Habitat and host factors associated with liver fluke ( Fasciola hepatica ) diagnoses in wild red deer ( Cervus elaphus ) in the Scottish Highlands

Background Red deer ( Cervus elaphus ) are a common wild definitive host for liver fluke ( Fasciola hepatica ) that have been the subject of limited diagnostic surveillance. This study aimed to explore the extent to which coprological diagnoses for F. hepatica in red deer in the Scottish Highlands, Scotland, are associated with variability among hosts and habitats. Methods Our analyses were based on coproantigen ELISA diagnoses derived from faecal samples that were collected from carcasses of culled deer on nine hunting estates during two sampling seasons. Sampling locations were used as centroids about which circular home ranges were quantified. Data were stratified by season, and associations between host, hydrological, land cover and meteorological variables and binary diagnoses during 2013–2014 ( n = 390) were explored by mixed effect logistic regression. The ability of our model to predict diagnoses relative to that which would be expected by chance was quantified, and data collected during 2012–2013 ( n = 289) were used to assess model transferability. Results During 2013–2014, habitat and host characteristics explained 28% of variation in diagnoses, whereby half of the explained variation was attributed to differences among estates. The probability of a positive diagnosis was positively associated with the length of streams in the immediate surroundings of each sampling location, but no non-zero relationships were found for land cover or lifetime average weather variables. Regardless of habitat, the probability of a positive diagnosis remained greatest for males, although males were always sampled earlier in the year than females. A slight decrease in prediction efficacy occurred when our model was used to predict diagnoses for out-of-sample data. Conclusions We are cautious to extrapolate our findings geographically, owing to a large proportion of variation attributable to overarching differences among estates. Nevertheless, the temporal transferability of our model is encouraging. While we did not identify any non-zero relationship between meteorological variables and probability of diagnosis, we attribute this (in part) to limitations of interpolated meteorological data. Further study into non-independent diagnoses within estates and differences among estates in terms of deer management, would improve our understanding of F. hepatica prevalence in wild deer.