Background

Human alveolar echinococcosis (AE), caused by the metacestode stage of the tapeworm Echinococcus multilocularis is considered as one of the most pathogenic zoonosis in temperate and arctic regions of Europe [1]. The life-cycle of E. multilocularis involves small rodent intermediate hosts such as arvicolids and wild or domestic canid definitive hosts such as the red fox (Vulpes vulpes), the Arctic fox (Vulpes lagopus), the raccoon dog (Nyctereutes procyonoides) or the dog (Canis lupus f. familiaris). Humans can act as aberrant intermediate hosts and are infected through the ingestion of eggs excreted in the faeces of definitive hosts. Such faecal-oral infection can be acquired by contact with definitive hosts or through contamination of soil, food or possibly water [2]. In humans, the metacestode stage resembles a malignant neoplasia as it proliferates indefinitely by exogenous budding and slowly invades the surrounding tissue to produce tumour-like lesions [3]. Human alveolar echinococcosis is characterized by an asymptomatic incubation period of around 5–15 years [4].

In Europe, the human risk of E. multilocularis infection was considered in the past to be restricted to certain geographical regions. In fact, until the 1990s, only a ‘core’ area consisting of Eastern France, southern Germany, parts of Switzerland and Austria were known to be endemic for the disease [5]. More recently, the expansion of the parasite into several new areas such as the Baltic regions, Denmark, the Netherlands, Poland, Romania, Slovakia, Slovenia and the increase of human AE incidence in ‘core’ areas such as Austria, France and Switzerland, suggested that the disease was spreading in Europe and the incidence of human AE increasing at least in some regions [6–10]. Although greater awareness and the use of advanced diagnostic tools may have contributed to an improvement in the detection of E. multilocularis infection in animals and humans, epidemiological research conducted over the past 20 years, suggested the expansion of this parasite in European countries [9]. Factors such as change in landscape composition and use, vegetation, climate change, presence of good intermediate hosts, urbanization of foxes, changing human behavioural attitudes toward foxes, wildlife reintroduction, E. multilocularis host population dynamics as well as globalization have all been proposed as potential factors influencing the increase of E. multilocularis infection risk for Europe [9, 11, 12].

In the light of these concerns, the European Commission (EC) adopted a Commission Delegated Regulation (EU) No. 1152/2011 (14 July 2011). This was considered as a preventive health measure to control E. multilocularis infection in dogs and decrease the potential risk of AE infection in humans, in order to ensure continuous protection of Finland, Ireland, Malta and the United Kingdom (UK), countries that have remained free from E. multilocularis [13]. Regulation 1152/2011 described the obligations of these four European Union (EU) member states in implementing a pathogen surveillance programme for the detection of E. multilocularis in accordance with specific requirements regarding sampling, detection and reporting procedures [14]. It also stipulated that the EC had to review this regulation by December 2016 to assess the justification of these preventive health measures, in the light of scientific developments regarding E. multilocularis infection in animals. In response to Article 29 of Regulation (EC) No. 178/2002, in addition to an EC request, the European Food Safety Authority (EFSA) was tasked with assessing E. multilocularis infection in animals within the EU and neighbouring Adjacent Countries (ACs) (Albania, Belarus, Bosnia and Herzegovina, Iceland, Kosovo, Liechtenstein, Macedonia, Moldova, Montenegro, Norway, Russia, Serbia, Switzerland, Turkey and the Ukraine). To fulfil this requirement, EFSA funded a project to provide a comprehensive and quantitative assessment of the current knowledge on E. multilocularis using a systematic review (SR) approach (GP/EFSA/AHAW/2012/01: Echinococcus multilocularis infection in animals).

The current SR provides an overview of the distribution and prevalence of E. multilocularis in the EU and ACs derived from both scientific and grey literature. In addition, the purpose of this review was to systematically determine the geographical distribution of E. multilocularis and the known wild and domestic definitive and intermediate hosts. The retrieved information was used to compile tables on the occurrence of E. multilocularis or highlight the lack of reliable reports. When available, data on E. multilocularis prevalence and worm burden of definitive hosts was reported. The importance of the various definitive and intermediate host species in the life-cycle of E. multilocularis in different parts of the EU and ACs was assessed and gaps in our knowledge were identified.

