Invasive Fascioloides magna and Its Italian “Alcatraz”

Simple Summary

The deer fluke (Fascioloides magna) is a liver parasite that came to Europe from Northern America in the last centuries. This study aimed to understand if the parasite has spread outside La Mandria Natural Park (LMNP) and which factors could help it spread. From 2012 to 2023, researchers checked the livers and faeces of wild herbivores in LMNP. They also collected snails, which are important for the parasite’s life cycle, and tested them. Cameras were used to see if animals move in and out of the park. The results showed that the parasite is still present inside LMNP, especially in red deer, and only rarely in roe deer. The parasite was not found outside the park, neither in animals nor in snails. Some snails inside LMNP carried the parasite, but snails outside did not. Even though some animals sometimes leave the park, the environment outside does not seem suitable for the parasite to survive and reproduce. Overall, the risk of the parasite spreading outside the park is currently low. However, regular monitoring and strong control measures are still important. This study shows that using different methods together helps manage and control invasive parasites in wildlife.

Abstract

Fascioloides magna, an invasive trematode introduced to Europe in the 19th century, persists in two main foci: the Danube basin and La Mandria Natural Park (LMNP) in northern Italy. This study assessed whether the parasite has spread beyond LMNP and evaluated environmental and host-related risk factors. Between 2012 and 2023, 331 wild ruminant livers were examined, and faecal samples were analysed for fluke eggs. Gastropods from the LMNP were sampled using a predictive habitat suitability model and screened for F. magna DNA. Camera traps monitored ungulate movements across LMNP boundaries. Results confirmed the parasite’s presence in red and fallow deer within LMNP and sporadic cases in roe deer, but no evidence of infection in wildlife or gastropods outside the park. Molecular screening detected F. magna DNA in 9.2%% of snails inside LMNP only. Despite occasional crossings by potential definitive hosts, ecological conditions outside LMNP appear unsuitable for sustaining the parasite’s life cycle. These findings suggest a low current risk of spread but highlight the need for continued surveillance and barrier reinforcement. The integrated approach combining parasitology, molecular diagnostics, and GIS-based risk mapping provides a valuable framework for managing invasive parasitic diseases in wildlife.

1. Introduction

The deer fluke (DF), Fascioloides magna, is a large digenetic trematode that infects various wild ungulates across its native distribution area in North America [1]. It was unintentionally imported to Europe between the 1860s and 1930s, following the import of exotic wapiti (Cervus canadensis) and white-tailed deer (Odocoileus virginianus) for sport hunting [2,3]. In Europe, F. magna occurs in two distinct foci. The historical focus was first reported in the second half of the 19th century at “La Mandria”, a royal hunting estate (now a nature reserve) on the outskirts of Turin in northern Italy (hereafter referred to as LMNP) [2,3]. Since then, this focus has remained localised, apparently due to a physical barrier in the form of a 34 km-long wall, which has prevented the invasive parasite from spreading further [2]. The second, much larger focus was identified during the first decade of the 20th century and is still expanding within the vast hydrographic basin of the Danube River [3]. It is affecting areas in the Czech Republic, southwestern Poland, in the floodplain forest of the Danube, in order of progressive invasion [4,5,6,7]. In both areas, the main hosts of the reservoirs are the red deer (Cervus elaphus), which are native to Europe, and the fallow deer (Cervus dama), which are exotic but have been present for a long time. Other hosts that occur in one focus or the other are considered dead-end hosts, as fluke eggs are rarely shed in their faeces. These include wild boar (Sus scrofa), moose (Alces alces), European bison (Bison bonasus), cattle (Bos taurus) and horses (Equus caballus) [1]. Alternatively, they are considered aberrant hosts, as excessive visceral migration is observed and flukes usually do not reach adulthood. These include sheep (Ovis aries), goats (Capra hircus), European mouflons (Ovis aries musimon) and northern chamois (Rupicapra rupicapra) [1,8,9].

