THE PARASITE AND ITS LIFE-CYCLE

Adults of A. perfoliata are ~5–8 cm in length and ~1·2 cm in width. The scolex of the worm is unusually large, 2–3 mm in diameter and has 4 large suckers, each with a muscular wall 0·14 mm wide (Spasskii, 1951). Four projections, known as lappets (whose function is not understood), protrude from the base of the scolex. Baer (1927) and Spasskii (1951) provided descriptions of the species for taxonomic purposes, while Kahane (1880), Becker (1921, 1922a,b) and Lee & Tatchell (1964) have provided detailed descriptions, including histological observations. More recently, Schuster (1991) undertook a morphometric analysis of an A. perfoliata population.

The embryonated anoplocephalid egg is passed via the faeces into the environment, where it can survive on pasture for up to 9 months (Dunn, 1978). The egg is ingested by an intermediate host, a pasture-dwelling oribatid mite of the families Carabodidae, Galumnidae and Oribatulidae (Bashkirova, 1941; Romero et al. 1989; Denegri, 1993; Denegri et al. 1998). Mechanical action of the mouth-parts of the intermediate host leads to the emergence of the oncosphere from the egg membrane. Activation of the oncosphere has not been studied in A. perfoliata but in Hymenolepis diminuta it appears to be stimulated by ions present in the gut of the intermediate host (Lethbridge, 1972), enabling the oncosphere to use its hooks to tear the oncospheral membrane. It penetrates the gut wall of the intermediate host using muscular activity, its hooks and secretions from penetration glands. The oncosphere then undergoes development to the cysticercoid (=metacestode stage) over a period of 8–20 weeks. The cysticercoid contains a retracted scolex and a tail (cercomer) with 3 pairs of oncospheral hooks (Bashkirova, 1941).

Horses acquire A. perfoliata infection by ingesting the infected mite. The processes involved in the development of the cysticercoid to the adult in the horse have not been studied, but it is likely that they are similar to those of other cestodes. Mastication and digestion of the mite releases the cysticercoid, which then undergoes 2 processes: (1) eversion of the scolex, involving orientation of the suckers in preparation for attachment to the intestinal mucosa and (2) activation, the initiation of muscular activity, allowing the worm to move and attach to the intestinal wall. These events in other cestodes are stimulated by the activity of a combination of enzymes and bile salts (cf. Smyth & McManus, 1989), which digest parts of the cysticercoid, leaving the suckers and the neck region. The neck begins to differentiate and produce proglottides (=segments) that, as they mature, become gravid and detach to be passed in the faeces. The pre-patent period of the parasite is ~6–10 weeks (Arundel, 1985).

PREVALENCE OF INFECTION

Numerous studies in various countries have been undertaken to determine the prevalence of A. perfoliata infection. Prevalence has been determined predominantly at autopsy, but ante mortem diagnostic approaches have also been used. These include treatment with pyrantel pamoate with subsequent faecal examination for expelled worms and faecal flotation. Of the three techniques, autopsy is the most reliable.

Early reports indicated that prevalence varied considerably. Neumann (1892) reported that the parasite was not common in France but was more common in Germany and Russia. He cited a prevalence of 28% in Copenhagen, and Spasskii (1951) reported prevalences of up to 97% in Russia. In the USA, multiple autopsy studies have been carried out. For example, 2 studies in consecutive years in Kentucky gave similar results: a prevalence of 54% in 1983 (Lyons et al. 1983) and 53% in 1984 (Lyons et al. 1984). Reinemeyer et al. (1984) recorded a prevalence of 18% in Ohio and Torbert et al. (1986) found that 47% of horses in Florida (with little exposure to anthelmintics) were infected with A. perfoliata. Bello (1979) in North Carolina used a faecal flotation method and reported a prevalence of 13% in 192 horses, and Slocombe (1979) in Canada, also using faecal flotation, recorded a similar prevalence (13·6%) in 580 horses. In the UK, higher prevalences of infection with A. perfoliata have been reported. For instance, Owen et al. (1988) recorded a prevalence of 69% in 103 horses in Wales and adjacent parts of England, while a study by Pearson et al. (1993) in the south west of England found a prevalence of 80% in 20 horses. Another study (Fogarty et al. 1994) found 51% of 363 horses slaughtered at an Irish abattoir to be infected. Imrie & Jacobs (1987) treated 80 horses from North London and South Hertfordshire with pyrantel pamoate and examined the faeces for expelled worms. A prevalence of 27·5% was reported. More recently, Nilsson et al. (1995) undertook a detailed abattoir survey in central Sweden and showed 65% of 470 horses to be infected with A. perfoliata (mean intensity: 79 worms per horse, with a maximum of >900 worms). Of the horses examined, approximately half had 1–100 worms, whereas 14% were infected with >100 worms; 72% of horses harboured adult and juvenile worms, and 20% and 8% harboured only juveniles and adults, respectively.

