Prevalence survey on lungworm (Angiostrongylus vasorum, Crenosoma vulpis, Eucoleus aerophilus) infections of wild red foxes (Vulpes vulpes) in central Germany

Background Angiostrongylus vasorum, Crenosoma vulpis and Eucoleus aerophilus are a source of increasing concern, potentially causing significant pulmonary and severe cardiac/systemic diseases in domestic dogs and wild canids, especially red foxes (Vulpes vulpes). To investigate the prevalence and geographical distribution of these parasites in central Germany, a total of 569 foxes were examined by dissection. Methods Pluck (heart and lung) and faecal samples of red foxes were collected from three regions of Germany. Lungs, hearts and adjacent vessels were processed for adult nematode detection. Parasitological diagnoses of faecal samples were performed by SAF technique, Giardia- and Cryptosporidium-Coproantigen-ELISAs and by a duplex copro-PCR for the detection of A. vasorum and C. vulpis DNA. Results Foxes originated from three Federal States of central Germany: Thuringia (n = 359); Rhineland-Palatinate (n = 121) and Hesse (n = 89). High prevalences for all three nematodes were detected, with E. aerophilus (69.4%; 395/569), followed by C. vulpis (32.3%; 184/569) and A. vasorum (14.1%; 80/569). In case of A. vasorum, prevalences varied significantly between Federal States, with the highest prevalence of 27.3% in Rhineland-Palatinate, followed by 19.1% and 8.4% in Hesse and Thuringia, respectively. The presence of A. vasorum in fox populations showed a rather patchy distribution, increasing from north-eastern to south-western regions. Analyses on C. vulpis revealed prevalences of 35.1%, 30.3% and 25.6% (Thuringia, Hesse and Rhineland-Palatinate, respectively). The most prevalent lungworm nematode was E. aerophilus, with a prevalence of 75.2%, 71.9% and 66.9% (Rhineland-Palatinate, Hesse and Thuringia, respectively) and an almost area-wide equal distribution. Significant differences for single parasite prevalences within geographical regions of the Federal States could be detected whilst no correlation between age or gender and parasite occurrence was estimated. Weak seasonality for the winter months for A. vasorum, stronger correlation to spring and late summer for C. vulpis and no correlation to any season for E. aerophilus were detected. The method of dissection revealed a significantly higher sensitivity for C. vulpis when compared with the results of the duplex copro-PCR. Conclusions A sylvatic cycle was confirmed for all three lungworm nematodes in the examined area. The prevalences for all three lungworm nematodes are some of the highest recorded so far in German foxes. The data suggest that A. vasorum might be spreading from south-western to north-eastern parts of Germany. Electronic supplementary material The online version of this article (10.1186/s13071-018-2672-4) contains supplementary material, which is available to authorized users.