Methods

Bibliographic searches

This SR and meta-analysis followed the Cochrane and PRISMA Group guidelines [15] and the systematic search was carried out using the Documentation Service for literature search at the Istituto Superiore di Sanità, Rome, Italy. The STN International-Fiz Karlsruhe platform [16] was used for database searching carried out on the 5th November 2013 and again on the 11th February 2015 in order to identify articles that had been published since the initial search. The results of these two searches were then combined. Searches were carried out using the Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica Database (EMBASE), Science Citation Index (SciSearch), Biological Abstracts (BIOSIS), Centre for Agricultural Bioscience International (CABI) and Google Scholar. Databases were searched using keywords associated with the Boolean operators “AND” and “OR”. The question mark (“?”) was used to expand searches by looking for words with similar prefixes using more than one letter (i.e. “echinococc?” was used to search for “echinococcus”, “echinococci”, “echinococcosis” and “echinococcoses”). The hashtag (“#”) was used to expand searches by looking for words with similar prefixes using one letter (i.e. dog# was used to search for “dog” or “dogs”).

Different combinations of words and Boolean operators were used in order to narrow results retrieved and maximise the number of relevant studies returned. The full electronic search strategy, including any limits used was: [Echinococcus multilocularis OR (Echinococcus AND Multilocularis) OR E# Multilocularis OR Alveolar Echinococcosis OR A# Echinococcosis] AND (Dog OR Dogs OR Cat OR Cats OR Canis OR Felis OR Canid? OR Felid? OR Wolf OR Wolves OR Animal OR Animals OR Fox OR Foxes OR Vulpes OR Ferret OR Ferrets OR Rodent OR Rodents OR Rodentia OR Nutria# OR Muskrat# OR Jackal# OR Arvicolid? OR Arvicolinae OR Worm Burden OR Host OR Hosts OR Hosted) AND (Occurrence# OR Geographic? Distribut? OR Geographic? Diffus? OR Incidence# OR Frequency OR Epidemic Outbreak# OR Endemic Outbreak# OR Prevalence# OR Epidemiology)]. If the title or abstract did not give a clear indication of relevance, the full text was screened. After this initial selection, full-text articles were evaluated for eligibility, in accordance with the inclusion/exclusion criteria described below. Data extraction was performed independently by two researchers and any disagreements were resolved either by consensus among researchers or through arbitration by an additional independent researcher. If database outcomes overlapped, all duplicated articles were removed. EU reports and conference proceedings were searched using the keywords “European Union report, “EU report”,”, “conference proceedings”, “Echinococcus multilocularis”, “E. multilocularis” and “alveolar echinococcosis”. Unpublished epidemiological data on E. multilocularis available within individual member states was collected from the National Reference Laboratories for Parasites in Europe [17] using a questionnaire (Additional file 1: Text S1). Searches for Bachelor, Masters and PhD theses were carried out using the keywords “Echinococcus multilocularis” and “alveolar echinococcosis”. A list of databases used for retrieving theses is available in Additional file 2: Text S2. Review Manager [18] software was used to prepare and maintain this SR.

Study selection

Studies eligible for inclusion were defined a priori and fulfilled the following criteria: (i) studies published from 1900 to 2015; (ii) studies based on cross-sectional or cohort design; (iii) primary research studies either published or in press; (iv) reports on wild or domestic hosts of E. multilocularis; (v) studies published in English, German, French, Polish, Finnish, Dutch, Spanish or Italian.

The list of included articles is available in Additional file 3: Text S3. Studies providing data from outside Europe and ACs, case reports, reports on E. multilocularis in humans, studies on agents other than E. multilocularis (e.g. Echinococcus granulosus), reviews and letters or editorials without original data were all excluded from this SR. The list of excluded articles is available in Additional file 4: Text S4. The study selection process was carried out according to the PRISMA statement [15] and is reported using the flow chart shown in Fig. 1.

Fig. 1
figure 1

The flow chart represents the algorithm of record/article selection

Full size image

Eligibility for inclusion in the meta-analyses

Studies included in the meta-analyses were those that reported prevalence data (total number of studied animals and number of positive animals) and studies with a definition of a geographical area (whenever possible the Nomenclature of Territorial Units for Statistics designated as NUTS level 1, 2 and 3 was used) [19]. When studies originated from different geographical areas or when they were conducted within the same geographical area but at different time intervals (e.g. during distinct years or months) they were divided into sub-studies. Data were extracted from each study independently. If the same samples were tested using different diagnostic methods, only data derived from the sedimentation and counting technique (SCT) or the intestinal scraping technique (IST) were included in the analysis. Studies reporting prevalence data obtained exclusively by enzyme-linked immunosorbent assays (ELISAs), designed to detect pathogen-specific copro-antigens in DHs, were subsequently excluded from the meta-analysis because of the low specificity of this test. When it was not possible to accurately assign the proportion of data reported per country, meta-analysis was not performed.