In Australia, autopsy studies conducted in the state of Victoria (Bucknell et al. 1995; Williamson et al. 1997) have reported prevalences of A. perfoliata to be 29% of 150 horses and 38·5% of 130 horses, respectively. Studies in other states of Australia have recorded prevalences of 62% in 138 horses from Queensland (English, 1979), 4·9% in 140 horses from Perth (Dunsmore & Jue Sue, 1985) and 32% in 57 horses from north Queensland (Mfitilodze & Hutchinson, 1989), whereas in New Zealand, Bain & Kelly (1977) reported a prevalence of 81·5% in 65 horses.

PREDILECTION SITES AND DISTRIBUTION

Anoplocephala perfoliata is found exclusively in the intestines of equids, having been reported in domestic horses (Equus caballus), donkeys (Equus asinus), Burchell's zebra (Equus burchelli) (Round, 1968) and zebras (species not stated) (Schmidt, 1986). While the ileocaecal junction has been recognized as the main predilection site of the parasite, it can be found in other regions, such as the ileum, caecum and colon. For instance, Owen et al. (1988) reported that 17% of 103 horses in England and Wales had tapeworms attached in the ileum and 10% in the colon. Horses with worms in these sites had intensities of infection (using the terminology of Margolis et al. 1982) higher than the mean intensity of infection of the entire sample, supporting a proposal by Lyons et al. (1986) of an association between heavy cestode burdens and the presence of tapeworms in regions other than the ileocaecal junction. Also, Fogarty et al. (1994) reported that of 184 horses infected with A. perfoliata, 91% had tapeworms attached to the caecal wall, whereas only 53% had worms at the ileocaecal junction, suggesting that a greater proportion of worms may attach in the caecum than previously thought. These findings and proposals were supported by a more recent study by Williamson et al. (1997) who reported a high percentage (81·1%) of A. perfoliata on the caecal wall compared with lower numbers (17%) in the ileocaecal junction, terminal ileum (1·7%) and ventral colon (0·2%).

PATHOGENIC EFFECTS AND ASPECTS OF THE PATHOGENESIS

Early reports of pathogenicity due to A. perfoliata varied. Cobbold (1873) considered the clinical signs insignificant, whereas Neumann (1892) reported 2 cases in which perforation of the bowel was associated with large numbers of cestodes. Brumpt (1913), Mönnig (1934) and Neveu-Lemaire (1936) noted that the cestode could cause hyperaemia, ulceration and transient oedema, and considered the parasite a significant pathogen. Also, Skrjabin & Schultz (1937) considered A. perfoliata as a significant cause of colic in horses, and Spasskii (1951) summarized the work of Russian authors on the pathogenesis of infection and the disease syndromes attributed to the parasite.

Lesions produced by A. perfoliata have since been well documented in parasitology and pathology textbooks (e.g., Dunn, 1978; Bello et al. 1982; Soulsby, 1982; Arundel, 1985; Urquhart et al. 1987; Georgi & Georgi, 1990; Barker et al. 1993) and case study reports (e.g., Rodgers, 1966; Rees, 1967; Beroza et al. 1983; Edwards, 1986). Studies involving the examination of significant numbers of infected horses include those of Bain & Kelly (1977), Beroza et al. (1985), Owen et al. (1988), Pearson et al. (1993), Fogarty et al. (1994) and Nilsson et al. (1995). Only some more recent studies (e.g., Fogarty et al. 1994; Williamson et al. 1997) have specifically described in detail lesions on the caecal wall, the earlier studies focusing on lesions in the ileocaecal junction.

Macroscopically, the most widely reported pathological lesion due to A. perfoliata has been ulceration at the ileocaecal junction (Rees, 1967; Bain & Kelly, 1977; Dunn, 1978; Bello et al. 1982; Soulsby, 1982; Arundel, 1985; Beroza et al. 1985; Cosgrove et al. 1986; Edwards, 1986; Klei, 1986; Owen et al. 1988; Georgi & Georgi, 1990; Barker et al. 1993; Pearson et al. 1993; Fogarty et al. 1994), often with associated diphtheritic membrane formation (Rees, 1967; Bain & Kelly, 1977; Arundel, 1985; Beroza et al. 1985; Barker et al. 1993; Fogarty et al. 1994). Other commonly recorded gross lesions include oedema (Rees, 1967; Soulsby, 1982; Arundel, 1985; Cosgrove et al. 1986; Fogarty et al. 1994), hyperaemia (Rees, 1967; Owen et al. 1988; Pearson et al. 1993), mucosal thickening (Dunn, 1978; Cosgrove et al. 1986; Edwards, 1986; Pearson et al. 1993; Fogarty et al. 1994) and polyps or raised nodular masses protruding from the ileocaecal junction (Bain & Kelly, 1977; Barclay et al. 1982; Soulsby, 1982; Barker et al. 1993; Pearson et al. 1993; Fogarty et al. 1994).