Background
The nematodes Angiostrongylus vasorum, Crenosoma vulpis and Eucoleus aerophilus (syn. Capillaria aerophila) are the most important lungworm species infecting wild and domestic canids in Europe [1,2]. Within the last years, they have become the focus of increased attention from the scientific community due to their occurrence in previously unreported regions in several European countries [2]. Angiostrongylus vasorum has been found in red foxes (Vulpes vulpes) [3], dogs [4][5][6], wolves [7], coyotes [8] and even mustelid species [9]. The distribution of the parasite is rather patchy, with 'hyperendemic' geographical regions being surrounded by areas in which the prevalence of A. vasorum-positive animals is rather low [3,10].
The fox lungworm C. vulpis has also been reported in raccoon dogs [11], wolves [12] and coyotes [13]. Crenosoma vulpis is distributed in several European countries as well as in Alaska and in the Atlantic site of Canada [14]. The distribution in Germany has a rather extensive character and is not focally restricted, but equally distributed [6,15].
For A. vasorum, which is the most pathogenic of the three species, clinical signs may vary from mild, such as coughing or exercise intolerance, to signs of a fatal cardiopulmonary disease (reviewed in [27,28]) and, furthermore haemorrhagic disorders [4] and even nervous complaints [29]. Similar to domestic dogs, pathology has also been observed in A. vasorum-infected foxes [30][31][32], but the effect of this nematode on fox health and population dynamics still remains uncertain [3]. It is widely suggested that foxes act as reservoir hosts of domestic dog infections [3,33], although the relative contribution of larvae shed by dogs or foxes to canine infections has not yet been completely determined [3]. According to Morgan et al. [3], the dynamics of parasite transmission across the wilddomestic animal boundary is likely to be influenced by a wide range of factors. Here for example, high fox population densities in areas of human habitation are discussed, even though an increased A. vasorum transmission from foxes to dogs is not necessarily evidenced for other parasites, such as the tapeworm Echinococcus multilocularis, and its associated rise in the incidence of human disease [34]. Recent geographical spatial reports on canine angiostrongylosis in Germany evidence that A. vasorum is potentially expanding its geographical range from its original occurrence in southern and western parts of Germany [6] to the North and East [35]. A recent nationwide epidemiological survey [36], including > 12,000 domestic dogs, on the distribution and risk factors of angiostrongylosis and crenosomosis confirmed these geographical expanding tendencies.
In the past, A. vasorum-infected foxes were frequently found in close proximity to foci of canine angiostrongylosis [3,30], but in the absence of genetic data, estimates of the degree of overlap in parasite populations between definitive host populations can only be speculative [1]. For E. aerophilus, first genetic analyses illustrated that some sub-populations of the parasite in fact co-infect pets and wildlife [37].
The current study aims to determine actual lungworm prevalences in fox populations in central Germany to better understand the complex epidemiology of these diseases.

Animals and sample collection
From November 2011 to February 2013, carcasses of red foxes (V. vulpes) (n = 569) killed by hunters, road accidents, for monitoring reasons or simply found dead in different counties of three central Federal States of Germany (Hesse, Rhineland-Palatinate and Thuringia) were subjected to necropsies at the regional Veterinary Authorities within the Federal States. Information on the find spot and the date of collection were recorded, wherever applicable. The carcasses were collected within 1-3 days after killing. This might not be applicable for foxes having died by road accidents or simply found dead. All fox carcasses were kept frozen at -80°C for at least 48 h to inactivate Echinococcus eggs. The animals were classified by gender and age [young (approximately up to 2 years): if incisors were clean and unworn; adult (over 2 years): if incisors showed loss of lobulation and brown dentin spots on occlusal surfaces]. At necropsy, heart and lungs were isolated from the carcasses, taking care not to rupture the pulmonary trunk.