Statistical approach and meta-analyses

Statistical analysis was conducted using the statistical software Stats Direct 2.8.0 (Stats Direct Ltd., Altrincham, UK). To perform the meta-analysis, animal species were divided into two main groups, definitive (DH) and intermediate (IH) hosts. The DHs included the red fox (Vulpes vulpes), the Arctic fox (Vulpes lagopus), the raccoon dog (Nyctereutes procyonoides), wild canids (wolf, Canis lupus; golden jackal, Canis aureus), the dog (Canis lupus f. familiaris) and the cat (Felis silvestris f. catus). The IHs included arvicolid rodents (including Arvicola spp., Myodes (syn. Clethrionomys) glareolus and Microtus spp. but excluding the muskrat Ondatra zibethicus), muskrat, nutria or coypu (Myocastor coypus), murid rodents (including Apodemus spp., Micromys minutus, Mus musculus and Rattus spp.), insectivores (including Sorex spp., Neomys fodiens and Talpa europaea) and swine (domestic Sus scrofa f. domesticus and the wild boar Sus scrofa). Each meta-analysis group included studies conducted in the same geographical area, at the European level, national level and using the three NUTS levels [19]. Meta-analyses were not stratified for the years/months in which the studies were conducted.

Since all included studies were cross-sectional, meta-analyses on proportions were performed. The Cochran’s Q test was performed to assess the degree of heterogeneity between studies, and the I2 index was used to describe the percentage of total variation across studies as a result of heterogeneity. If the p-value from the Cochran’s Q test was < 0.05 and the I2 statistic was > 50 %, heterogeneity was found and a random-effect model was applied. However, if heterogeneity was not detected, a fixed-effect model was used. A forest plot was produced to describe the pooled analysis; this showed the single prevalence of the studies and the pooled proportion with relative 95 % confidence intervals (CIs). Publication bias was quantified by inspection of funnel plots and computation of Begg and Egger’s probability values [20, 21].

Quality assessment

The quality of all included studies was assessed independently by two researchers using the Newcastle-Ottawa Scale (NOS) according to the Cochrane Handbook for Systematic Reviews [22, 23]. The NOS was modified for use on an animal model. Quality assessment could not be performed on grey literature.

Results and discussion

Bibliographic searches identified 2,805 scientific papers, of which 1,429 were deleted due to duplications. At the end of the search, 1,376 papers were identified of which 974 were excluded based only on title and abstract screening. A total of 402 full-text papers were assessed for eligibility, data were extracted from 255 studies and it was possible to perform meta-analyses on 244 studies (Fig. 1). The quality assessment carried out using the modified NOS, allowed the allocation of a maximum 7-star rating to any one individual study. A score of 5 or 6 was given to 108, 9, 1, 17 and 8 studies on foxes, raccoon dogs, wild canids, dogs and cats, respectively. A lower score (4 or 3) was assigned to 79 studies on foxes, 7 on raccoon dogs, 2 on wild canids, 9 on dogs and 12 on cats. A similar scoring for intermediate hosts showed that 2 studies on muskrats and 5 on arvicolids had a 5 or 6 rating. Four or three star ratings were assigned to 6 studies on muskrats, 11 on arvicolids, 4 on murids, 2 on nutria, 1 on insectivores and 1 on swine, respectively.

Geographical distribution and prevalence of Echinococcus multilocularis

Red foxes

Data regarding the geographical distribution and prevalence of E. multilocularis in red foxes were extracted from reports published for the period between 1968 and 2014 (Table 1).

Table 1 Pooled prevalence of Echinococcus multilocularis in red foxes
Full size table

