Helminth infections in Italian donkeys: Strongylus vulgaris more common than Dictyocaulus arnfieldi

Abstract Donkeys have been used as working animals for transport and farm activities worldwide. Recently, in European countries, there has been an increasing interest in donkeys due to their use as pets, onotherapy or milk production. During 2014–2016, a countrywide survey was conducted to determine prevalence and risk factors of principal helminth infections in 1775 donkeys in 77 Italian farms. A questionnaire on management and parasite control practices was filled out for each farm. Faecal samples were examined using a modified McMaster technique, a centrifugation/flotation method and a sedimentation technique. Pooled coprocultures were performed for differentiation of strongylid eggs. Strongyles were the most common parasites detected (84.9%), followed by Dictyocaulus arnfieldi (6.9%), Oxyuris equi (5.8%), Parascaris spp. (3.6%), Anoplocephala spp. (1.0%), Strongyloides westeri (0.3%). Coprocultures revealed an omnipresence of cyathostomins (100%), followed by Strongylus vulgaris (31.0%), Poteriostomum spp. (25.0%), Triodontophorus spp. (9.0%), Strongylus edentatus (7.0%), Strongylus equinus (5.0%). Logistic regression analysis identified breed, co-pasture with horses, living area, herd size and number of treatments as significantly associated with strongyles. Sex, age, living area and herd size were significantly associated with Parascaris spp. Dictyocaulus arnfieldi was significantly associated with sex, grass, co-pasture with horses, living area and herd size. Strongylus vulgaris was significantly associated with living area and herd size. The mean number of anthelmintic treatments/year was 1.4; most of the donkeys (71.8%) were dewormed using an ivermectin drug. It is important to design parasite programs to specifically address both D. arnfieldi and S. vulgaris in donkeys, and this is especially important if donkeys co-graze with horses.


Introduction
Donkeys (Equus asinus) are traditionally used as working animals for transportation and farm activities in developing countries (Ali et al., 2014). In Europe, donkey populations have declined from the 1960s and nowadays, according to the Food and Agriculture Organization, the population is estimated to be about 395,910 animals (FAO, 2020). In Italy, the donkey population has gradually decreased due to industrialization and urbanization from 500,000 in the 1960s to about 23,000 in the early 2000s (FAO, 2020).
In recent years, donkeys have found increasing use in leisure activities, onotherapy (animalassisted therapy) and milk production (Veneziano et al., 2011). This has caused an increase of the Italian donkey population to currently 92,965 donkeys (AIA, 2020). In Italy, there is a long-established tradition in donkey breeding as confirmed by the presence of several autochthonous breeds such as Amiatino, Asinara, Martina Franca, Ragusano, Romagnolo, Pantesco, Sardo, Grigio Siciliano and Viterbese (Bigi & Zanon, 2008). However, despite the economic and sanitary relevance of parasitosis in farm animals, there are studies on ectoparasites and protozoa, but limited epidemiological data regarding helminths affecting this species available in the scientific literature (Matthews & Burden, 2013).
Donkeys represent the main competent host and reservoir for D. arnfieldi and the major source of pasture contamination for horses (Beelitz et al., 1996), although infection from horse to horse is reported (Matthews, 2002). Clinical signs in donkeys are quite rare, but in horses, tachypnea and persistent cough could be present (Matthews, 2002).
Moreover, horses and donkeys can be parasitized by the liver fluke, Fasciola hepatica, in wetter areas (Matthews & Burden, 2013). Although massive parasitic infections are often subclinical in donkeys, the impact of endoparasites on their health, welfare and production is still unclear (Buono et al., 2018). Generally, donkeys with a high parasite load of small strongyles appear healthy, and mainly overworked animals or those with poor physical status show clinical signs, as reported in developing countries where substandard environmental conditions exist (Burden et al., 2010;Matthews & Burden, 2013;Tavassoli et al., 2015). Regarding large strongyles, in an eight-year-old donkey, S. vulgaris caused obstruction of the cranial mesenteric artery, which led to death of the animal (Borji et al., 2014).
Anthelmintic therapy represents a main strategy to control endoparasitosis in horses , as well as in donkeys (Buono et al., 2018); however, only few anthelmintic drugs are registered for use in donkeys, so extra-label use of products registered for horses and ruminants is the norm (Gokbulut & McKellar, 2018).
Previously in Italy, only three reports on a small number of donkeys (approximately 100 animals) were available about helminth infections (Giannetto et al., 2008;Trentini et al., 2010;Garippa et al., 2016). In Europe, robust parasitological studies have been carried out by The Donkey Sanctuary (UK) but did not aim to conduct an epidemiological study on a national level (Matthews & Burden, 2013).
The aims of the present survey were to: (1) determine the prevalence and distribution of main helminth infections in Italian donkeys; (2) investigate the risk factors associated with infection; and (3) describe current parasite control practices adopted for donkeys.