Infestation by F. magna is usually well tolerated by wild reservoir hosts, except in young or debilitated individuals [1,10]. However, signs such as depression, poor appetite, anaemia and weight loss may occur [1,8,9], and fatal outcomes have occasionally been reported in naturally infested white-tailed deer [11] and red deer [12] as well as in experimentally infested wapiti [8] and mule deer (Odocoileus hemionus) [9]. In contrast, the outcome of infestation in aberrant hosts is often fatal within six months, even at a low intensity of migrating flukes. This has been observed in naturally and experimentally infested small domestic ruminants, as well as a variety of wild caprine species in North America and Europe [13]. Accordingly, the endemic occurrence of DF among competent wildlife is considered to be a limiting factor in the profitability of extensive sheep and goat farming. Livestock economics in DF-endemic zones are also affected by income losses due to liver condemnations in subclinically infested cattle [14].

The epidemiological role of the roe deer (Capreolus capreolus), traditionally categorised as a non-typical host of deer fluke [1], has recently been debated, as chronic infestations were recorded in a significant proportion of naturally exposed individuals in Poland [15]. The lower lethality at both the individual and population levels than previously thought, together with the significant prevalence and intensity of fluke egg shedding by roe deer alongside infested red and fallow deer, raises the question of whether roe deer should be considered an additional, previously unrecognised DF reservoir host, with all the resulting management implications. Roe deer can reach much higher population densities than other cervids considered reservoir hosts of the disease [16], and can colonise a variety of habitats, including riparian and agricultural landscapes [17]. They show an obvious tendency to disperse from their birthplace [17], which could facilitate the spread of this invasive parasite in Europe.

There is anecdotal evidence that roe deer became extinct in the LMNP during the early phases of a two-year fascioloidosis outbreak in the late 1970s, which almost halved the local red and fallow deer herd. This was the most severe episode of mortality ever reported amongst F. magna-infested wildlife [12]. At that time, no roe deer were necropsied and there were no population estimates to assess fluctuations in roe deer mortality. Since the end of the 20th century, the number of roe deer in the area around LMNP has clearly increased, and personnel of the protected area have progressively recorded sightings of dispersing individuals. In fact, the wall surrounding the fenced part of the LMNP may be permeable to roe deer and other wildlife crossing at creek entry and exit points. More recently, DF environmental DNA was detected outside the fenced part of LMNP [18], raising concerns about F. magna potentially escaping its ‘inviolable prison’ and establishing itself in the unfenced zone. This would represent an unfortunate event, likely paving the way for the spread of DF in northern Italy, to the detriment of the health of susceptible wild and domestic ruminants. Roughly 150 years after the accidental introduction of the invasive disease to the LMNP, this study aimed to investigate its presence in the immediate vicinity of the historical Italian focus. The presence of F. magna was investigated in a range of potential definitive hosts, both inside and outside the fenced zone, as well as in Galba truncatula, the principal vector in Europe [16]. To optimise sampling of snails for testing for the presence of F. magna DNA, original habitat suitability mapping was developed. Finally, camera traps were used to quantify the flow of potential definitive hosts into and out of the fenced zone of LMNP at permeable points.

2. Materials and Methods

2.1. Study Area

‘La Mandria’, which was managed as the Savoia royal family’s hunting estate for centuries, was designated a natural park in 1978. The protected area covers 6,500 hectares of flat to slightly undulating land with predominantly clayey soils, close to the town of Turin in northern Italy (45°18′02″ N; 7°51′62″ E). Interestingly, approximately half of the park is enclosed by a 34 km wall, which is crossed by the Ceronda river at two opposite points. The enclosed part is dominated by broadleaf woodlands, permanent meadows and lakes, whereas agricultural fields and meadows prevail outside. LMNP is home to a rich fauna, including a diverse community of large mammals such as wild boar, red deer, fallow deer, roe deer and, more recently, wolves (Canis lupus italicus). The number of red and fallow deer is traditionally kept under control by park rangers in line with management guidelines approved by national and regional public agencies. According to the most recent estimates, there are approximately 200 red deer and 50 fallow deer. As the culled deer are sold as venison, their carcasses and viscera are inspected by a veterinary officer at a nearby slaughterhouse. Roe deer, still in low numbers (approximately 10 to 20), are fully protected within the NP. The LMNP borders two public hunting areas (ATCTO2 and CATO4), which are covered by a mix of broadleaf woodlands, meadows, and agricultural fields. Flat and slightly undulating terrain similar to that in LMNP is dominant in ATCTO2, whereas slopes are present in the portion of CATO4 bordering LMNP. Overall, the area of the investigated regions surrounding the LMNP totals 12,500 hectares. Figure 1 provides a global view of the location of the LMNP in relation to the hunting areas. Soil diversity is greater than in LMNP and includes clay, rock, gravel, sand and silt. Wild ruminants (wild boar, roe deer, red deer, northern chamois and Mediterranean mouflon) are harvested according to culling plans authorised by the regional public game agency on a yearly basis. The rules governing the hunting of wild ruminants in the two units stipulate that the culled game must be identified with a plastic band immediately after shooting and presented to technicians at dedicated check stations on the same day for administrative tasks and the collection of biometric data and biological samples.