Parasite isolation from the pulmonary trunk and parasite identification
After defrosting, the heart, pulmonary arteries and further pulmonary vessels, trachea, bronchi and larger bronchioles were opened via incision. The heart was separated from the lungs by transecting the major blood vessel as close as possible to the lungs, and all ventricles as well as the pulmonary arterial trunk were opened and inspected visually for nematodes. Both heart auricles and ventricles were incised transversally between the base and the apex of the heart. Blood clots were dissolved by digital pressure and the heart chambers were rinsed with tap water through a sieve of 150 μm pore size. The trachea was opened longitudinally and inspected carefully for nematodes. The lung lobes were dissected by opening all visible large pulmonary vessels down to the narrowest practicable point (c.1 mm diameter). All detected nematodes were collected and further identified (see below). Afterwards, bronchi were likewise dissected and then rinsed with tap water over a sieve of 150 μm aperture. Then, water was pumped vigorously through the lung tissues using tap pressure, and flushing was continued until lungs appeared pale and devoid of blood and the washing fluid appeared clear. All washings were sieved. The residues of each sieve were resuspended in tap water and examined under a binocular dissecting microscope (Leica MZ 75, Wetzlar, Germany) using 0.63 × 10.0× magnification for nematodes, larvae or eggs. The nematodes were further identified using a light microscope (Olympus BH-2, Hamburg, Germany) equipped with a digital camera (Olympus SC30, Hamburg, Germany). The parasites were characterized morphologically as described elsewhere [38]. Adult female and male nematodes of all species found were washed with sterile phosphate-buffered saline (PBS) before freezing at -80°C.
Copro-PCR for simultaneous detection of A. vasorum and C. vulpis DNA After defrosting of the fox carcasses, faecal samples were taken from the rectal content (565/569 foxes). In four foxes (three from Thuringia and one from Hesse) the rectum was missing, which was the reason why these foxes could not be examined by copro-PCR. Copro-DNA was isolated from 1 g faeces according to Nunes et al. [39] performing a combination of homogenization by horizontal vortexing with glass beads (Ø 4 mm, Carl Roth, Karlsruhe, Germany; 30 beads/sample; Vortex-Genie 2 with the MoBio horizontal 15 ml tube holder adapter, 10 min vortexing) and an extraction protocol using a commercial kit (QIAamp DNA stool-kit, protocol for larger volumes, Qiagen, Hilden, Germany).
For C. vulpis a diagnostic real-time PCR based on the ribosomal DNA of the internal transcribed spacer 2 (ITS2) sequence was designed. Only two sequences of C. vulpis (18S and 28S ribosomal RNA gene) were available in GenBank. These sequences are highly conserved among different genera of metastrongyloid nematodes making them less suitable for species-specific PCR assays. Therefore, the more variable ITS regions of the ribosomal RNA genes using the universal nematode forward primer NC16 (5′-AGT TCA ATC GCA ATG GCT T-3′ [40]) and reverse primer NC2 (5′-TTA GTT TCT TTT CCT CCG CT-3′ [41]) were amplified and sequenced (GenBank: KF836608). A probe-based real-time PCR assay was designed with Beacon Designer 8.01 (Premier Biosoft, Palo Alto, USA). The specificity of primers and probe was verified by BLAST analysis of the GenBank database and with DNA isolated from C. vulpis and A. vasorum. This assay was combined with the A. vasorum real-time PCR [42] as duplex PCR. The efficiency of the duplex PCR -determined by the standard curve method of 1:10 dilutions of A. vasorum and C. vulpis genomic DNA in a background of DNA from a negative fox faecal sample -was 100% for A. vasorum and 83% for C. vulpis (not shown).

Statistical analysis
The statistical analysis of the data was performed with the help of the validated statistical programme packages BMDP/Dynamic, Release 8.1 [43] and BiAS [44]. The presence of nematodes or DNA detection during PCR for every variable was dichotomised into negative (= not present) and positive (= present) to calculate the estimated prevalences and their corresponding 95% confidence intervals (CI).
In a first step of the analysis the raw association between the presence of the three nematode species and the potential epidemiological influencing factors 'surveyed region' resp. 'geographical area within the surveyed region' , 'gender' , 'age (young or adult)' and 'date of hunting resp. detection (both: monthly and quarterly)' was analysed using Pearson's chi-square test for contingency-tables. For these global comparisons differences were considered as significant at a level of P ≤ 0.05.
In the case of statistically significant global differences, pairwise group comparisons followed with either the chisquare test or Fisher's exact test, controlling the global type I error rate using the Bonferroni-Holm-procedure.
Within the single Federal States the categorization into governmental districts (Darmstadt, Kassel, Giessen for Hesse) or geographical areas (North, Central, South for Rhineland-Palatinate and North, West, East, Central for Thuringia) was performed and the prevalence of the three nematode species was compared in just the same manner.
In a second step of the analysis a multi-way logistic regression model was applied (programme BMDPLR) to consider the possible overlay effects of the independent variables region, gender, age and date. The dependent variables were built by the presence of the examined nematodes (present or absent). In this way adjusted regression coefficients and odds ratios (OR) could be computed. The number of samples being analysed via the multiple logistic regression model had to be reduced to n = 192, because this method needs complete observations, which was not the case for all samples.
In the third part of the analysis the association between the common occurrence of the different nematodes was analyzed. This was done by cross-tabulation, the computation of the kappa-coefficient (κ) as a measure of concordance and computing the chi-square test.