Microscopically, ulceration of the ileocaecal junction is the most widely reported finding (Bain & Kelly, 1977; Beroza et al. 1986; Georgi & Georgi, 1990; Pearson et al. 1993; Fogarty et al. 1994). Multiple studies report the presence of a diphtheritic membrane overlying the ulcerated site (Rees, 1967; Bain & Kelly, 1977; Beroza et al. 1985; Pearson et al. 1993; Fogarty et al. 1994; Williamson et al. 1997) comprised mainly of fibrin and an inflammatory cellular exudate dominated by eosinophils. Luminal plant matter and bacteria often adhere to the membrane (Pearson et al. 1993; Fogarty et al. 1994). Infiltration of various types of inflammatory cells is amongst the most commonly described change at the site of ulceration. In particular, eosinophilic or lymphocytic infiltrations have been reported (Rees, 1967; Bain & Kelly, 1977; Beroza et al. 1983, 1985; Pearson et al. 1993; Fogarty et al. 1994), and in some cases infiltration of the two cell types was reported to occur concurrently (Bain & Kelly, 1977; Beroza et al. 1983, 1985; Pearson et al. 1993). Other changes including oedema (Rees, 1967; Beroza et al. 1985; Fogarty et al. 1994), goblet cell hyperplasia and hypertrophy (Rees, 1967; Beroza et al. 1985), fibrosis (Beroza et al. 1983; Fogarty et al. 1994) and abcessation have been reported in a small number of cases by Oxspring (1934) cited by Beroza et al. (1985), Pearson et al. (1993) and Fogarty et al. (1994).

A. perfoliata attaches to the intestinal mucosa by grasping the surface with its muscular suckers. When the tapeworms are removed from the gut wall a small ‘tent’ or plug of mucosa protrudes from the site of attachment (Dunn, 1978; Williamson et al. 1997). Thus, the mode of attachment of the worm results in significant damage to the gut wall. A correlation between the intensity of infection and the degree of mucosal damage at both the gross and histological levels has been documented (Pearson et al. 1993; Fogarty et al. 1994). In these studies, where fewer than 20 worms were present per horse, macroscopic lesions ranged from congestion and cellular infiltration to focal ulceration and diphtheresis. Histologically, the cellular infiltration and ulceration present was concentrated at the attachment site. Fogarty et al. (1994) found that the ulceration was generally confined to the mucosa in light infections. Gross lesions were more extensive in horses harbouring more than 21 tapeworms. In all cases, the mucosa appeared thickened and ulcerated, with diphtheresis. The studies included a small number of horses with polyps or raised nodular swellings on the ileocaecal junction. Histologically, there was ulceration, in some cases extending into the submucosa, and associated diphtheresis. Cellular infiltration was usually dominated by eosinophils. In addition, these authors observed mucosal haemorrhage, and Fogarty et al. (1994) described oedema as a characteristic feature. The latter study included a large number of horses in which fibroplasia damaged the mucosal architecture resulting in a distortion of the glandular pattern.

Fogarty et al. (1994) reported that the general features of the lesions on the caecal wall were similar to those on the ileocaecal junction, with macroscopic and microscopic ulceration and diphtheresis being the major findings. Histologically, fibroplasia was also described. However, the authors distinguished between the severity of the lesions produced by equivalent numbers of worms on the ileocaecal junction and in the caecum. Where small numbers of worms were present, 81% of lesions in the ileocaecal junction exhibited diphtheresis, compared with 30% of lesions on the caecal wall. If more than 21 worms were attached, there was no difference in the gross lesions at the two sites. At this intensity of infection, all attachment sites were described as grossly thickened, oedematous and ulcerated with diphtheresis. Histologically, however, 9 of the 21 horses with worms at the ileocaecal junction exhibited ulceration extending to the submucosa, whereas no horse with caecal worms had ulcers of that severity.

In a detailed study, Williamson et al. (1997) examined the attachment sites of A. perfoliata within the gastrointestinal tract and reviewed the major features of lesions produced at sites of attachment, concentrating particularly on a comparison of lesions at the ileocaecal junction with those on the caecal wall. The worms were attached in the ileocaecal junction (17%), caecal wall (81·1%), terminal ileum (1·7%) and in the ventral colon (0·2%). The severity of lesions produced at the sites of attachment was associated with the number of worms attached. Due to the small surface area in the ileocaecal junction, worms at this site were attached in close proximity, resulting in more severe lesions. The predominant features of the lesions included ulceration, diphtheresis and thickening of the mucosa, submucosa and lamina propria, and there was an increase in the number of eosinophils and a decrease in the number of lymphocytes present at the sites of lesions.

Based on the distribution of the parasite within the intestine and the characteristics of the lesions produced at attachment sites, Williamson et al. (1997) proposed 2 main ways in which this tapeworm causes disease. Firstly, a large number of worms attached to the ileocaecal junction, the narrowest part of the intestinal tract, may cause an obstruction to the flow of ingesta resulting in impaction colic. Secondly, clustering of worms at an attachment site may result in increased thickness and inflammation of the mucosa. Such changes may predispose to the development of an intussusception (Reymond, 1977). The clustering of worms on the caecal wall in horses with high intensities of infection would place the host at greater risk of intussusception than would a population of lower density. Clustering of worms is most common at the ileocaecal junction (Williamson et al. 1997). Thus, it seems that worms at this junction, in spite of their lower numbers, may be of greater clinical significance than those on the caecal wall. The cellular response to this parasite appears to be limited to the influx of eosinophils, suggesting that the damage which results at attachment sites may be predominantly mechanical, caused by the large scolex of A. perfoliata and the depth to which it penetrates the gut wall rather than being immunologically mediated (Williamson et al. 1997).