2.2. Sampling of Potential DF Definitive Hosts

Between 2012 and 2013, and again between 2020 and 2023, the livers of 331 wild ruminants that were culled or found dead within a radius of 10 km around the centroid of the LMNP and bordering hunting units were examined externally for gross lesions that could be indicative of F. magna infestation (e.g., perihepatitis villosa, migration tracks, black pigmented streaks and pseudocysts filled with brown fluid). The livers were then cut into slices approximately 1 cm thick in order to search for adult and/or immature flukes. Morphological identification of the flukes was performed at species level according to Stiles (1895) [19] and Ward (1917) [20]. The sampled wildlife included red deer, fallow deer, roe deer, northern chamois and Mediterranean mouflon. No mouflon were harvested after 2015 due to a population decline following the establishment of one or two wolf packs in their focal distribution area.

Additionally, fluke eggs were searched for in faecal samples collected from the rectums of a subsample of infected wild ruminants using a sedimentation technique [21]. For samples testing positive, egg counts were subsequently performed using a modified McMaster technique with a saturated zinc sulfate solution with a specific gravity of 1.35 [22,23].

2.3. Development of a Model to Assess the Suitability of Areas for the Presence of DF Vectors and Facilitate Sampling Efforts

To optimise the search for DF vectors (G. truncatula and R. peregra), a predictive suitability map of their potential presence in the study area was created using QGIS software version 3.10.0, ‘A Coruña’ [24]. The risk map considered the following two variables, which, as indicated in the literature, can influence the presence of gastropods:

(i)

Soil permeability;

(ii)

Land use type. Soil permeability characteristics were derived from a regional geodatabase (https://www.regione.piemonte.it/web/temi/agricoltura/agroambiente-meteo-suoli/suoli-paesaggi-agrari-piemonte, accessed on 25 January 2025). The soil permeability classes were classified into three risk classes (high, medium, and low) in relation to the possible presence of gastropods, with associated numerical values ranging from 10 (low risk) to 30 (high risk). The type of land use was assessed using the same regional database. The land-use classes were then reclassified into three risk classes: low (agricultural areas and rivers—10 points), medium (forest areas—20 points) and high (predominantly pastoral areas—30 points). This was based on their suitability for hosting microhabitats favourable to F. magna vector gastropods. Urbanised areas were given value = 0. The risk map was generated by overlaying layers and summing the values for each of the two parameters. The risk of gastropod presence varies between 10 and 60. These values were then reclassified into three categories of total risk: low (10–20 points), medium (30–40 points) and high (50–60 points). The risk map was then subjected to field validation to verify its ability to accurately predict the presence and abundance of gastropods.

To validate the model, snails of the genus Lymnaea were searched for in the LMNP, both inside and outside the fenced area. In both zones, locations were randomly selected based on the identified DF vector suitability classes. Using a Garmin eTrex^® SE GPS (© 2023 Garmin Ltd., Olathe, KS, USA) navigation device, we identified all possible snail habitats (mostly ponds and irrigation channels) within a radius of 50 m and then we devoted a maximum of 5 min in each identified possible habitats, to searching for snails in the mud and vegetation of each habitat. Any snails measuring ≥4 mm in length were placed in vials containing ionised refrigerated water. The vials were labelled according to their location and the snails were examined morphologically under a stereomicroscope according to Jackiewicz (1998) [25]. Finally, each location was classified as vector positive or negative based on the presence or absence of G. truncatula and/or R. peregra. In each location, snails were collected in three sessions: June, August, and October.