A total of 192 papers describing the distribution and prevalence of E. multilocularis in foxes were used in the meta-analyses. A preliminary ranking of E. multilocularis infection in red foxes based on pooled prevalence allowed us to identify three main groups (Table 1). A low prevalence group included countries with a pooled prevalence of ≤1 %, namely Denmark [24–27], Slovenia [28, 29] and Sweden [24, 30–36]; a medium prevalence group with a pooled prevalence of > 1 % but ≤ 10 %, which included Austria [37–44], Belgium [24, 45–55], Croatia [24, Relja Beck, personal communication], Hungary [24, 56–59], Italy [24, 60–66], the Netherlands [30, 31, 40, 49, 54, 67–72], Romania [73–75] and the Ukraine [76, 77], whereas the high prevalence territories had a pooled prevalence of > 10 % and included the Czech Republic [30, 31, 40, 78–83], Estonia [84–86, L. Laurimaa, personal communication], France [30, 31, 40, 87–104], Germany [24, 30, 31, 39–41, 105–151], Latvia [152], Lithuania [153, 154], Poland [155–171], Slovakia [24, 28, 31, 40, 164, 172–184], Liechtenstein [70] and Switzerland [24, 30, 39, 40, 185–198]. The occurrence and pooled prevalence of E. multilocularis in foxes in the EU and ACs is shown in Fig. 2. The highest prevalence estimates for E. multilocularis in red foxes seem to be concentrated in central and north-eastern Europe. A more detailed map of the geographical distribution and pooled prevalence of E. multilocularis in red foxes at a NUTS 1 level is shown in Fig. 3. Studies from four countries, namely Finland, Ireland, the UK and Norway, reported the absence of E. multilocularis in red foxes [24, 31, 32, 158, 196–203]. Echinococcus multilocularis in Arctic foxes in Norway was documented only for the Arctic Archipelago of Svalbard [207, 208].

Fig. 2
figure 2

Pooled prevalence of Echinococcus multilocularis in red and Arctic foxes within the European Union and adjacent countries at national level (data obtained from studies performed after 2000). Note: the pooled prevalence data for Norway originated only from Arctic foxes on the Svalbard islands; prevalence data from Spain originated from single studies

Full size image
Fig. 3
figure 3

Pooled prevalence of Echinococcus multilocularis in red foxes within the European Union and adjacent countries at NUTS 1 level (data obtained from studies after 2000). Note: prevalence data from the Netherlands and Sweden originated from single studies; only studies reporting NUTS information were taken into account

Full size image

Other definitive hosts

Five potential DHs of E. multilocularis other than red foxes were identified in the screened literature; four wild animal species, the Arctic fox [21, 24, 209], the raccoon dog [24–26, 32, 77, 152, 154, 168, 169, 175, 177, 199, 210–213], the golden jackal [24, 214] and the wolf [31, 77, 215] and two domestic animal species, dogs [24, 30, 72, 80, 95, 101, 122, 142, 168, 169, 177, 193, 195, 199, 200, 216–225] and cats [24, 37, 66, 80, 113, 115, 121, 122, 132, 133, 142, 168, 169, 177, 216, 219, 223, 226–230]. The geographical distribution and prevalence of E. multilocularis for these DHs are summarized in Table 2.

Table 2 Pooled prevalence of Echinococcus multilocularis in Arctic foxes, raccoon dogs, wild canids (golden jackal and wolf), cats and dogs
Full size table

Pooled prevalence results showed that sylvatic animals, excluding red foxes, are more frequently infected than domestic species. The two species showing high E. multilocularis prevalence were the raccoon dog and the golden jackal. In general, high E. multilocularis prevalence in these two species correlated with high infection rates in foxes. Importantly, the raccoon dog is currently not established in some areas that are deemed free of E. multilocularis (e.g. Ireland, Malta and the UK), but is present in high numbers in Finland. A third species, with high prevalence rates was the Arctic fox [21, 24, 209], which is only present in a few northern countries, namely northern Russia, Iceland, and the Norwegian Arctic Archipelago of Svalbard, in addition to a small population on the Scandinavian peninsula.

Dogs and cats do not seem to be important in terms of prevalence and are found to be infected only in some areas of high E. multilocularis pooled prevalence in red foxes such as Czech Republic [80, 212], Germany [24, 113, 115, 121, 122, 132, 133, 142, 216, 219, 227, 228], France [24, 30, 95, 101, 216–218, 226] and Switzerland [24, 193, 195, 223–225]. However, dogs can be regarded as potentially relevant hosts considering E. multilocularis introduction into areas that are free of the parasite by travelling from endemic to distant (non-endemic) areas with their owners, and also with regard to transmission in endemic areas because of their closer association with humans than sylvatic DHs.

Information on E. multilocularis worm burden in definitive hosts from EU countries and ACs was only available in a few studies for red foxes (43/190) [25–27, 29, 34, 38, 55–59, 61, 63, 68, 75–77, 79, 80, 85, 99, 102, 110, 119, 127, 140, 152–154, 161, 165–167, 173–176, 179–181, 189, 191, 192], raccoon dogs (3/17) [32, 154, 213], dogs (1/23) [224] and cats (5/19) [80, 113, 132, 169, 228]. In contrast, no data were available on E. multilocularis worm burden of wild canids and Arctic foxes for the same regions.

