Angiostrongylus vasorum is an emerging parasite in dogs, with frequent new reports throughout Europe and beyond [1, 2]. Within several endemic Europe countries, its distribution appears to be expanding, but in a highly heterogeneous or patchy manner. In the United Kingdom (UK), A. vasorum was first reported in Cornwall in southwest England in the 1970s [3, 4], later in Wales and southeast England [5–7], and recently in central and northern England and Scotland [8–12]. National questionnaire surveys of veterinary practices confirmed both northward spread and persistent heterogeneity in disease incidence in dogs [13]. Better understanding of the factors underlying this patchy distribution is needed to guide risk assessments by veterinary clinicians and other stakeholders, in order to prevent and control the severe disease that can ensue from infection [14], as well as to build on currently sparse fundamental understanding of the epidemiology of nematode parasites in gastropod intermediate hosts [15, 16].

Angiostrongylus vasorum has an indirect life-cycle with canids such as domestic dogs (Canis lupus familiaris) and red foxes (Vulpes vulpes) as the definitive host, terrestrial gastropod molluscs (slug or snail) acting as intermediate hosts, and frogs acting as paratenic or intermediate hosts [17]. Heterogeneity in parasite abundance and disease risk to definitive hosts could arise from variable environmental influences on mollusc populations and infection rates. Using a recently developed real time polymerase chain reaction (PCR) assay [18], surveys of prevalence in slugs are feasible. Previous studies assessing the presence of A. vasorum in slugs and snails have reported variable prevalence. In Denmark, 29 % of molluscs sampled contained larvae and prevalence varied by locality [19]. The species found to be infected in this study were Arion lusitanicus, Ar. ater, Ar. aterrufus and Limax maximus. A survey in south Wales (UK) reported 43 % prevalence using PCR, with Ar. ater aggregate, Ar. hortensis aggregate and Deroceras sp. positive [18]. In London (UK), prevalence of larval recovery from slugs was 1.6 %, with a broad host range including two species previously not recorded as acting as intermediate hosts for A. vasorum: Ar.distinctus and Tandonia sowerbyi [20]. In a PCR survey of slugs in west Scotland, an area newly colonised by A. vasorum, the slug fauna and positive slugs were dominated by large Arion species [11], with overall prevalence of 7 %.

The aims of the present study were to determine the prevalence of infection in slugs in contrasting areas within a region endemic for A. vasorum, and to determine whether differences in slug species composition could explain differences in overall prevalence. This information is needed to better understand the role of slugs as intermediate hosts for A. vasorum, the factors underlying local risks of transmission, and the most important slug species involved as sources of infection for dogs.

These results demonstrate for the first time a rural–urban gradient in A. vasorum prevalence in intermediate hosts in the UK. Higher prevalence of A. vasorum in suburban areas is consistent with studies in Denmark in gastropods [19] and in foxes [23]. Peak prevalence in suburban areas could be explained by overlap of high fox population density and high density of suitable intermediate hosts, as proposed for the fox tapeworm, Echinococcus multilocularis, in Europe [24].

Wide variation in the prevalence of A. vasorum at different sites and between the cities of Swansea and Bristol demonstrates a high level of spatial heterogeneity in parasite distribution among intermediate host populations. Previous studies have noted marked geographical patchiness of angiostrongylosis in definitive hosts [25–28]. The reasons for this patchy distribution are unknown [30], and could include spatial variation in intermediate host density and species composition, and in environmental factors influencing transmission, as well as local movement of infected intermediate hosts around new parasite introductions [30–32].

The prevalence of 29.4 % in Swansea in the current study is considerably higher than the 1.6 % recorded by Patel et al. [20] in London, by extraction and identification of larvae from gastropod tissues. Larval extraction methods are likely to underestimate true infection prevalence, through low sensitivity and because pre-infective stage larvae are usually not recovered [Abdul-Aziz, unpublished observations]. Conversely, prevalence estimated from PCR does not accurately reflect the presence of mature larvae able to propagate infection in definitive hosts, since earlier larval stages and, potentially, residual DNA from non-viable larvae can also be detected [11]. Prevalence estimates using these different methods are, therefore, not comparable. Low prevalence in a large sample size in the present study in Bristol suggests that false positive results are not common using this PCR assay.

In the present study, slug species were identified by external and internal morphology based on Rowson et al. [22], a new comprehensive guide to British and Irish slugs validated by PCR amplification and sequencing of mitochondrial (16S, COI) and nuclear (ITS-1) DNA. Previous work on A. vasorum in slugs [18–20] examined ‘aggregates’ of superficially similar species. The present study therefore provides more accurate information on species distribution of A. vasorum based on more specific identification.

