Source: Dissertation Abstracts International, Volume: 72-03, Section: B, page: .

Adviser: Maria A. Diuk-Wasser.

West Nile virus (WNV) was introduced into the United States in 1999 and has since become endemic with a range covering most of temperate North America. It is maintained in an enzootic cycle between passerine birds and several species of mosquito vectors. Transmission to humans occurs with infrequent disease outcomes resulting from infections. But, according to the Centers for Disease Control and Prevention, over 29,700 human cases and 1,180 deaths have been reported in the US since 1999. As a disease of public health importance, researcher is needed to seek to identify foci of transmission and predict outbreaks. Simple epidemiological models are unlikely to accurately model transmission dynamics as several of the components of the transmission cycle are highly heterogeneous in both space and time. For example, differences in reservoir host competence may lead to differential transmission depending on the local avian host community structure and host species preferences by vector mosquitoes. These affect host-vector contact rates and may result in differential transmission depending on which host species are differentially fed upon by vectors.

In this dissertation, I present four studies that examine several sources of heterogeneity influencing transmission of WNV in the northeastern United States. In the first chapter, I describe experimental 2-host choice trials in which I determined that Culex pipiens, a major vector for WNV in the northeast, exhibits host preferences for American robins ( Turgus migratorius) over two other commonly co-occurring species, European starlings (Sturnus vulgarus) and house sparrows (Passer domesticus). In chapter 2, I use original data on host preferences exhibited by Culex pipiens at four field sites in Connecticut (CT) to parameterize a mathematical transmission model. The model quantifies the influence of host preference on enzootic transmission and shows that host-preference induced heterogeneity is a major factor determining intensity and timing of enzootic transmission. In chapter 3, I use geographic information system methodology to determine whether increased host diversity results in reduced WNV transmission due to a 'dilution effect.' I also used logistic regression models to identify several key environmental predictors for WNV infection prevalence in Culex vectors; community competence, edge habitat, landscape diversity and mosquito abundance. In chapter 4, I present a field study showing that late-season communal bird roosts play a role in WNV amplification. During late-season, Culex vectors fed often on American robins and were significantly more infected with WNV at communal roosts than non-roosting sites.

The findings presented in this dissertation demonstrate that the interaction between vector host preferences and host community structure results in highly heterogonous transmission patterns across different landscapes. Thus, efforts to predict and prevent WNV in humans will require study at a local rather than regional scale. This research highlights the benefits of interdisciplinary approaches that combine field and laboratory studies with modeling techniques to investigate the complex interactions inherent in multi-host, vector-borne zoonoses.