Intermediate hosts

Potential IHs of E. multilocularis screened in this study included the muskrat [51, 113, 121, 136, 231–241], arvicolids [24, 30, 32, 39, 51, 79, 93, 95, 101, 121, 141, 187, 189–191, 207, 209, 223, 224, 242–257], murids [51, 79, 95, 101, 168, 169, 224, 229, 242, 246, 250, 251, 253, 256], nutria [231, 233], swine [32, 73, 220, 258] and insectivores [24, 79, 101, 251] (Table 3). For the majority of countries, the distribution of the prevalence of E. multilocularis in muskrats and arvicolids matched that (although the prevalence was lower) in red foxes and was similar to the pooled prevalence of E. multilocularis in other sylvatic DHs (Table 2). Muskrats and arvicolids are thus potentially good sentinels to investigate the presence of E. multilocularis in specific settings.

Table 3 Pooled prevalence of Echinococcus multilocularis in arvicolids, muskrat, nutria, swine (domestic and wild), insectivores and murids
Full size table

Among murids, Apodemus spp. was the host with the highest E. multilocularis prevalence [24, 79, 224, 242, 246, 250, 251, 253, 256]. In France, E. multilocularis prevalence in these species was similar to that reported for Microtus spp. [256]. Only one study on E. multilocularis infection in Mus musculus in France is known to exist [95]. In general, murids have not frequently been found positive for E. multilocularis [24, 79, 101, 168, 169, 224, 229, 242, 246, 250, 251, 253]. However, the number of studies (n = 14) and the number of murids examined remains small (n = 2,610). None of the screened insectivores were positive for E. multilocularis [24, 79, 101, 251] but the number examined was small (n = 531). Although swine seem to play no role in the life-cycle of this parasite, E. multilocularis infections in swine were reported from Germany [259], Lithuania [220] and Switzerland [258] and therefore this animal species could potentially be regarded as a domestic IH sentinel (Table 3). Data regarding E. multilocularis in definitive and intermediate hosts in EU countries and ACs are summarised in Table 4.

Table 4 Data on Echinococcus multilocularis infection in definitive and intermediate hosts
Full size table

Ranking of hosts (other than red foxes) in the life-cycle of Echinococcus multilocularis

Definitive hosts

In order to clarify the importance of other screened DHs in the life-cycle of E. multilocularis, pooled prevalence for each DH, other than red foxes were generated (Table 5). The ranking of pooled prevalence in DHs could be used to hypothesise the importance of the different DHs in the life-cycle of E. multilocularis.

Table 5 Pooled prevalence of Echinococcus multilocularis in definitive hosts other than red foxes
Full size table

Ranking based on an E. multilocularis pooled prevalence of > 3 %, resulted in the following order (high to low rank): red fox, Arctic fox, golden jackal, raccoon dog and wolf. Although data on the golden jackal and the Arctic fox are scarce [21, 24, 209, 214], they provide evidence in support of these two animal species serving as potentially important DHs of E. multilocularis. Despite some uncertainties due to the low number of studies regarding these two species, data have nevertheless been included in this report for the following reasons: (i) these are the only data available for the golden jackal and the Arctic fox; and (ii) parasite prevalence in the studied individuals was high (Arctic fox, 9 %, 95 % CI: 6–12; golden jackal, 4.7 %, 95 % CI: 0.1–15.3), which is indicative of the potentially important role that these species could play in the maintenance and transmission of E. multilocularis. Interestingly, Arctic foxes are restricted to the northern area of the EU and ACs because of their habitat needs, but the golden jackal population seems to have an increasing trend of migrating from eastern EU countries and ACs towards the west, which should be taken into account when considering the potential future spread of E. multilocularis.

Intermediate hosts

In order to clarify the importance of the screened IHs in the life-cycle of E. multilocularis, the pooled prevalence for each IH group was determined (Table 6). Pooled prevalence in the screened IH groups showed that muskrats and arvicolids (muskrats, n = 25,985; arvicolids, n = 65,956) (and more specifically Arvicola spp.) are important in the life-cycle of E. multilocularis. For nutria (n = 650) and murids (n = 2,610), the number of animals screened was too low to draw any robust conclusions, although it seems that they could play a role in the life-cycle of E. multilocularis in areas with a sustained medium to high pooled prevalence in red foxes [24, 79, 95, 101, 168, 169, 224, 229, 231, 233, 242, 246, 250, 251, 253, 256]. Swine and insectivores seem to play no role in the life-cycle of E. multilocularis within the EU and ACs.