In spite of detailed descriptions of lesions, the pathogenesis of A. perfoliata remains largely conjectural. The lack of understanding of the pathogenesis of this condition arises in part from the paucity of knowledge of the ultrastructural features of the worm. For example, the presence of scolex glands, which have been identified in numerous other species of cestodes (reviewed by Smyth & McManus, 1989), would suggest an enzymatic factor in the pathogenesis, at least providing a mechanism by which such a factor may reach the host intestine. Bain & Kelly (1977) suggested that the enzyme acetylcholinesterase might decrease parasympathetic transmission by reducing acetylcholinesterase to acetic acid and choline before it can reach the post-synaptic membrane and initiate an action potential. A decrease in the number of action potentials would decrease gut motility, predisposing the animal to intestinal abnormalities, such as intussusception. This hypothesis was based on the study of Lee & Tatchell (1964), who found that acetylcholinesterase was present in A. perfoliata in large amounts in many different parts of the worm, but remains to be tested. Barclay et al. (1982) suggested another possible mechanism, directly involving pathological changes in the gut wall at the site of tapeworm attachment. Such changes, which may include ulceration and granulation tissue formation, were proposed to alter the contractility of the gut wall, predisposing the animal to intestinal intussusception.

CLINICAL ASPECTS

There is now ample evidence that chronic infection with A. perfoliata can cause clinical disease (Proudman & Trees, 1999; Matthews et al. 2004). This evidence is contrary to earlier beliefs, such as those of Rooney (1970) who stated that no apparent pathogenic significance had been observed to result from A. perfoliata infection and of Dunn (1978) who stated that this species had been wrongly incriminated as a pathogen. Evidence from reports of clinical cases has shown an association between the presence of tapeworms and various abdominal conditions (Rodgers, 1966; Slocombe, 1979; Foerner et al. 1980; Barclay et al. 1982; Beroza et al. 1983; Beroza et al. 1986; Cosgrove et al. 1986; Edwards, 1986; Carmel, 1988; Owen et al. 1989; Ryu et al. 2001; Little & Blikslager, 2002). Common clinical conditions reported are ileocaecal intussusception (Rodgers, 1966; Barclay et al. 1982; Beroza et al. 1985; Cosgrove et al. 1986; Edwards, 1986; Owen et al. 1989), caeco-caecal intussusception (Foerner et al. 1980; Barclay et al. 1982; Beroza et al. 1986; Edwards, 1986; Owen et al. 1989), caecal perforation leading to peritonitis (Barclay et al. 1982; Beroza et al. 1983, 1985), intestinal obstruction caused either by mural masses produced in the wall of the caecum or terminal ileum (Slocombe, 1979; Beroza et al. 1983; Carmel, 1988), or by mechanical obstruction of the ileocaecal junction by larger numbers of tapeworms (Bello, 1979; Carmel, 1988).

Proudman & Edwards (1993) demonstrated an association between the presence of tapeworms and the occurrence of ileocaecal colic. The type of ileocaecal colic most commonly associated with the presence of tapeworms was ileal impaction, with 7 of the 9 cases studied harbouring A. perfoliata. The intensity of infection was not determined in this study, as infection was diagnosed using a faecal flotation method. Intensity of infection is an important parameter because it has been linked to clinical significance by several authors (Slocombe, 1979; Bello et al. 1982; Soulsby, 1982; Urquhart et al. 1987; Carmel, 1988; French & Chapman, 1992). All authors agreed that a small number of tapeworms does not produce clinical disease but that larger numbers can.

Edwards (1986) suggested that there has been an increase in the mean intensity of infection of A. perfoliata, resulting in increased numbers of clinical cases attributable to this tapeworm. Edwards (1986) postulated that the cause of such an increase may be the current, widespread use of macrocyclic lactones, a highly effective nematicide which is ineffective against cestodes. It was also thought that the reduction in nematode infections may decrease competition against the tapeworms, resulting in an increased intensity of infection. This suggestion was supported by Owen et al. (1988), who used a regime of alternating pyrantel embonate and fenbendazole treatments (before the advent of macrocyclic lactones). Once macrocyclic lactones replaced these anthelmintics, tapeworm segments were seen in faeces by some owners, a situation not encountered previously. The association between ivermectin usage and intensity of A. perfoliata infection was investigated by French et al. (1994) who recorded the prevalence of tapeworms in 3 groups treated with different anthelmintic regimes. One group remained untreated, another received ivermectin and a third was treated with a rotational drenching regime involving pyrantel pamoate, ivermectin, oxibendazole and thiabendazole-piperazine. In an analysis of faecal samples using a flotation technique, a significant difference between the 3 groups was demonstrated in the number of samples containing A. perfoliata eggs. The highest number of ‘positive’ samples was obtained from the ivermectin-treated group and the lowest number from the group for which anthelmintics were rotated. The authors claimed that the absence of strongyle eggs in the ivermectin-treated horses made A. perfoliata eggs more visible, explaining the greater number of positive samples. However, the rotational programme included ivermectin and benzimidazoles, both of which should have been effective against nematode parasites in the horses treated using that regime. When 7 mares from the untreated and ivermectin-treated groups were removed from their groups and treated with pyrantel pamoate, cestodes were recovered from the faeces of all except 1 of the animals. Foals treated with the same two regimes were necropsied and found to have a similar prevalence of A. perfoliata infection. On the basis of the results of treatment with pyrantel pamoate and the necropsy data, these authors concluded that ivermectin did not cause an increase in the intensity of A. perfoliata infection.