2.4. Vector Analysis for DF DNA

Additional snails measuring at least 4 mm in length were collected in late spring and early summer from natural ponds and irrigation channels within and outside the fenced area of the LMNP. The snails were provisionally stored at −18 °C and then molecularly processed for the presence of F. magna DNA. The soft tissues of each snail were cut into small pieces (around 1–2 mm), after which genomic DNA was extracted using a NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany). Extraction was carried out according to the manufacturer’s instructions, with the DNA eluted with 30 μL of buffer BE instead of 100 μL to increase its concentration. For every 20 samples, one tube containing only reagents was used to monitor for contamination. F. magna genomic DNA was amplified using species-specific ITS1 primers designed by Králová-Hromadová et al. [7]. Briefly, each reaction consisted in 10-20 ng of the DNA, 20 pmol of each primer, FM_ITS1_SPEC_F, 5′-TGTC ATGCGATAAAAATGTTTT-3′ and FM_ITS1-SPEC_R, 5′-CTGGACCCGCGCCCGAAGGA-3′ (Sigma Aldrich, Milano, Italy), 0.2 mM of each of the deoxynucleotide triphosphate, 2.5 mM of MgCl₂ and 1 U of HotStar Taq (QIAGEN, Milano, Italy). The reaction volume was 25 μL. Samples were subjected to the following thermal profile for amplification in a 2720 thermal cycler (Applied Biosystems, Foster City, CA, USA): 15 min at 95 °C (initial denaturing), followed by 35 cycles of three steps of 30 s at 94 °C (denaturation), 45 s at 55 °C (annealing) and 1.5 min at 72 °C (extension), before a final elongation of 10 min at 72 °C. The PCR products were loaded on the 1.5% agarose gel stained with ethidium bromide for DNA visualisation under UV light.

2.5. Camera Trapping

LMNP has only a few permeable points along the rivers, where semi-permeable barriers have been installed. These structures do not entirely prevent animal movement or migration. At key locations—specifically the inbound and outbound sections of the Ceronda River—we deployed two camera traps at each gate. The camera traps (IR-PLUS model) recorded video footage both day and night, enabling us to monitor animal crossings at the barriers. Each camera was positioned at human height, facing the barrier and set to a wide-angle to capture a broad field of view and accurately document interactions between animals and the barriers. The cameras were configured to record one-minute videos with high sensitivity settings. Monitoring was conducted for one week per month over a two-year period, totaling 672 nights of observation (4 × 7 × 24).

2.6. Statistical Analysis

Statistical analyses and graphical representations were performed using R 4.3.0 [26]. F. magna prevalence and eggs count per gram of faeces were calculated. p-values below 0.05 were considered statistically significant. Spatial analysis was performed using QGIS software version 3.10.0, ‘A Coruña’ [24].

3. Results

3.1. Occurrence of Flukes in Wild Ruminant Hosts

Between 2012 and 2023, we examined the livers of 280 wild ruminants that were either sport hunted (n = 273) or killed by vehicles (n = 7) near the LMNP. Of these, 202 belonged to roe deer (100 from CATO4 and 102 from ATCTO2), 41 to Mediterranean mouflons, and 37 to northern chamois (all from CATO4; see Figure 2). This sample represented a significant proportion of the wild ruminants culled in the relevant hunting areas (56.7% in total). No liver tested positive for F. magna pseudocysts and/or gross lesions indicative of F. magna infestation. In parallel, we examined the livers of 32 red deer and nine fallow deer, as well as the entire carcasses of ten roe deer, all of which originated from the LMNP. The red and fallow deer were culled into the fenced area as part of herd reduction programmes, whereas the roe deer were found dead (four in the fenced area and six outside). Mature and immature flukes belonging to F. magna were identified in the majority of red deer (n = 25; 78.1%) and fallow deer (n = 7; 77.7%), with the only exception being fawns aged less than six months. Fluke eggs were observed in 13 of 15 infected red deer and 3 of 3 infected fallow deer, with an intensity of 93.3 ± 77.1 range 0–300) and 133.3 ± 104.1 (50–250) EPG. A F. magna infestation was also recorded in three roe deer from the fenced area: two adult does in 2012 and 2019, and an adult buck in 2020. The first infested roe deer was in poor body condition. On opening the abdomen, a diffuse peritonitis with dark pigmented exudate was visible. The liver was adhering to the diaphragm and was enlarged, firm and dark red with a few nodules protruding from it, measuring 2–4 cm in diameter. The liver capsule was opaque and irregular (perihepatitis villosa), spotted with dark iron-porphyrin pigment. Upon cutting, fibrotic, dark-stained fluke migratory tracks and several pseudocysts measuring 2–6 cm in diameter were observed. The latter were filled with dark brown fluid, from which 34 flukes were collected, mostly adults measuring 4.5–6.5 cm. The regional lymph nodes were enlarged, with a dark-stained cutting surface. Numerous F. magna eggs were found and counted in the roe deer faeces (650 EPG). The other two infested roe deer also exhibited poor body condition, peritonitis and similar liver lesions, as well as the presence of adult flukes in pseudocysts. Fluke intensity was not recorded for these two roe deer. The doe had subacute catarrhal pneumonia with hardening of the apical lobes.