Study animals
The study was performed during 2014-2016 and included 1775 donkeys raised on 77 farms located in 13 Italian regions (nine farms from northern Italy; 28 from central Italy; 26 from southern Italy; 14 from islands). The sample size was calculated using the formula proposed by Thrusfield (2007), including the following information: study population at the time of the start of the study (58,647 donkeys; data supplied by Italian Breeders Association -AIA, 2013); expected prevalence (80%) based on the results of a regional study on intestinal strongyles in donkeys conducted in Sicily, southern Italy (Giannetto et al., 2008); desired absolute precision (3%); and confidence interval (CI) (99%).
The age of donkeys was determined matching owners' information with examination of teeth, according to the guide proposed by The Donkey Sanctuary (2016). The animals were classified into four age groups: ≤1 year; 1-4 years; 4-10 years; and >10 years. Furthermore, body condition scores (BCS) were determined using the chart developed by The Donkey Sanctuary (2008). Briefly, a scale of 1-5 was applied (1 = poor, 2 = moderate, 3 = ideal, 4 = fat and 5 = obese).
All procedures on the donkeys were performed with the owner's consent and according to the European Communities Council Directive (86/609/EEC). The investigation was approved by the Ethical Animal Care and Use Committee of the University of Naples 'Federico II' number 889/04.

Questionnaires and data collection
On each farm, owners were interviewed using a purpose-designed questionnaire to obtain a complete history of the farms and examined animals. The questionnaire consisted of 36 closed-ended questions (multiple choice and yes/no options) and it was divided into three sections (general information, grazing practices and farm management, helminth control procedures). The questionnaire can be found in the supplementary material.
The second section included information about farm management: availability of box stalls and access to grazing (yes or no); months of grazing per year; and co-grazing with horses (yes or no) or ruminants (yes or no).

Coprological analysis
Individual fecal samples were collected from all donkeys included in the study (1775) and, according to general recommendations proposed by Nielsen et al. (2010a), faeces were taken directly from the rectum of each animal, or alternatively picked up off the ground from fresh deposits (especially from nervous animals and foals) using plastic gloves. Fecal samples were stored in portable refrigerators (about 4°C) and individual fecal egg counts (FECs) were performed within 48 h, using a special modification of the McMaster method with a detection limit of ten eggs per gram (EPG) (Zajac & Conboy, 2011). The floatation medium used was the Sheather's sugar solution with a specific gravity of 1.250 (Lester & Matthews, 2014). Based on visual morphological identification (Zajac & Conboy, 2011), each egg was classified as belonging to intestinal strongyles, Parascaris spp., S. westeri, O. equi, Anoplocephala spp. Furthermore, a qualitative centrifugation/flotation technique (Proudman & Edwards, 1992) using 2 F. Buono et al.
Sheather's sugar solution and a qualitative sedimentation technique (Ambrosi, 1991) were used for the detection of Anoplocephala spp. and F. hepatica eggs, respectively.

Fecal cultures and speciation of intestinal strongyle
Pooled fecal cultures (from a maximum of five donkeys) were set up for each farm. Briefly, 20 g of faeces from samples with FECs greater than the suggested 300 EPG selective therapy cut-off (Matthews & Burden, 2013) were used and then incubated at 27-30°C for 7-10 days for larval development. At the end of the incubation period, third-stage larvae (L3s) were recovered by the Baermann technique (MAFF, 1986) and identified according to their specific genera and/or species, based on the shape and number of gut cells, using the keys proposed by the Atlas of Diagnosis of Equine Strongylidosis (Cernea et al., 2008). Where possible, up to 100 larvae (L3s) were identified.

Baermann technique and diagnosis of D. arnfieldi
For each donkey, individual Baermann technique was also used to ascertain the occurrence of lung worm infection on 30 g of faeces, in a double layer of gauze, closed with an elastic band and placed in a funnel (Zajac & Conboy, 2011). The funnel was filled with tap water until the faeces were immersed. After 24 h, the content of each Baermann was collected in 15 ml tubes and centrifuged at 400 g for 5 min. The supernatant was discarded, and 1 ml of each pellet was transferred to nematode slides and examined under microscope.