Geographical information system (GIS) database and data visualisation
The data were analysed by a GIS-based approach using the program RegioGraph 10 (GfK GeoMarketing, Bruchsal, Germany) to visualise the regional distribution of collected and analysed fox carcasses and positive results of the necropsies for the three different nematodes on administrative maps. For Hesse and Rhineland-Palatinate the place of hunting or discovery was associated to the corresponding hunting ground with its postal code. For Thuringia the place of hunting or discovery was directly associated to the corresponding postal code. Using these postal codes as points of reference prevalences were displayed on maps with administrative and postcode boundaries as differently coloured areas.

Nematode burden and co-infections
In all three nematode populations found in the hearts and lungs of foxes, both genders were present and gravid female nematodes carrying eggs were frequently detected. Worm burdens ranging between 1-39 (A. vasorum), 1-114 (C. vulpis) and 1-95 (E. aerophilus) were recorded, varying greatly between the nematode species (see Table 1 for details). In total, 440 out of 569 foxes were positive for lungworms.
The highest number of mono-infections was recorded for E. aerophilus (45.7%; 201/440), followed by double infections of E. aerophilus and C. vulpis (30.7%; 135/ 440). For a detailed overview including all recorded double and triple infections see Fig. 1.
The number of positive samples per county or city and per governmental district or geographical region in the Federal States is depicted in Additional files 1, 2 and 3: Tables S1, S2 and S3.
Statistical calculation on county/city basis was not performed due to too low fox numbers in some of the counties/cities. After categorization into larger regions (see also Additional files 1, 2 and 3: Tables S1, S2 and S3) significant differences could be observed in the occurrence of all three nematodes as detailed below.
Statistical comparison on the median numbers of worms found in lungworm-positive samples (worm burden excluding negative samples) between the three Federal States revealed no statistically significant differences for A. vasorum and for C. vulpis Comparing the absence of positive samples in the different counties/cities, altogether 25/58 counties/cities (11 Hessian, 7 Rhineland-Palatinian and 7 Thuringian) were found with a negative result for A. vasorum. Sixteen out of 58 (5 Hessian, 9 Rhineland-Palatinian and 2 Thuringian) counties/cities were negative for C. vulpis and none of the examined counties/cities were negative for E. aerophilus, indicating an almost equal, area-wide distribution of E. aerophilus (see also Fig. 2), a slightly less area-wide, but still widespread distribution of C. vulpis (see also Fig. 3) and a focal, patchy distribution of A. vasorum (see also Fig. 4) in the examined area.   Table 4. Furthermore, raw prevalence differences between genders could be shown only for E. aerophilus (χ 2 = 5.454 df = 1, P = 0.0195; Table 4). Thus, male foxes (78.2%) were more often infected than female ones (65.5%). Referring to age groups, no statistically significant differences were detected for any lungworm species.
To take into account that there are possible overlapping effects between the considered epidemiological parameters, a multiple logistic regression model was fitted to the data. However, complete observations on all criteria were not available for all foxes (reduced sample size: n = 192). This resulted in weaker indications of significance when compared to analyses on raw prevalence differences. Nonetheless, significant seasonal differences in prevalences were still observed for A. vasorum (F = 3.80, df = 3, P = 0.0112) and C. vulpis (F = 6.88, df = 3, P = 0.0002) infections as well as differences between the Federal States for A. vasorum (F = 3.07, df = 2, P = 0.049).
In this multiple analysis the formerly found prevalence differences for A. vasorum between spring and late summer in comparison to the winter months could be confirmed in the form of low odds ratios (OR): OR (spring to winter) = 0.21 (95% CI: 0.05-0.82); OR (early summer to winter) = 0.24 (95% CI: 0.08-0.75). In contrast, for C. vulpis higher odds ratios (directly connected with prevalences) in spring and late summer in relation to the winter months were detected: OR (spring to winter) = 5.9  Table 5.