3.2. Development of a Suitability Model for the Presence of DF Vectors

Figure 2 illustrates the distribution of high, medium and low risk areas for the presence of DF vectors, both within and outside the fenced area of the LMNP. As expected, high-risk areas were more prevalent within the fenced area, which corresponds to the historical focal point of the disease.

Using the risk map, vectors for F. magna were searched for at 30 sampling points (10 for each risk area). Within each sampling area, the presence of an environment suitable for the development of the intermediate host was investigated. In some areas, several suitable humid environments were found and were therefore sampled to investigate the presence or absence of gastropods (in other words more than one sampling point was investigated in each area), while in others no suitable environments were found. Using the predictive models to do sampling on the field, we identified 29 environments in low-risk areas with a total absence of gastropods; 28 humid environments in medium-risk areas with the actual presence of gastropods in 5 cases; and 37 humid environments in high-risk areas with the presence of gastropods in 10 points. The number of positive and negative sampling points for the presence of gastropods by risk areas is reported in

Table 1. Presence/absence and prevalence of gastropods in different risk areas.

	High Risk	Medium Risk	Low Risk
Presence	10	5	0
Absence	27	23	29
Gastropods prevalence	27%	18%	0%

.

3.3. Vector Analysis for DF DNA

We collected 226 G. truncatula specimens from suitable areas in the unfenced portion of LMNP. Inside the enclosed part, we collected a control sample of 141 G. truncatula specimens. No F. magna DNA was found in any of the potential intermediate hosts collected outside the enclosed part of the park. However, 13 positive samples were recorded in the enclosed part, giving a prevalence of 9.2% (95% CI: 5.4–15.1).

3.4. Inflow and Outflow of Roe, Red and Fallow Deer in the LMNP

A total of 218 crossings in front of the trail cameras were recorded using camera trapping. Figure 3 shows the number of crossings broken down by species.

Foxes (58%) and wild boar (19%) accounted for most of the movements in front of the cameras. Roe deer accounted for 3% of the total number of video/images. Of the species that can act as definitive hosts for F. magna, less than 6% of crossings were recorded.

When considering only species that can act as definitive hosts for the parasite, roe deer accounted for half of all videos and images (Figure 4). Of the seven roe deer passages, two were clearly identified as moving from inside to outside the LMNP.

4. Discussion

The results of this analysis suggest that, even if F. magna was introduced to our study area in the second half of the 19th century, the parasite has not been able to ‘escape’ the area of its original introduction (La Mandria Natural Park). Our results confirmed that the parasite is still circulating at a low prevalence in intermediate hosts in LMNP, indicating that the parasite’s life cycle is being maintained. Although the developed risk model has highlighted suitable areas for the parasite’s life cycle and its predictive power has been confirmed by detecting intermediate hosts in areas classified as high and medium risk, there is evidence that the parasite is absent outside the study area at both definitive and intermediate host levels. In fact, none of the 280 wild ungulates hunted outside the study area or any of the 226 gastropods were found to be infected. However, it is difficult to explain why the parasite has not spread outside the infected areas, given that camera traps observed potentially infected ungulates moving both inside and outside the park (approximately 6% of animals crossed the park boundaries) and roe deer have recently been proven to be competent hosts for F. magna infestations [27].