Statistical analysis
Animal-level prevalence was computed with the associated 95% CIs for all the considered parasites (intestinal strongyles, Parascaris spp. and D. arnfieldi) with the exception of S. vulgaris, whose prevalence was estimated at farm level. All the answers from the questionnaires were included as independent variables in a multivariate logistic regression analysis developed by a stepwise regression process to explore factors that could be considered predictors for positive status (>0 EPG) for intestinal strongyles, Parascaris spp. and D. arnfieldi (>0 larvae per gram) at the animal level (sex, breed, presence of pasture, co-pasture with ruminants, co-grazing with horses, BCS, age, living area, herd size and number of treatments/year) and S. vulgaris at the farm level (presence of pasture, co-pasture with ruminants, co-pasture with horses, presence of young stock (donkeys <1 year) living area, herd size and number of treatments/year). Results were presented as adjusted odds ratios. All the statistical analyses were performed using commercial software (SPSS, Version 22.0, Chicago, IL, USA).

Risk factor analyses
Results of logistic regression analysis for risk factors associated with patent infections of intestinal strongyles, Parascaris spp. and D. arnfieldi are summarized in tables 2-4. The results of the analysis at farm level of the presence of S. vulgaris are summarized in table 5. In the logistic regression analysis of intestinal strongyles, breed, co-pasture with horses, living area, herd size and number of treatments were significantly associated with the outcome variables (table 2).
In the logistic regression analysis of Parascaris spp., sex, age, living area and herd size were significantly associated with the outcome variables (table 3).
In the logistic regression analysis of D. arnfieldi, sex, grass, co-pasture with horses, living area and herd size were significantly associated with the outcome variables (table 4).
In the logistic regression analysis of S. vulgaris, living area and herd size were significantly associated with the outcome variable (table 5).