Table 6 Pooled prevalence of Echinococcus multilocularis in intermediate hosts
Full size table

The importance of different definitive hosts in countries classified as having low, medium and high prevalence rates of Echinococcus multilocularis

Definitive hosts

Considering that the number of studies and the number of animals screened in many cases were too low for drawing robust conclusions, the following comments should be regarded as tentative.

The importance of each screened DH, according to country, was stratified by the pooled prevalence of E. multilocularis in red foxes (or Arctic foxes in Svalbard, Norway). The resulting classification, with regard to E. multilocularis infection, enabled us to group countries into zero, low, medium or high prevalence regions (Table 7). The raccoon dog [24–26, 32, 199, 200], the wolf [31], the dog [31, 216] and cat [216] were screened in countries with low (including absence of the parasite) E. multilocularis prevalence in foxes. None of these DHs, at this level of fox prevalence, seem to sustain the life-cycle of E. multilocularis, although issues relating to the representativeness of the sample number should be taken into account since, occasionally, the number of screened animals was low (raccoon dogs, n = 3,833; dogs, n = 27,638; cats, n = 13,498).

Table 7 Grouping of countries according to Echinococcus multilocularis prevalence in red foxes in relation to definitive (DH) and intermediate (IH) hosts
Full size table

For countries stratified in the medium E. multilocularis prevalence group, golden jackals [24, 214], if present, seem to participate in the life-cycle of the parasite, with prevalence estimates roughly similar to those reported for red foxes in the same countries [56–59, 120]. By contrast, wolves [77], dogs [24, 72, 216] and cats [216] seem to play no role in countries with medium E. multilocularis prevalence levels in foxes  [24, 30, 31, 37–44, 49, 54, 60–72]. For countries with high E. multilocularis prevalence levels, raccoon dogs [24, 152, 154, 168, 169, 175, 177, 210–213] are also important in the life-cycle of the parasite, with prevalence estimates of between one-seventh and two-thirds of the pooled prevalence in foxes. An exception is evident in Slovakia, where the pooled E. multilocularis prevalence in foxes [24, 31, 40, 164, 172–184, 260] was similar to the prevalence found in raccoon dogs (~27 %) [24, 175, 177]. Importantly, in countries with a high prevalence, an additional DH (i.e. wolf) seems to join the life-cycle of E. multilocularis, although with a lower prevalence (one-sixth) than that reported for foxes and raccoon dogs [215].

With regard to domestic hosts (dogs and cats), only a very low prevalence of E. multilocularis could be found and only in highly endemic situations (Table 7), and thus these hosts seem to be of minor importance in the life-cycle of the parasite in Europe and ACs, especially when a zero, low or medium E. multilocularis prevalence is found in foxes. In addition, cats have been shown to be unsuitable hosts for E. multilocularis, because full maturity of the parasite is often not attained in the feline intestine [261].

Intermediate hosts

In countries with a low (including 0) E. multilocularis pooled prevalence in foxes, only two types of IHs have been screened, namely arvicolids (in Finland) [24] and swine (in Finland and Sweden) [32] whereas in other countries such as Ireland, Slovenia and the UK no IHs have been inspected for the prevalence of E. multilocularis. Therefore, to interpret these results, the potential importance of those IHs in medium- and high-prevalence situations should first be assessed. Muskrats and arvicolids seem to be the only IHs for E. multilocularis in medium-prevalence rated countries. In muskrats, a pooled prevalence of 16 % was recorded in Belgium [51, 232] and a prevalence of 0.06 % in the Netherlands [236] where the pooled prevalence for E. multilocularis in foxes was 8 and 4.7 %, respectively. Similarly, in Norway (Arctic fox pooled prevalence 5.8 %), Romania (fox pooled prevalence 4.5 %) and Belgium (fox pooled prevalence 8 %), the pooled prevalence for E. multilocularis in arvicolids was 27.5 % [30, 39, 207, 209], 1.4 % [262] and 0.2 % [51], respectively. In countries with a high E. multilocularis prevalence, the prevalence estimates were high for arvicolids (13.3 %) [187, 189–191, 223, 224, 252–254] and pigs (10 %) [258] in Switzerland (fox pooled prevalence 17 %), muskrats (3.8 %) [113, 121, 136, 231, 235–237, 240, 241] and pigs (5.3 %) [259] in Germany (fox pooled prevalence 29.2 %) and arvicolids (4.8 %) [93, 95, 101, 242–249, 255, 257] and nutria (5.8 %) [233] in France (fox pooled prevalence 13.9 %) (Table 7).

