of viruses transmitted by Aedes aegypti mosquitoes, such as chikungunya, dengue, and Zika viruses, has resulted in coinfection of humans with multiple viruses. Despite the potential impact on public health, we know only little about the occurrence and consequences of such coinfections. Here, we review the impact of coinfection on clinical disease in humans, discuss the possibility for co-transmission from mosquito to human, and describe a role for modeling transmission dynamics at various levels of co-transmission. Solving the mystery of virus coinfections will reveal whether they should be viewed as a serious concern for public health.

Model equations for humans are as follows, with parameter values and their meaning given in S2 Table : − )S dt dS = (λ 1 + λ 2 + λ 12 S )I dt dI 1 = λ 1 − (λ 2 + λ 12 + r  Simulations of the model are initiated by introducing a single infectious human into a population of 1,000,000. After 30 days, a single person with virus 2 is introduced into the population. We assume that there is no cross-protective immunity between the two viruses.

S3 Fig
shows how the total number of co-infections over the course of the outbreak changes as we vary p 1 , p 2 , and p 12 . The highest number of co-infections occur when p 12 is highest (bottom left corner), which is also when p 1 and p 2 are lowest as p 1 + p 2 + p 12 = 1 .

S4 Fig
shows how the number of co-infections (left panel) and the proportion of co-infections due to co-transmission (right panel) varies as we change the probability of co-transmission, p 12 , and keep p 1 = p 2 . more than half of co-infections are due to co-transmission when p 12 =0.175 (right panel).

Probabilities of co-transmission
A rough estimate of the probabilities of co-transmission from co-infected humans to mosquitoes can be obtained with data from Rückert et al. [4] . This study fed mosquitoes on blood that contained a single virus or all combinations of chikungunya, dengue, and Zika viruses. They then tested which mosquitoes were single-or co-infected, and also which mosquitoes had either or both virus in their saliva after 3, 7, and 14 days (a proxy for transmission potential).
For mosquitoes co-exposed to dengue and Zika virus, there were 197 mosquitoes that were infected, of which 118 (60%) were co-infected. Similarly, using the data underlying Figure 4 in Rückert et al. [4] , after 14 days there were 26 dengue/Zika virus co-infected mosquitoes which had at least one virus in their saliva, of which 7 (27%) had both. The equivalent figures for the other pairs of viruses are shown in S3 Table. For simplicity of analysis, our model assumes that virus 1 and 2 have equal probabilities of co-transmission ( ), although in practice it seems likely these values will differ depending on p 1 = p 2 the pair of viruses in question ( S3 Table ). Table. Parameter names, meanings and values. See the next section of the appendix for a discussion of the probabilities of transmission from co-infected Ae. aegypti mosquitoes and humans. Transmission parameters specific for Ae. aegypti mosquitoes were chosen broadly in line with previous dengue modeling studies [1][2][3] , and to produce a final attack rate of 60% for a single invading virus.  Table. Proportion of transmission events leading to co-infection from a co-infected human to a mosquito ( ) or from a co-infected mosquito to a human ( ). In the latter case we assume that p 12 M p 12 co-infection does not affect the transmission probability, and that hence co-transmission occurs in the same proportion as which it is found in the saliva.