COPROLOGICAL DIAGNOSIS

The lack of an accurate test for the diagnosis of A. perfoliata infection in individual horses has been a significant impediment to investigating various aspects of equine cestodiasis. Faecal examination for eggs has been a common technique, although some authors (Dunn, 1978; Arundel, 1985) suggested the examination of faeces for proglottides or entire worms as an alternative approach. When present in faeces, the eggs can be readily identified, being 65–80 μm in diameter, often with one side that appears flattened. The egg contains a hexacanth embryo surrounded by a refractile pyriform apparatus (Spasskii, 1951; Thienpont et al. 1986).

The sensitivities of copro-diagnostic tests for A. perfoliata infection vary considerably. A suggested cause of the low sensitivity of faecal flotation techniques is the shedding of tapeworm eggs in gravid proglottides (Slocombe, 1979). Other possible explanations may be that proglottides are shed irregularly or that immature worms are present. However, for taeniid cestodes, it has been demonstrated that the majority of eggs are released from their proglottides before the faeces pass out of the host (Coman & Rickard, 1975), and it is thus possible that a proportion of A. perfoliata eggs are also released from proglottides while still within the intestine. The highest sensitivity for a diagnostic test recorded to date, 61% (Proudman & Edwards, 1992), is lower than desirable for a diagnostic test. Other studies have recorded sensitivities of 40–54% (Slocombe, 1979; Beroza et al. 1985, 1986; French et al. 1994; Meana et al. 1998; Martins et al. 2003); some of these investigations examined relatively small numbers of horses, limiting the assessment of the reliability of the techniques used.

Various approaches have been used for the coprological diagnosis of A. perfoliata infection. For instance, Carmel (1988) reported centrifugation/flotation techniques that were more sensitive than standard flotation methods. Beroza et al. (1986) compared a centrifugation/flotation technique with gravitational flotation, using the same saturated sucrose solution (specific gravity not specified) for both, by mixing known numbers of eggs obtained from formalin-preserved crushed worms in 30 g faecal samples. The percentage of eggs recovered from the centrifugation technique was 10 times higher than that of the flotation technique. While the centrifugation technique was clearly superior in terms of recovery of eggs from spiked samples, the two methods were equally sensitive (50%) in evaluating samples from 6 infected horses. Martins et al. (2003) recently tested 82 infected horses and claimed to achieve a diagnostic sensitivity of 34% using the original method described by Beroza et al. (1986) and 40% employing a modified method.

A number of flotation solutions has been assessed, the most effective being saturated sucrose. The sensitivity of the flotation technique does not appear to depend on the specific gravity of the flotation solution. For example, Beroza et al. (1985) found that of 5 solutions tested, a saturated sucrose solution gave the best results, despite having the lowest specific gravity. French et al. (1994) also achieved their best results with 2 sucrose solutions, both being equally successful in analysis of 6 faecal samples and more sensitive then 4 other methods applied to the same samples.

To date, Proudman & Edwards (1992) have achieved the highest diagnostic sensitivity (61%) using a combined centrifugation/flotation technique. The sensitivity increased to 92% when horses harbouring 20 or fewer worms were eliminated from the sample. Hence, the test is a diagnostic tool for horses at the highest risk of clinical disease. In spite of the increased test sensitivity in horses with higher intensities of infection, there was no correlation between intensity of infection and number of eggs found in this study (Proudman & Edwards, 1992). In a subsequent study, Williamson et al. (1998) compared the sensitivity of 3 coprological (1 sedimentation and 2 sucrose flotation) methods for the diagnosis of A perfoliata infection and related the results with worm numbers and mucosal damage in each infected horse, thus allowing an assessment to be made of their diagnostic value in horses likely to be at risk of clinical disease. Sensitivities of 23–38% were achieved. A simple sucrose flotation method achieved the highest sensitivity (38%) at all intensities of infection compared with the other two methods. Almost half of the horses examined harboured fewer that 10 tapeworms, exerting a significant effect on the overall sensitivity of the test. This sensitivity (38%) was higher than those used by some other workers (Slocombe, 1979; Lyons et al. 1983, 1984) and lower than others. For instance, the centrifugation technique of Beroza et al. (1986) achieved a sensitivity of 50%. However, that technique was validated on a sample of only 6 infected horses. Beroza et al. (1985) also achieved a higher sensitivity using a flotation technique with saturated sucrose, detecting tapeworm eggs in 8 of 18 faecal samples (44%). Other studies, such as those of Slocombe (1979) (40% of 10 samples) and French et al. (1994) (40% of 6 samples), obtained similar sensitivities. However, the sample sizes in the latter studies were insufficient to establish the value of the techniques. Another factor that may contribute to the low sensitivity is horses harbouring a pre-patent infection (i.e. juvenile worms). The lack of information regarding the intensities of infection in some studies has made interpretation and comparison of results difficult. A large proportion of horses harbouring many tapeworms is expected to maximize the sensitivity of any test performed. A number of studies (e.g., Proudman & Edwards, 1992; Williamson et al. 1997; Höglund et al. 1998) have shown a clear association between sensitivity and intensity of infection and that these two parameters increase concurrently.