In our study, the risk map was effective in identifying areas in the environment that were suitable for intermediate hosts, and thus areas at risk of potential parasite spread. Although our risk maps are based on parameters such as proximity to water sources, high humidity and grazing areas, other studies [28,29] emphasise the importance of incorporating additional factors, such as precipitation and terrain slope. Another potentially important parameter in risk mapping is the floodability of the land; however, in our study area, this seems ineffective due to its geological conditions. Depending on the natural morphology of the landscape, other types of parameters could improve the identification of intermediate hosts on a larger scale (e.g., in northwestern Italy). This would help to prevent massive parasite infestations, such as outbreaks of Fasciola hepatica in livestock production systems [29].

The number and prevalence of gastropods collected in our study was lower than that of snails found in the Danube floodplain [5], with differences observed in the types of marsh. In our study, almost all intermediate hosts were found in irrigation canals used for agricultural purposes, rather than in natural ponds formed by rivers such as the Danube [30]. Many swamps providing optimal conditions for the development of intermediate hosts were found outside the park, where a considerable number of snails were also recorded. By contrast, the number of natural aquatic environments inside the park is significantly higher than in the surrounding areas where this study was conducted. None of the snails collected outside the fenced part of LMNP in the present study tested positive for intermediate forms of F. magna. Although the number of gastropods tested is limited, thus requiring the resulting prevalence estimates to be interpreted with caution, the data remain relevant for the surrounding area and support the robustness of our molecular findings, which provide evidence of the absence of trematodes in intermediate hosts outside the LMNP [31,32].

Despite the limited number of ungulates recorded, the observation of potentially infected animals moving in and out of the park was confirmed, which can represent a potential cause for a new outbreak. The animals detected by the traps were roe deer, red deer and fallow deer, which highlights the possibility that definitive hosts could spread the giant liver fluke beyond the LMNP fence. The steady increase in roe deer numbers in the area could result in the parasite being carried outside the park by a new host in the future, as increasingly reported in other endemic areas [33]. Furthermore, red deer movement across the river was low due to movable fences that the deer found difficult to navigate. This suggests that, at present, red deer may not be effective in spreading DF eggs outside the park. Therefore, the presence of different host species in an area does not demonstrate stable colonisation, as occurred in the Danube floodplains [5,34]. Similarly, the detection of environmental DNA (eDNA) outside the park boundaries cannot be considered conclusive evidence of parasite establishment, as eDNA traces may result from transient hosts, water flow transport or environmental persistence, rather than an active, self-sustaining parasite population [18].

In conclusion, the likelihood of parasites spreading in our study area is very low for the following reasons: (i) only a limited number of definitive hosts escape from the fenced area and release parasite eggs outside the enclosure; (ii) there are fewer potentially infected intermediate hosts in irrigation canals than in natural forest swamps; (iii) swamps are less numerous outside the park than inside, and are located in areas with different land use; (iv) agricultural habitats do not provide the same natural environmental conditions as forest habitats for ungulate definitive hosts, resulting in limited spatial overlap with intermediate host populations.

The main consequence of these differences between inside and outside the park is that it is currently impossible to guarantee the parasite’s full life cycle outside the LMNP. While an encounter between eggs, intermediate hosts and definitive hosts is potentially possible, it is very unlikely. In the future, the re-naturalisation of these areas on a large scale could lead to an increase in the deer population, which would provide more opportunities for them to act as vectors for giant liver fluke. Conversely, a decrease in agricultural activity could create a more suitable natural environment for intermediate hosts.

The first step to reducing the risk of parasite escape should be to restrict the movement of the definitive host by installing new barriers. To protect the area surrounding the river and prevent access to definitive hosts, the installation of an electrified fence could also be considered. At the same time, passive surveillance of definitive and intermediate hosts must be maintained to identify and verify any positive cases. Monitoring at three levels—final hosts, intermediate hosts and suitable environments—is the right way to keep DF enclosed in LMNP.

Further studies are needed to understand the key risk factors associated with the spread of F. magna in non-native areas, where the role of host species as maintenance or sentinel hosts may differ from that of North American ungulates. Ungulate management plans around the LMNP should take into account the importance of active and passive surveillance for monitoring giant liver fluke, with the continuous development of the epidemiological situation involving both wildlife and health authorities. In conclusion, this study highlights the crucial role of an integrated, multidisciplinary approach involving geographical data, hunting management plans, the support of pathologists and parasitologists, and molecular analysis to monitor parasitic diseases in wildlife. This approach has already been demonstrated to be effective in other studies [35].