Predicting the prevalence of tick-borne pathogens is an important challenge in disease ecology. One such pathogen in the southeastern US is Ehrlichia chaffeensis, the causative agent of ehrlichiosis in humans. E. chaffeensis is transmitted by the lone star tick (Amblyomma americanum) and maintained by the white-tailed deer (Odocoileus virginianus) as its primary reservoir. Reported cases of E. chaffeensis infection in the US in humans has more than doubled since 2000, but its dynamics remain poorly understood. Studies tend to be limited spatially (few sites) and temporally (only one to two years), making predictions difficult. To develop robust models of E. chaffeensis prevalence, we studied dynamics of E. chaffeensis, its hosts, and environmental factors at 130 hierarchically spaced sites across southeastern Virginia. During June and July of 2012, 2013, 2015, and 2016, we collected lone star ticks and used PCR to test for presence of E. chaffeensis. Here, we present an analysis of temporal heterogeneity in E. chaffeensis prevalence within the nymphal tick population. We build generalized linear mixed-effects models to ask what factors drive inter-annual variation in E. chaffeensis prevalence and lone star tick abundance, and discuss potential biological mechanisms.
Results/Conclusions
E. chaffeensis prevalence was positively associated with cumulative rainfall during the current spring and with temperature during the previous fall and winter, and tick abundance was positively associated with rainfall during the current spring and with temperature during the previous spring. Spring represents the end of quiescence and beginning of adult and nymphal questing, and fall and winter represent the mating season of the white-tailed deer. We hypothesize increased precipitation to affect tick survival through the relief of water stress, and decreased temperature to affect deer health or fecundity by increasing caloric demand. Ticks’ ovipositing or hatching success may also be affected by temperature. Additionally, tick abundance also appeared correlated with rainfall during the summer and early fall, which represents the larval questing period, but random slope models show this effect to vary spatially. This differential response may be a function of topography, where larval populations occupying elevated areas respond more strongly. In all, our results highlight the complexity of tick-borne disease dynamics, but also represent an important first step toward modeling changes in E. chaffeensis prevalence through time. This is important for predicting disease risk short term, but also for suggesting long-term trends due to climate change.