Discussion
The present study is the first European countrywide survey providing data on parasite dynamics in donkeys and provides   (Veneziano et al., 2015), where deworming was carried out by almost all owners (94.0%) and the frequency was mainly two or three times/year (68.6%) and comparable with other equine surveys (Schneider et al., 2014;Nielsen et al., 2018;Tzelos et al., 2019). Moreover, MLs (mainly IVM) were the most common drugs used to control endoparasites in Italian donkey farms, similarly to those used in horses (Veneziano et al., 2015;Becher et al., 2018;Nielsen et al., 2018;Wilkes et al., 2020). This supports the theory that deworming in donkeys is commonly performed with the same drugs and dosages used in horses (Gokbulut & McKellar, 2018). In Italy, given that a sizeable proportion of donkeys are raised mainly for milk production for human consumption, the use of MLs (IVM and MOX) should be avoided in lactating animals given the lack of established withdrawal time for milk in equids (Gokbulut & McKellar, 2018).  It is important to keep in mind that donkeys are not small horses with long ears, as they are characterized by a greater level of activity of some P450 isoenzymes than horses (Lizzaraga et al., 2004). Thus, for certain drugs, the recommended dose and posology for horses may not be appropriate for donkeys (Grosenbaugh et al., 2011). Regarding anthelmintic drugs, some studies confirm the use of horse dosage in donkeys, despite the apparent differences in intestinal absorption and plasma disposition (Gokbulut & McKellar, 2018). PYR pamoate in oral paste and granule formulations administered to donkeys at the horse dose rates, showed different bioavailability and persistence compared to horses, but were effective in controlling intestinal strongyle infections (Gokbulut et al., 2014). In donkeys, mebendazole was reported to be highly effective against small strongyles both at the horse dose rate and at double the horse dosage, and egg reappearance period was five weeks for both dosages (Gokbulut et al., 2016). In donkeys, a recent  trial showed that in well-managed farms, anthelmintic drugs administered at the horse dose rate were generally effective in the control of small strongyle infections (Buono et al., 2018). Based on the results of these studies, the horse dosage for some anthelmintics may also be effective for the donkey dosage. In the present survey, owners and veterinarians generally treated donkeys like horses.
Intestinal strongyle infections are common in grazing horses and donkeys; however, in this survey, access to pasture was not associated with a higher prevalence of positive strongyle FECs. A similar finding was made in a previously published equine study (von Samson-Himmelstjerna et al., 2009). Donkeys raised in box stalls with an annexed paddock showed high intestinal strongyle prevalence, and it is possible that parasites could be transmitted in stalls in cases of suboptimal hygiene (Nielsen et al., 2010b). Moreover, it should keep in mind that encysted and arrested larvae will gradually develop into adult parasites, causing positive FECs over time, and several anthelmintic treatments could be needed to achieve a negative FEC.
In the examined farms, donkeys generally grazed with a low number of ruminants, and the ratio between these two species was very low. This would explain why a 'dilution effect' was not observed and why there were no significant differences in strongyle egg excretion between donkeys that co-grazed and those that did not co-graze with ruminants.
Donkeys raised in farms in northern Italy had a significantly higher prevalence of strongyle infection. This is probably due to the climatic difference between northern and central-southern Italy. Pasture contamination with small strongyles is widely influenced by climatic conditions (Leathwick et al., 2015); in northern Italy, rainfalls are more abundant than central-southern Italy and this would allow a greater migration of the L3 on the pastures.
Small (≤25 animals) and large farms (≥50 animals) were more associated with strongyle infection than medium-sized farms (25-50 donkeys). Similar findings have been reported in horses (Kornaś et al., 2010). In small farms, donkeys are usually housed with horses and are often neglected and just sheltered into the farms without being involved in management and anthelmintic control practices, whereas in larger farms a high stocking density could be correlated to higher infection pressure.
The distribution of strongyle egg count levels reported herein are different from those reported in managed horses, where an overwhelming majority (50-75%) is in the low contaminator category (Relf et al., 2013;Lester et al., 2018). Consequently, the 80:20 distribution rule that is often cited for horses (Becher et al., 2010;Nielsen et al., 2018), according to which 20% of animals shed about 80% of intestinal strongyle eggs, does not appear directly applicable for this donkey population. Here, about 40% of adult donkeys were shedding approximately 80% of the eggs, suggesting an 80:40 distribution ( fig. 3). In this study, 770 donkeys (43.4%) had <300 EPG intestinal strongyles, and, thus, these animals may not require deworming.
Faecal cultures documented a wide biodiversity with a high prevalence of large strongyles, most notably S. vulgaris. Given the distinct pathogenicity of this parasite (Pihl et al., 2018), more attention should be given to controlling it in donkey operations. It was reported that S. vulgaris caused obstruction of the cranial mesenteric artery, which led to death of the animal in an eight-year-old donkey (Borji et al., 2014). Moreover, in a two-year-old intact male donkey, aberrant migrations of S. vulgaris through the spinal cord resulted in progressive paraparesis and tetraplegia (Mayhew et al., 1984). The farm-level analysis showed a significant association between S. vulgaris and location, suggesting that donkeys raised on farms located on the island of Sicily are less exposed to S. vulgaris infection. The reason for this difference is not clear, but climatic influence and management practices may play a role.