Several authors have associated intensity of infection with probability of clinical disease (Slocombe, 1979; Bello et al. 1982; Carmel, 1988; French et al. 1992). The studies of Pearson et al. (1993) and Fogarty et al. (1994) have also shown that the intensity of infection has a significant effect on the severity of lesions produced on the gut wall. Thus, the correlation between intensity of infection and sensitivity of a diagnostic test is important. A test identifying horses most at risk of clinical disease is considerably more valuable to the clinician than the overall sensitivity would indicate. Proudman & Edwards (1992) also demonstrated an increase in sensitivity at higher intensities of infection (an increase from 61% overall to 92% when horses with fewer than 20 tapeworms were eliminated from the analysis). The authors were, however, unable to relate the finding to pathogenicity, as the damage caused by the tapeworms was not assessed in that study.

The correlation found in the study by Williamson et al. (1998) between intensity of infection and A. perfoliata eggs recovered using a simple flotation method is of value because it gives an indication of the intensity of infection and thus of the potential for clinical disease. The only other study in which such a relationship has been investigated is that of Proudman & Edwards (1992), in which a correlation between these factors was not shown. Williamson et al. (1998) demonstrated that a correlation existed between intensity of infection, pathogenicity and the sensitivity of a simple sucrose flotation method. Although the overall sensitivity of the method was relatively low, its reliability at intensities of infection that may predispose to clinical disease was high. The correlation identified between intensity of infection and eggs per gram detected gives an indication of the significance of the number of eggs recovered in a test sample. While a high egg count is a reliable indication of a horse at possible risk of intestinal lesions, a low egg count does not preclude the possibility that the horse is harbouring a large number of tapeworms.

SEROLOGICAL DIAGNOSIS

In order to attempt to overcome the limitations of coprological diagnosis, various workers have evaluated serodiagnostic approaches using somatic or excretory/secretory (ES) products from A. perfoliata as antigens (Höglund et al. 1995, 1998; Proudman & Trees, 1996a,b). For example, Höglund et al. (1995) assessed an enzyme-linked immunosorbent assay (ELISA) using scolex antigen for the detection of serum antibody and used it to test sera from slaughtered horses (n=426) with variable infection intensities of the parasite and with varying degrees of intestinal lesions. The findings showed a strong relationship between specific IgG antibody levels and both the number of tapeworms present and the extent of pathological change. No serological cross-reactivity with antigens from parasitic nematodes was apparent. However, there was a considerable degree of variation in antibody levels among horses with similar infection intensities. Interestingly, a seasonal pattern in antibody levels was measured, interpreted to reflect the establishment of newly acquired infection. It was concluded that the ELISA had the potential for monitoring A. perfoliata infection in groups of horses and that it could be used as a complementary diagnostic tool for studying the epidemiology of the parasite (cf. Höglund et al. 1998). Subsequently, Proudman & Trees (1996a) evaluated the use of whole (somatic) extract or excretory/secretory (ES) products from A. perfoliata for the specific detection of serum antibody in the ELISA. Sera from horses known to be infected with A. perfoliata (n=38), control sera from horses with infections of Anoplocephaloides mamillana (n=3), of multiple species of strongylid (n=6) and Parascaris equorum (n=1) and ‘negative control sera’ from ponies raised and maintained helminth-free (n=20) were used. In an optimized ELISA, the ES products gave the best differentiation between sera from A. perfoliata-infected and those not infected with this parasite. The ELISA achieved a diagnostic sensitivity of 68% and a specificity of 95% for sera from the 20 helminth-free horses. Cross-reactivity in ELISA was detected in 6 of the sera from horses infected with parasites other than A. perfoliata. Western blot analysis of ES products indicated that components of ~12/13 kDa were recognized by antibodies in (pooled and individual) sera from tapeworm-positive but not from selected tapeworm-naïve horses.