Discussion
A. vasorum is parasitizing the heart and pulmonary arteries, and has been documented in dogs in Germany with an endemic presence in certain areas [6,35,45]. Western Federal States have been in the focus of canine examinations for A. vasorum infections [35,45] with prevalences between 0.3% and 7.4%. The whole of Germany has been the focus of a recent GIS-supported epidemiological analysis on canine angiostrongylosis/crenosomosis investigating in total 12,682 dogs with positive samples mainly originating from western and southern areas and showing prevalences of 2.3% (A. vasorum) and 2.2% (C. vulpis) [36]. The aim of the underlying study was to gain data on lungworm infections in red foxes from Germany, with a focus on endemic areas of canine lungworm infections. Past German wild reservoir studies with sample sizes between 100 and 400 foxes reported no A. vasorum occurrence, a maximum prevalence of 35% for C. vulpis and up to 77.8% prevalence for E. aerophilus [17,[46][47][48][49], while a recent study examining foxes in Brandenburg (eastern Germany) detected 9.0% prevalence for A. vasorum via PCR in lung tissue [50]. Thus, the overall prevalence of the current study for A. vasorum (14.1%) is the highest reported so far for foxes in Germany. The prevalences for C. vulpis (32.3%) and E. aerophilus (69.4%) are some of the highest reported so far for foxes in Germany.
On a European level the current A. vasorum prevalence was comparable, e.g. to data obtained from some southern European countries [51,52], but higher than data in some of the direct European neighbour countries [53][54][55]. 27.3%) the prevalence of A. vasorum in the last one was already comparable to other highly endemic regions, even though it does not reach Danish levels (48.6%) [56]. Concerning the C. vulpis data, this is one of the highest prevalence reported so far for Europe, surmounted only by data from a few European countries [11,16,[57][58][59][60][61]. Prevalence of E. aerophilus is higher than in most of the previous European studies for foxes and comparable to numbers detected in the Netherlands [54], Hungary [60] and Denmark [56].
Based on our data, a sylvatic cycle can be assumed for all three lung nematodes. The importance of these cycles for the epidemiology in dogs cannot be judged by the actual data alone, but the study helps to fill a gap within the whole epidemiological context of lungworm infections in Germany. As recorded in other studies, prevalences in dogs are generally lower than the ones found in foxes [6,22,35,45,62,63]. These findings could be linked to differences of host innate/adaptive immune reactions of foxes and domestic dogs against the same parasites. Consistently, foxes develop a variable and non-protective immunity against A. vasorum, and this might explain parasite tolerance and long-term survival in these wild canids [64].
Concerning the seasonal distribution of the three lungworms, a highly significant occurrence in summer was detected for C. vulpis, whilst A. vasorum occurrence showed the highest levels in the autumn and winter months. These findings could be linked to differences observed in the biology of the different intermediate host species (i.e. Arion ater, A. distinctus, Tandonia sowerbyi and Limacidae species) [65], particularly on their innate immune response in form of haemocyte-derived extracellular traps against infective lungworm larvae to be involved in gastropod-parasite interactions [66]. Unfortunately, from our own investigations on dissected foxes, it cannot be deduced when natural infections occurred.
In the case of E. aerophilus infections, seasonal data revealed no statistical significance. Endogenous factors, such as prepatency period as well as exogenous factors (micro-and macroclimatic conditions, intermediate host species and slug/snail population dynamics), might additionally influence the seasonality of the parasites [6,65]. There was no significant correlation between prevalences  [67], but also in contrast to studies [31,56], which confirmed explicitly young foxes to be infected with A. vasorum. The same was true concerning the gender of foxes. No significant correlation was detected, as previously described by Magi et al. [67] and in opposite to studies by Saeed et al. [56] where more positive males were identified. The precise parasite burden is often not examined in lungworm-related studies due to the fact that faecal examination is used as proof of infection. Concerning methods for the detection of A. vasorum in the final host, Houpin et al. [68] recently indicated sensitivities of 84% for dissection and visual examination of plucks, 69% for nested PCR and 76% for an antigen detection test, thus showing the advantages of the dissection method.
The underlying study revealed parasite burdens of 1-39 for A. vasorum, 1-114 for C. vulpis and 1-95 for E. aerophilus. For A. vasorum, parasite burdens in dogs have been reported between 1 and 7 in Newfoundland and Labrador [69], and between 1 and 88, with a mean of 17.4 adults in Danish foxes [31]. Thus, the mean infection load for A. vasorum was rather low in the current study. Double and triple infections have already been recorded for the three lung nematodes [31,67]. Apart from fox infections with C. vulpis and E. aerophilus, which showed a significant correlation to one another, no other parasite correlations were statistically significant.
From the 1960s onwards, A. vasorum infection in dogs and foxes was reported at an accelerating rate, and the parasite became established in areas other than the longterm known hyperendemic foci, such as Ireland [70], the UK [71] and Denmark [33]. Since then, there has been a series of new records from different countries in Europe, increasing reports on angiostrongylosis in endemic areas of Europe and North America [1,69]. For Germany, the geographical dispersal of A. vasorum infections in dogs has been reported in a number of studies, demonstrating the highest prevalences for southern and south-western regions [6,35,36,45]. The prevalence of A. vasorum in definitive hosts is strongly affected by different factors, such as host density, parasite aggregation and sample structure, while microclimatic factors rather influence the intermediate host abundance and activity [72]. An expansion of canine A. vasorum endemic areas and/or emergence of this parasite in regions previously believed to be free was recently also reported for Belgium, Bulgaria, Great Britain, the Netherlands, Portugal, regions of Scandinavia or Slovakia [73][74][75][76][77][78][79].
For C. vulpis and E. aerophilus, the distribution pattern in foxes seems to be more generalized in Germany since similar prevalences of 25.6% and 75.2% (Rhineland-Palatinate), 30.3% and 71.9% (Hesse) and 35.1%and 66.9% (Thuringia) were found in the three examined Federal States, respectively. In line with these data, C. vulpis infections in dogs have been described with rather equal, but regionally piled areas [35].
The current geographical dispersal as well as the pattern of dispersion of A. vasorum and C. vulpis in foxes is also mirrored in dogs [6,35,45] with increasing tendencies   [80]. In addition, the different microclimatic conditions and the cold hardiness of the first larvae of C. vulpis [56] may be of importance here.
Overall, E. aerophilus showed an almost area-wide, equal distribution with rather stable, high prevalences, as also reported for Great Britain [75] and Hungary [81].
Finally we analysed the value of a biomolecular approach for the detection of A. vasorum and C. vulpis in comparison to the dissection technique. Overall, the duplex copro-PCR reliably detected DNA of A. vasorum and C. vulpis in fox faeces without prior larval concentration. However, comparing on prevalence level, the dissection results outweighed those of the PCR, demonstrating a lower sensitivity for the copro-PCR. False negative and positive PCR results for the detection of A. vasorum-DNA in faecal samples have also been reported by others [42]. Besides intermittent larval shedding [82] other factors, such as the infection stage, the sample size, the DNA extraction method, the quality of the clinical samples, but also competitive inhibition by other DNAs or the presence of DNases in each faecal sample or inefficient cell lysis during the extraction process were discussed as basis of PCR failure [42].