The low anthelmintic treatment intensity reported in this study could be a risk factor for S. vulgaris infection in donkeys as also demonstrated in horses (Nielsen et al., 2012;Hedberg-Alm et al., 2020). Considering the high prevalence of S. vulgaris in the donkey population and that many donkeys co-graze with horses, it is necessary to assess the presence of large strongyles in donkey species to avoid the risk of infection in horses. Thus, if coprocultures show the presence of S. vulgaris, it is important to also treat the low contaminators (FEC <300 EPG) to limit its spread on the pasture. It is also important to highlight that IVM treatment may not be completely effective against S. vulgaris migrating fifth-stage larvae . Although horses demonstrate different degrees of immunity to Parascaris spp. in relation to age and exposure with patent infections (Clayton & Duncan, 1979), adult donkeys can harbour ascarids without clinical signs, and could be playing an important role in pasture contamination (Matthews & Burden, 2013). The reason for female donkeys having a lower ascarid prevalence than males is unclear, but could be due to differences in management strategies. A large proportion of females were kept for milk production purposes, whereas males were used for a variety of purposes, including meat production. Consequently, males and females could be managed substantially differently, which could affect the risk of acquiring patent ascarid infection. Although ascarid eggs were found in faeces of all age groups, younger donkeys (≤1 year) had a significantly higher prevalence than other age groups. This contrasts findings among working donkeys in Ethiopia, where a high prevalence (>50%) regardless of the animal age was reported (Getachew et al., 2008). This discrepancy could be explained by the inability of the animals to develop immunity against ascarids due to poor farm management and poor nutrition for working equids in developing countries (Getachew et al., 2010). Donkeys on farms located in southern Italy had significantly higher Parascaris spp. prevalence, and this can be explained considering that in these farms donkeys were kept mainly for milk production and, therefore, characterized by a greater presence of animals aged less than one year.
Donkeys and mules are natural hosts for D. arnfieldi, whereas horses are primarily at risk of getting infected while co-grazing with these animals. Although the prevalence was relatively low in this study, it could be underestimated considering that the firststage larvae (L1) recovery of D. arnfieldi by Baermann technique is strongly influenced by the method of storage of faecal samples. Considering that faecal samples were stored in portable refrigerators, temperature fluctuations during the first 48 h could adversely affect the parasitological examination (Rode & Jorgensen, 1989). Female donkeys were at higher risk of infection than male and gelding, probably due to the fact that they spent more time grazing. Donkeys that had access to pasture were at major risk of infection by D. arnfieldi, suggesting that animals acquire infections when grazing on a contaminated pasture (Matthews, 2002). Donkeys co-grazing with horses had a lower prevalence of D. arnfieldi, which is probably due to the dilution principle, considering that horses rarely harbour adult stages and considering that horses are not permissive hosts of the full life cycle, they generally do not excrete larvae in the faeces (Matthews & Burden, 2013). The highest prevalence was reported in animals aged less than one year and in animals with BCS<3. This makes sense as young age and poor BCS represent potential risk factors, probably related to impaired immunity to infections (Tihitna et al., 2007).
It is advisable to perform diagnosis for lungworm infection in all donkeys co-grazing with horses, and the animals should be treated using MLs to reduce pasture contamination by L1 of D. arnfieldi (Matthews, 2002). In our study, IVM was the compound mostly used for deworming and this could explain the low prevalence of lungworm infections. Moreover, the frequency of anthelmintic treatments was associated with a lower infection rate, suggesting that the routine deworming scheme in the donkey farms could be correlated to a reduction of lungworm parasitic load.
The prevalence of Anoplocephala spp. was low (1.0%) but in agreement with data reported in donkeys in few studies on the presence of tapeworms in Europe (Matthews & Burden, 2013). In horses, Anoplocephala perfoliata was detected with higher prevalence (Tzelos et al., 2019;Hedberg-Alm et al., 2020). This low prevalence could be correlated with the irregular excretion of tapeworm eggs in the faeces (Gasser et al., 2005). Furthermore, considering that the Proudman test can result in false negatives if the parasitic load is less than 20 worms (Proudman & Edwards, 1992), it is possible that tapeworm infection was not diagnosed in those animals with a low parasitic load.
In Europe, F. hepatica is widespread in ruminants, but with a clustered spatial distribution, and regional variation in epidemiology and prevalence (4% in southern Italy) (Beesley et al., 2017). In Italy, the low prevalence of this liver fluke could justify its absence in the studied donkey population.
In conclusion, this study represents the first large-scale epidemiological survey in Europe on the prevalence of helminth infections in donkeys, adding useful data to the scarce literature available about the parasite dynamics in this neglected animal. Our findings revealed the presence of several helminth species, representing the most important pathogenic parasites of equids and donkeys worldwide. Considerations for parasite control in donkeys could be different to those suggested in horses, because large strongyles in donkeys are more common than in horses. Furthermore, in donkeys the 80:20 strongyle egg shedding  (1775), 1005 animals (56.6%) were above the red threshold line, with FEC > 300 eggs per gram (EPG), while 770 (43.4%) were below the red threshold line, with FEC < 300 EPG.
distribution rule is not applicable, because about 40% of donkeys were shedding around 80% of the eggs. Furthermore, this study highlighted the prominent occurrence of S. vulgaris in managed donkeys and a farm prevalence exceeding that of D. arnfieldi. Thus, parasite control programs in donkeys require specific consideration of strongyle egg shedding patterns as well as occurrence of individual parasite species with pathogenic potential for both the donkeys and possible co-grazing horses.
Acknowledgements. This work would not be possible without the willingness of the donkey owners who agreed to participate in the survey and Il Rifugio degli Asinelli (The Donkey Sanctuary branch) that supported the research.
Author contributions. VV, BF, PL and RC contributed conception and design of the study. BF, FV, PL, NE, SG, NB, MU and ZSA collected the samples, performed the laboratory activities and organized the database. BF, PD and VV wrote the manuscript. VF performed statistical analysis. MKN critically oversaw substantial revisions of the manuscript. All of the authors revised the manuscript and approved the submitted version.
Financial support. This study was partially funded by grants from the Ministry of Health of the Italian Republic (IZSME 14/11 RC and IZSLT 11/ 15 RC, IZSLT 9/18 RC).

Conflicts of interest.
None.