Gaps and conclusions

Generally gaps were found in the literature regarding the following aspects (i) NUTS level specifications beyond the national level were absent in many reports, making it difficult to localise infection foci within specific areas for each country; (ii) many EU countries and ACs (n = 18) had no data on E. multilocularis prevalence in definitive or intermediate hosts, even in cases where E. multilocularis infection was probable because the parasite had been found in surrounding countries; (iii) data on the prevalence of the parasite in DHs, other than red foxes, and in some IHs were scarce and often reported in only one single study; (iv) the number of screened animals was considered insufficient in some reports in which the estimated prevalences were low; and (v) publication bias (for example there may be unpublished studies regarding the absence of E. multilocularis within the EU and/or ACs).

Furthermore, inadequacies were identified with regard to the assessment of E. multilocularis prevalence in red foxes. Specific gaps were that (i) the vast majority of studies were concentrated in six countries (Germany, France, Slovakia, Switzerland, Poland and Hungary, whereas the estimates of the pooled prevalence for other countries was based on few studies or further to this, in two cases (Liechtenstein and Spain) on single studies; (ii) sampling in some countries had been done in only specific areas in which it was assumed that the prevalence might be high and thus extrapolation of the data at national level could be biased; and (iii) bias may arise as a result of the sampling strategy used. The sampling strategy data for red foxes are summarised in Additional file 5: Table S1. In the current SR, 50/190 studies relating to fox sampling and E. multilocularis control programmes, excluding those based on coproELISA (n = 10), were included in the analysis. In addition, data were obtained from 20/190 papers describing rabies control programmes, in which foxes were probably mainly obtained by shooting. A further 38/190 papers included in this study did not report the type of sampling methods utilized. Additionally, in 133 studies examined, fox carcasses were made available to authors/authorities through other sources (road kill; hunting season). This type of sampling strategy can cause bias with regard to restrictions in sampling locations, since hunting for example is generally conducted in areas distant from human habitation. Therefore, in more than half of the prevalence studies, synanthropic fox populations living in villages, towns or cities were not included in the sampled animals. This may be the case for all fox sampling within the EU and ACs.

Specific gaps and weaknesses were also found for data relating to DHs other than red foxes. These were that (i) the number of studies was very low (n = 44) for the five DHs; (ii) some of the DHs are geographically restricted, for example, Arctic foxes are limited to northern latitudes [21, 24, 209] and golden jackals are found in only a few countries [24, 93, 209, 214]; (iii) some of the DHs such as raccoon dogs were not found on island countries (Ireland, Malta and the UK) and (iv) some of the DHs are protected species (e.g. the wolf). Specific gaps and deficiencies in data for IHs were that (i) the number of studies were very low (n = 27) for all screened IHs, excluding arvicolids and muskrats; and (ii) some of the IHs were geographically restricted.

In addition, in terms of the importance of definitive and intermediate hosts of E. multilocularis, this systematic review identified gaps regarding the following aspects: (i) the number of studies for the different hosts and the number of screened animals was very low, excluding red foxes, muskrats and arvicolids; and (ii) data on worm burden and worm maturity for the different DHs or fertility of protoscoleces in different IHs were lacking, precluding the assessment of the real role of each host in the maintenance of the life-cycle of E. multilocularis. However, the ranking of animals according to their importance as hosts may be useful in providing recommendations for the screening of DHs to better ascertain the presence of E. multilocularis in a given area. Host screening strategy should be as follows: in the absence of the most important DH, the second most important DH should be screened and so forth. Nevertheless, both the presence of hosts and the protected status of some species (e.g. wolves) are a matter to be taken into account when a recommendation for screening is given.

When conducting epidemiological studies, particularly if the absence of the parasite or a low to medium prevalence is expected and if red foxes cannot be screened, sylvatic animals should, preferably, be screened if the aim is to demonstrate the absence or presence of E. multilocularis. When the presence or maintenance of the life-cycle is to be assessed, the suitability of each DH to allow the full maturation of the parasite (worms producing infective eggs), and the evaluation of worm burden, should be taken into account. In a similar manner, when the presence or maintenance of the life-cycle needs to be assessed, the suitability of each IH to allow full maturation of the parasite (protoscolex production) should be considered.