The present study found broadly similar slug species composition in the nearby cities of Swansea and Bristol. Together with the wide host range observed here and in previous studies, this suggests that the overall prevalence of A. vasorum is not closely related to the presence or absence of particular intermediate host species. Nevertheless, some differences in slug fauna were noted. Sampling in Swansea yielded fewer species overall, even allowing for lower sample size, and less variability in slug species between sites compared with Bristol. Arion rufus was relatively more common across locations in Swansea, and Ar. flagellus and some of the smaller slug species were more common in Bristol. Reasons for these differences could include ecological factors (e.g. climate, soil, habitat, or predation) or a history of accidental introduction (perhaps especially for those species found at single sites). However, the consistent differences in abundance of widespread species present at all sites (Ar. rufus and Ar. flagellus) seem more likely to be due to ecological rather than historical factors. Arion rufus, widely found and commonly infected in the Swansea samples, has been acknowledged to act as an efficient host of A. vasorum and to be a major potential source of infection in various locations [18, 33, 34]. It appears that in urban areas of Swansea, L. maculatus acts as additional or secondary host. In Bristol, the lower availability of Ar. rufus as the preferred or optimal host of A. vasorum in Bristol might explain the low overall infection prevalence. Nevertheless, Ar. flagellus and L. maculatus were abundant and could be important alternative hosts in such areas.

The broad intermediate host range observed in this study is in agreement with previous data on experimental and natural infections [18–20, 35]. Around 25 species have been successfully infected in an experimental setting. These are: (i) terrestrial slugs (Deroceras reticulatum, D. agreste, D. laeve, Arion lusitanicus, Ar. hortensis, Limax flavus, Laevicaulus alte); (ii) terrestrial snails (Arianta arburstorum, Bradybaena similaris, Cochlodina laminata, Cepea nemoralis, Euparypha pisana, Prosopeas javanicum, Helix aspersa, Helix pomatia, Subulina octona, Succinea putris, Achatina fulica,Vitrea diaphana); (iii) aquatic snails (Lymnaea peregra peregra, Lymnaea tomentosa, Planorbis planorbis, Biomphalaria glabrata, Anisus leucostomus, Physa sp.) [33, 35–40]. In natural infections, mollusc species reported to be hosts for A. vasorum were: Arionidae: Arion ater aggregate (including A. ater seg., A. rufus seg., A. lusitanicus auctt.), Arion hortensis aggregate (including A. distinctus), Arion subfuscus aggregate (including A. subfuscus seg.), Geomalacus maculatus; Agriolimacidae: Deroceras sp., Deroceras reticulatum, Deroceras caruanae auctt., Deroceras laeve, Limacidae: Limax sp., Limax marginatus, Limax maximus, Limacus maculatus; Milacidae: Milax sp., Tandonia sowerbyi; Snails: Cornu aspersum [4, 11, 18–20, 33–35, 38, 41]. However, the extent to which these types of mollusc are accessible to definitive hosts may vary. A broad intermediate host range has also been reported for other Angiostrongylus species, such as A. cantonensis [42, 43] and A. costaricensis [44, 45].

The slugs most commonly infected in this study were larger Arion spp., which have been previously found to have higher larval burdens than smaller slugs [41, 46]. Rosen et al. [38] reported average infection levels in Ar. rufus of 750 L3 and over 1000 in some cases. Arion flagellus and L. maculatus, alternative hosts for A. vasorum in the present study, are also among the larger slug species. It is possible that these larger slugs are more commonly infected than others through greater coprophagic tendencies. Arion flagellus was found to be a new host record for A. vasorum. There is evidence of ongoing distributional spread of both Ar. flagellus and L. maculatus in the UK and elsewhere [21, 22].

The current study demonstrated infection of slugs with A. vasorum in summer and autumn, confirming previous records in summer [18], autumn [19] and summer and winter [20]. Thus, slugs are likely to present a source of infection to dogs all year round. The much lower prevalence in Bristol in summer, compared with that in Swansea in autumn, might stem from a seasonal rather than a spatial effect, and results are therefore not comparable between locations. Higher prevalence in autumn might be expected based on accumulation of infection and maturation of larvae through the summer months. More information is needed on seasonal variation in A. vasorum prevalence in intermediate hosts, to contribute to better understanding of the seasonal epidemiology and risks of infection to dogs, and also to guide efficient sampling for surveillance or spatial studies.

In conclusion, the results presented here demonstrate heterogeneity in the distribution of A. vasorum infection in slugs in two nearby cities, and variation in prevalence along the rural-urban gradient. The composition of the slug fauna appears to play a role in parasite distribution at these scales. Further studies are needed to account for ecological and host factors underlying species differences in intermediate host infections, and to separate spatial from seasonal drivers of prevalence.