ABSTRACT: A broad range of rodent species are described as potential intermediate hosts for Echinococcus multilocularis, a wide-spread zoonotic cestode causing alveolar echinococcosis. However, little is known about the relative contribution of these species for parasite reproduction and the maintenance of its life cycle. In a comparative study in a high endemic region in Zurich, Switzerland, we investigated prevalence rates and fertility of E. multilocularis in the most abundant vole species as well as the predation rate of foxes on these species. To ensure that the fox families had access to different vole species and that these voles were exposed to the same environmental contamination with parasite eggs, we selected eight study plots where at least two rodent species co-occurred. The parasite prevalence in Microtus arvalis [11.0%, confidence intervals (CI) 8.9-13.4] was significantly higher than in Arvicola scherman (5.3%, 3.9-7.1) and Myodes glareolus (3.9%, 2.0-6.7). None of the, only 29 individuals of, Microtus agrestis was infected (0%, 0.0-9.8) and the species was excluded for further analyses. Logistic regression models for the prevalences revealed significant differences between nearby study plots and higher infection rates for females, heavier individuals, and individuals trapped during spring, when the prevalence in M. arvalis peaked up to 65% (CI 50-79) in one plot. Furthermore, we detected significantly higher percentages of fertile infections in M. arvalis and M. glareolus than in A. scherman (OR 11.2 and 6.4, respectively) and a significantly higher protoscolex number in M. glareolus (median 100,000) than in M. arvalis (13,500) and A. scherman (4,290). The most abundant fox prey remains were of the genera Microtus (12.3%, CI 8.4-17.2) and Arvicola (11.5%, 7.7-16.3), whereas Myodes was never recorded as prey (0.0-1.3%). We conclude that M. arvalis and to a lesser extent A. scherman can be regarded as key intermediate hosts in Western and Central European high-endemic regions whereas M. glareolus and M. agrestis play a marginal role. We, therefore, postulate that distribution models of these species could contribute to predict parasite occurrence on a more detailed spatial scale than models of the distribution of foxes which have a very broad and uniform distribution.