Proudman & Trees (1996b) affinity-purified the 12/13 kDa component from ES products, and assessed it in the ELISA for the specific detection of anti-12/13 kDa IgG and IgG(T) serum antibody in horses with known intensities of A. perfoliata infection (n=94) and horses (n=33) for which there was no evidence of A. perfoliata infection (based on coprological examination; Proudman & Edwards, 1992) but which harboured other parasites, including strongylids, P. equorum or A. mamillana. Analysis of the results showed that there was a positive correlation (0·56 and 0·63, respectively) between IgG and IgG(T) serum antibody and the intensity of infection with A. perfoliata. Linear regression analysis indicated that the IgG(T) antibody response was the best predictor of the intensity of infection. The correlation between anti-12/13 kDa IgG(T) antibody level and intensity of infection was interpreted to be relevant clinically, because the risk of A. perfoliata-associated colic is considered to be proportional to the intensity of infection, which is also consistent with the finding that the degree of pathological change in the ileo-caecal junction in infected horses is linked to the intensity of infection (e.g., Pearson et al. 1993; Forgarty et al. 1994; Williamson et al. 1997; Rodriguez-Bertos et al. 1999). Also, anti-12/13 kDa IgG(T) antibody levels were shown to decrease in horses (n=4) as early as 28 days after the elimination of A. perfoliata by cesticidal treatment. Immunoblot analysis of the 12/13 kDa component and other helminth extracts using pooled sera from A. perfoliata-infected horses (n=10) and A. perfoliata-naïve horses (n=10) suggested that IgG(T) antibody responses to the 12/13 kDa antigen component were specific. Proudman & Trees (1996b) concluded that the ELISA for the detection of IgG(T) antibody provided a useful diagnostic tool. Extending this work, studies were conducted to investigate the epidemiology of A. perfoliata infection (Proudman et al. 1997), the relationship between A. perfoliata infection and clinical disease (Proudman et al. 1998), and populations of horses with a history consistent with A. perfoliata-associated colic (Proudman & Holdstock, 2000).

Proudman et al. (1997) utilized the correlation between IgG(T) response and intensity of infection to examine the variation in natural infection intensity with age. Infection intensity, based on the serological responses, showed a triphasic age-dependency profile, with a peak intensity in young horses (0·5–2 years), reaching a plateau in horses of 3–15 years of age, and rising again in older horses. The determinants of this age-intensity pattern were interpreted to be immunological or behavioural. In a case-control study, Proudman et al. (1998) investigated the relationship between IgG(T) antibody levels against the 12/13 kDa antigenic component and the risk of spasmodic colic. Sera from horses with spasmodic colic were tested serologically and compared with those from unaffected (control) horses. Conditional logistic regression modelling indicated a dose response relationship between the intensity of infection with A. perfoliata and the risk of spasmodic colic. These findings were supported by a subsequent study of an outbreak of tapeworm-associated colic in a training/rehabilitation yard where the anthlemintic control program consisted of oral treatment with ivermectin alone every 1–2 months (Proudman & Holdstock, 2000). In this investigation, horses for which there was serological and coprological evidence of a high intensity of infection were observed to be at an increased risk of colic. Cesticidal treatment of some horses resulted in a marked decrease in anti-12/13 kDa IgG(T) antibody levels, which was linked to a considerable decrease in the incidence of colic in the horse population. The study indicated that anthelminthic regimens employing a macrocyclic lactone alone might lead to an increased intensity of A. perfoliata infection and associated colic. The results provided support for the hypothesis that the risk of ileal impaction colic and spasmodic colic increases with increased intensity of A. perfoliata infection. The authors concluded that the findings of the study supported the practical application of the serological assay, which has been reinforced by a recent investigation in the USA (Little et al. 2002). The anti-12/13 kDa IgG (T) ELISA (Proudman & Trees, 1996b) has been developed commercially (www.diagnosteq.com.uk) and is presently offered as a diagnostic service through the University of Liverpool, Leahurst, UK. This ELISA is considered to provide a complementary tool to support clinical diagnosis, for research purposes and for the prevention of colic associated with A. perfoliata.

TREATMENT AND CONTROL

Treatment is a central component in the control of equine cestodiasis. Effective drugs are dichlorophen given at a 180 mg/kg and niclosamide (100 mg/kg) (Arundel, 1985). Morantel tartrate is probably effective at high dose rates (e.g., 12·5 mg/kg bodyweight) (Arundel, 1985), but Drudge et al. (1984) showed poor removal (25%) of A. perfoliata using morantel-trichlorfon paste at a dose rate of 6 mg/kg. Key anthelmintics which are ineffective against A. perfoliata are ivermectin, moxidectin and oxfendazole (Kingsbury & Reid, 1981; Monahan et al. 1996). In North America, the drug most commonly used for the treatment of equine cestodiasis is pyrantel pamoate. Numerous studies have been conducted on the efficacy of this drug (e.g., Slocombe, 1979, 2004; Lyons et al. 1986, 1989). Slocombe (1979) confirmed that pyrantel pamoate is more effective at 13·2 mg pyrantel base/kg, which is twice the recommended dose for nematodes. Slocombe's results are supported by those of Lyons et al. (1986) who showed that a dose rate of 13·2 mg pyrantel base/kg had a significantly higher efficacy than 6·6 mg pyrantel base/kg, the recommended dose rate for nematodes. More recently, the effectiveness of praziquantel for the treatment of A. perfoliata has been demonstrated (Lyons et al. 1992, 1995, 1998); efficacies of >80% were achieved with dose rates of 0·75 mg/kg and 1·0 mg/kg. Drug combinations, such as pyrantel pamoate/trichlorfon and praziquantel/moxidectin (Grubbs et al. 2003) or praziquantel/ivermectin (Coles et al. 2003; Rehbein et al. 2003; Mercier et al. 2003), can also be used for (endo/ectocidal) treatment against a broader spectrum of parasites. In addition to treatment strategies, control of A. perfoliata may also be complemented with pasture management. For instance, Bello et al. (1982) suggested cropping or rotational grazing by cattle. The first part of this recommendation is based on the fact that oribatid mites are rare on ploughed ground, also acknowledged by Bain & Kelly (1977). Rotational grazing would ensure that infected mites were not ingested by horses.