The prevalence in muskrats and arvicolids seems to parallel those found in red foxes and if foxes cannot be screened, a larger number of muskrats and arvicolids than foxes would need to be screened to confirm the absence of E. multilocularis. This is necessary because the prevalence in foxes as compared to Arvicola spp. appear to correlate at a ratio of around 3:1. Similarly, in areas where both M. glareolus and Microtus spp, were found, E. multilocularis prevalence correlated with that in foxes at a ratio of 1:4–6 (Table 7). An exceptional case is Svalbard in Norway, where Microtus spp. had a 27 % E. multilocularis prevalence and the DH (Arctic fox) showed around 9 % prevalence [207, 209]. This could be attributed to ecological variables specific for this DH-IH interaction, since the IH (Microtus levis) has a very limited spatial distribution, while Arctic foxes are able to stroll on ice and can cover long distances and are therefore not limited to either the Spitsbergen Island nor to the Svalbard Archipelago. The only additional potential DH in this area is the dog, but this DH has to date not been screened in this region.

This SR has also highlighted gaps in our knowledge regarding mustelids and the role they may potentially play in the life-cycle of this parasite. Studies on E. multilocularis infection in mustelids (including Mustela spp., Neovison vison, Lutra lutra, Meles meles and Martes spp.) from Europe initially formed part of this meta-analysis. None of the studied mustelids from the Czech Republic (n = 6) [80], Denmark (n = 29) [24], Germany (n = 1142) [24, 122, 133, 142, 263], Poland (n = 22) [168, 169, 251], Slovakia (n = 18) [175, 177] and the Ukraine (n = 26) [77] were found infected with E. multilocularis. Interestingly, mustelids (Martes spp.) from Ryazan district, Russia were recently found to harbor adults of E. multilocularis [264]. As far as we are aware this is the only known report on the occurrence of E. multilocularis in mustelids and is the only known study that identified mustelids as ‘definitive hosts’ based on the presence of E. multilocularis adults in the intestine of 4/31 Martes species. While this infection can evidently occur, no information on E. multilocularis worm maturity, worm burden or prepatency was provided. Additionally, no mustelid-derived faecal samples have been unequivocally confirmed by molecular methods to be positive for E. multilocularis. Furthermore, we speculate that this infection may occur as a result of the predator-prey relationship of carnivorous mustelids and small rodents. In the absence of studies in which a larger number of mustelids are examined and/or experimental data we were reluctant to include data on mustelids in this analysis. Although the absence of E. multilocularis in mustelids in Europe suggests that they may not be important hosts of this parasite, further studies are required in order to clarify their role.

Importantly, the presence of E. multilocularis in red foxes cannot be excluded from countries where data may have been published using languages other than those represented by this SR consortium, but where this host is known to be present. For example there are many publications (albeit in Russian) on E. multilocularis in animals in the former Soviet Union ([265], Paul Torgerson personal communication]). High E. mulitlocularis infection rates in foxes (33.1 %) and raccoon dogs (15.4 %) were reported from Ryazan district [264] and in foxes from Bryansk Oblast (41 %) [266] and Kamchatka (14.7 %) in the east, respectively [267]. Similarly, high E. multilocularis infection rates of 40 % and 98 % were found in Arctic foxes from Krasnoyarsk region [268] and Sakha, Yakutia [269], respectively. In addition, reports on rodents have documented E. multilocularis infection in Apodemus uralensis and Microtus arvalis [270] in Kabardin-Balkar and Clethrionomys spp. in Sakha [271]. In a similar manner, the absence of E. multilocularis in foxes in countries for which only a few studies were available, may not be representative of the infection status of foxes in those particular areas.

Conclusion

In conclusion, this SR confirmed the status of the red fox as the most important definitive host of E. multilocularis in the EU and ACs. If the prevalence in foxes was zero or low in a given area, there was no indication that the life-cycle of E. multilocularis was maintained by other DHs. In contrast, when the prevalence level in red foxes was greater than 3 %, both raccoon dogs and golden jackals, if present, seemed to play a similar role as the fox in the life-cycle of the parasite. In terms of IHs, muskrats and M. glareolus, if present, are important hosts in the life-cycle of E. multilocularis. Under specific conditions, Arvicola spp. and Microtus spp. could be important in the life-cycle of the parasite. Swine and insectivores seem to play no role in the life-cycle of E. multilocularis within the EU and ACs.