Conclusions
In conclusion, this study presents an overview on the occurrence of three important canid lungworm nematodes in a wide area and a large fox population in Germany. It thus provides essential epidemiological data from wildlife acting as natural reservoirs, which is urgently needed for the identification of endemic foci of canid lungworm infections and for predictions on the actual spread of A. vasorum. A sylvatic cycle could be here clearly confirmed for all three lungworm species investigated, with far higher prevalences than in the domestic canine cycle. A regional dispersal and prevalence increase from north-eastern to south-western areas, as observed in dogs (mainly for A. vasorum and C. vulpis), could also here be illustrated for German foxes. These data should provide new insights into the epidemiological situation of lungworm infections in free-ranging foxes and call for further investigations on lungworm infections associated to other wild canids (i.e. wolves) of Germany.

Funding
The study was partially funded by Bayer Animal Health GmbH. FK receives funding for a Bayer project at Leipzig University.

Availability of data and materials
The datasets supporting the conclusions of this article are included within the article and its additional files.
Authors' contributions KS performed the experiments, analyzed the data and generated the maps. FK assisted in the documentation of the study data, drafted the manuscript and worked on the presentation of the maps. RS co-financed the study and revised the manuscript. JH established and validated the PCR assays. KF performed all statistical analyses. CH and AT designed the study and coordinated the work. All authors read and approved the final manuscript.