CONCLUSION

Currently, A. perfoliata infection of horses is considered to be of clinical significance. While there has been significant research on some aspects of the parasite's biology, epidemiology and the clinical disease caused by it, there is still a paucity of information regarding the pathogenesis of the disease, and the ecology and population genetics of the parasite in different geographical regions, all of which have implications for a better understanding of the parasite and its transmission pattern(s). While a significant limitation for epidemiological and ecological studies has been the lack of an accurate method for the diagnosis of infection in the live horse, some studies (e.g., Proudman & Edwards, 1992; Proudman & Trees, 1996a,b; Proudman et al. 1997, 1998; Williamson et al. 1997; Proudman & Holdstock, 2000) have shown that coprological and/or serological approaches can be useful for the detection of horses at risk of clinical disease (i.e., those with high intensities of infection). While a serological technique has been established for the detection of horses at risk of clinical tapeworm disease, and for estimating the prevalence of anti-A. perfoliata (12/13 kDa) IgG(T) antibody in horse populations (Proudman et al. 1996b), such an assay does not allow the reliable differentiation of current (pre-patent or patent) and recent past infections in individual horses. The demonstration that chronic A. perfoliata infection of high intensity places horses at considerable risk of clinical disease, combined with the inability to reliably diagnose current infection in individual horses, means that regular preventative treatment with an effective cesticide is central to the health and welfare of horses. Nonetheless, the development of an accurate test for the diagnosis of current (patent) A. perfoliata infection in individual horses would be a significant advance, and would clearly assist in prevention and control. There has been some discussion of the potential of an immunological assay for the detection of A. perfoliata antigen(s) in the faeces from infected horses, as effective assays have been developed for taeniid cestodes in dogs and humans (e.g., Allan et al. 1992, 1993; Deplazes et al. 1992, 1999), but unpublished findings suggest that the sensitivity of an assay for the detection of A. perfoliata ‘coproantigens’ is low (Proudman, personal communication, 2003).

Given the improvements in the performance of some molecular methods (i.e. the ability to specifically detect or amplify tiny amounts of DNA from biological samples), it is possible that polymerase chain reaction (PCR) or other amplification techniques could be adopted and optimized to specifically detect A. perfoliata DNA (sloughed from the active tegumental surface and from eggs) in the faeces from infected horses. Since the spectrum of other species of tapeworm is limited in horses in most countries, the definition of specific genetic markers for A. perfoliata seems feasible (cf. Drogemuller et al. 2004). While the intensity of A. perfoliata infection in horses can be low, the development of a semi-nested or nested PCR approach (cf. Verweij et al. 2000, 2001; Yamasaki et al. 2004) for the detection of current infection in horses at clinical risk seems a useful option. This statement is supported, in particular, by a recent report (Traversa et al. 2004) describing the exquisite diagnostic specificity (100%) and sensitivity (~97%) of a semi-nested PCR for the specific diagnosis of gastric habronemiasis in horses, estimated to be able to specifically detect a minimum of ~0·02 fg of parasite DNA. In spite of the very low reproductive potential of species of Habronema, H. microstoma and/or H. muscae, DNA could be amplified specifically from faecal samples from horses with as few as 3–7 adult specimens in the stomach (confirmed by necropsy) (Traversa et al. 2004). Such findings provide prospect for improved diagnosis of A. perfoliata infection by PCR, which would further enhance present prevention and control efforts. If successful, such a molecular tool could also be applied to evaluate the efficacy of anthelmintics against A. perfoliata by monitoring the decline or absence of parasite-specific DNA in faeces after treatment in live horses, thus circumventing the need to sacrifice them. A PCR tool for the specific detection of the parasite in the intermediate hosts would also be useful for fundamental investigations into the ecology of the parasite. Although very labour intensive, the prevalence, seasonal occurrence and intensity of A. perfoliata infection in intermediate host populations could be assessed, which would contribute to understanding transmission patterns. Also, mutation scanning methods, employing sufficiently variable genetic markers, could be used to investigate the genetic structure(s) of A. perfoliata populations (adult worms and cysticercoids) in order to establish whether different genetic variants exist which have distinct transmission patterns in particular intermediate host species.