Experimental evidence for opposing effects of high deer density on tick-borne pathogen prevalence and hazard

Abstract

Background: Identifying the mechanisms driving disease risk is challenging for multi-host pathogens, such as Borrelia burgdorferi sensu lato (s.l.), the tick-borne bacteria causing Lyme disease. Deer are tick reproduction hosts but do not transmit B. burgdorferi s.l., whereas rodents and birds are competent transmission hosts. Here, we use a long-term deer exclosure experiment to test three mechanisms for how high deer density might shape B. burgdorferi s.l. prevalence in ticks: increased prevalence due to higher larval tick densities facilitating high transmission on rodents (M1); alternatively, reduced B. burgdorferi s.l. prevalence because more larval ticks feed on deer rather than transmission-competent rodents (dilution effect) (M2), potentially due to ecological cascades, whereby higher deer grazing pressure shortens vegetation which decreases rodent abundance thus reducing transmission (M3).

Methods: In a large enclosure where red deer stags were kept at high density (35.5 deer km^-2), we used an experimental design consisting of eight plots of 0.23 ha, four of which were fenced to simulate the absence of deer and four that were accessible to deer. In each plot we measured the density of questing nymphs and nymphal infection prevalence in spring, summer and autumn, and quantified vegetation height and density, and small mammal abundance.

Results: Prevalence tended to be lower, though not conclusively so, in high deer density plots compared to exclosures (predicted prevalence of 1.0% vs 2.2%), suggesting that the dilution and cascade mechanisms might outweigh the increased opportunities for transmission mechanism. Presence of deer at high density led to shorter vegetation and fewer rodents, consistent with an ecological cascade. However, Lyme disease hazard (density of infected I. ricinus nymphs) was five times higher in high deer density plots due to tick density being 18 times higher.

Conclusions: High densities of tick reproduction hosts such as deer can drive up vector-borne disease hazard, despite the potential to simultaneously reduce pathogen prevalence. This has implications for environmental pathogen management and for deer management, although the impact of intermediate deer densities now needs testing.

Keywords: Borrelia burgdorferi sensu lato; Dilution effect; Ecological cascades; Ixodes ricinus; Lyme disease.

Fig. 1

Conceptual diagram to illustrate three pathways through which high deer density might affect nymphal infection prevalence (NIP) with Borrelia burgdorferi sensu lato (s.l.), and how the density of infected nymphs (DIN-Lyme disease hazard) depends on a combination of NIP and the density of questing nymphs (DON)

Fig. 2

Mechanism 2—dilution effect. NIP for Borrelia burgdorferi s.l. (%) [± 95% confidence interval (CI)] in high deer density and deer-exclusion plots

Fig. 3

Mechanism 1—increase in transmission potential. DON (nymphs 10 m⁻²) (± 95% CI) in high deer density and deer-exclusion plots. For abbreviations, see Fig. 2

Fig. 4

a–d Mechanism 3—ecological cascades linking high deer density with Lyme disease pathogen prevalence in Ixodes ricinus ticks. Graphs show predicted outputs from generalized linear mixed-effects models of ground vegetation height (a), bank voles per 100 trap nights (TN) with ground vegetation height (b), bank voles per 100 TN in high deer density and deer-exclusion plots (c), and NIP with Borrelia afzelii (%) with bank vole abundance the previous year (d). Error bars and shaded areas represents 95% CI. For other abbreviations, see Fig. 2

Fig. 5

Effects of deer density on DIN (nymphs 1000 m⁻²) (± 95% CI) in high deer density and deer-exclusion plots. For abbreviations, see Fig. 2