Surveillance of tick-parasitized voles, mice and roe deer in Germany: Arboviral infection rates in relation to population densities and host characteristics

Abstract

This dissertation is focussed on the ecology of tick-borne encephalitis (TBE), an infectious disease of major medical significance. The thesis is organised as a series of 9 self-contained chapters, enframed by a general introduction to the topic and the research questions (Chapter 1) and a concluding general discussion (Chapter 11). Chapter 2 preludes the chapters 3-9. Using human incidence data from 140 counties in southern Germany over a period of 8 years, several abiotic and biotic factors contributing to the infection risk of TBE were identified in a spatial explicit, predictive framework. By controlling for unexplained spatial and temporal variation, the analysis indicated that type and amount of forest cover, and indexed population densities of roe deer (Capreolus capreolus) in the previous year were positively associated with TBE incidence in humans. An index of forest fragmentation and red deer (Cervus elaphus) and red fox (Vulpes vulpes) population densities in the previous year were negatively associated with TBE incidence in humans. Unexpectedly, spring warming, a suspected key driver for seasonal synchrony of immature ticks and thus for TBE transmission dynamics, was negatively associated with TBE incidence. The results are discussed with respect to their biological significance and the current resolution (county level) of the public health system is critically discussed. Building up from this analysis, the following chapters investigate specific issues on a finer scale using field data. Chapter 3 provides a brief overview of the study area in southern Hesse, where a large fraction of the data collection was carried out. This chapter presents the methods used and highlights the importance of reservoir studies for disentangling the often complex interactions of zoonotic vector-borne diseases. In Chapter 4, extrinsic and intrinsic factors affecting individual tick burdens (Ixodes spp. and Dermacentor spp.) on roe deer were examined. Controlling for season, spatial autocorrelation and individual correlates, the population size of roe deer appears to be positively correlated with nymphal Ixodes spp. tick burdens which lends further support to the notion that roe deer amplify tick densities. In Chapter 5, linear models are presented which allow rapid tick abundance estimations on individual roe deer. For estimating larval burdens, counts can be restricted to the front legs, for nymphal ticks, counts can be restricted to the head and for estimating burdens of adult ticks, counts can be restricted to the neck. The presented models explain considerable parts of the observed variation and thus display reliable alternatives to total body counts. Chapter 6 is focussed on attachment site selection of ticks. Feeding site niches and interstadial feeding site overlap were quantified with standard ecological indices and the distribution of ticks was tested against the ideal free distribution hypothesis. Levels of inter- and intrastadial aggregation were relatively high and ticks of the same stage appear to attract each other which contradicts the ideal free distribution hypothesis. The results are discussed with respect topathogen transmission. Chapter 7 presents a flexible modelling approach which also models the dispersion of a distribution as a function of dependent variables. Similar to Chapter 4, two basic questions (1) Is tick parasitism in roe deer sex biased and (2) Does tick parasitism in roe deer scale with body mass? are addressed. The results are largely in concordance with those of Chapter 4 and support the resource trade-off hypothesis and provide only limited support for the sex-bias hypothesis. Chapter 8 investigates the extrinsic and intrinsic factors affecting tick burdens on rodents. The study confirmed that yellow-necked mice (Apodemus flavicollis) experience higher Ixodes spp. tick burdens than bank voles (Myodes glareolus). Unexpectedly, roe deer densities did not explain the observed variation in individual tick burdens. The analyses provided only limited support for the sex-bias hypothesis and highlight multiple correlates (e.g. vegetation cover, body mass, age) for individual tick burdens. Further on, the results suggest that the individual tick burden declined with increasing rodent population density i.e. ticks were diluted among many hosts. Co-feeding (feeding of nymphal and larval ticks on the same host) was correlated with spring temperature warming and mainly occurs in yellow-necked mice and rarely in bank voles. Also, the occurrence of Dermacentor spp. ticks was recorded and appeared to be positively associated with areas experiencing relatively warm temperatures during the vegetation period. In Chapter 9, roe deer sera were screened for TBE-antibodies and correlates for the presence of these antibodies were identified. The results suggest that the TBE virus mainly circulated in areas with high roe deer densities and that spatial autocorrelation was an important factor. Spring warming was not statistically associated with TBE virus antibody presence in roe deer. Among the individual correlates, hind foot length and number of female Ixodes spp. ticks were identified as best predictors. Yet, the interpretation of these characteristics is not unambiguous. More importantly, results of this chapter demonstrate how variable TBE virus presence in roe deer (and thus in ticks) can be within a county. This strongly suggests scaling down the resolution of public health units to ecologically meaningful units, e.g. from the county level to the forest patch level. Chapter 10 summarises the activities, research questions and preliminary results of the Network Rodent-borne pathogens , mainly focusing on hanta virus research in Germany. With regard to the ecology of TBE, the presented results reframe the importance of spring warming for the maintenance of TBE and suggest that other factors are instrumental in TBE virus transmission (e.g. species composition, presence of the virus in the system). Moreover, strong statistical support for the importance of roe deer densities has been found and it is highly recommended to explicitly identify the biological mechanism behind this association.