CMR FigSearch
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dubey, J. P.
Right arrow Articles by Ortega-Mora, L. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Dubey, J. P.
Right arrow Articles by Ortega-Mora, L. M.

 Previous Article  |  Next Article 

Clinical Microbiology Reviews, April 2007, p. 323-367, Vol. 20, No. 2
0893-8512/07/$08.00+0     doi:10.1128/CMR.00031-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Epidemiology and Control of Neosporosis and Neospora caninum

J. P. Dubey,1* G. Schares,2 and L. M. Ortega-Mora3

Animal Parasitic Diseases Laboratory, Animal and Natural Resources Institute, Agricultural Research Service, United States Department of Agriculture, Beltsville, Maryland 20705,1 Institute of Epidemiology, Friedrich-Loeffler-Institut—Federal Research Institute for Animal Health, Seestrasse 55, D-16868 Wusterhausen, Germany,2 SALUVET, Animal Health Department, Faculty of Veterinary Sciences, Complutense University of Madrid, Ciudad Universitaria s/n, 28040 Madrid, Spain3

SUMMARY
INTRODUCTION
LIFE CYCLE
HOST RANGE AND GEOGRAPHIC DISTRIBUTION
    Hosts Proven by Isolation of Viable N. caninum by Bioassays with Animals, Cell Culture, or Both
    Hosts with N. caninum-like Parasites Demonstrated by Immunohistochemical (IHC) Staining of Parasites by Specific Antibodies, by N. caninum DNA, or by Both but Not by Isolation of Viable Parasites
    Serologic Prevalence of N. caninum Antibodies in Animals and Humans
    Zoonotic Aspects of N. caninum
OOCYST SHEDDING BY DOGS AND OTHER DEFINITIVE HOSTS
    Oocyst Shedding by Naturally Infected Dogs
    Coyotes and Other Definitive Hosts of N. caninum
STRAIN VARIATION AND PATHOGENICITY
TRANSMISSION
    Transmission in All Hosts
    Transmission of N. caninum in Dogs
    Transmission of N. caninum in Cattle
        Transplacental (vertical) transmission.
        Post-natal (horizontal) transmission.
RISK FACTORS FOR BOVINE NEOSPOROSIS
    Epidemic and Endemic N. caninum-Associated Abortion
    Risk Factor Studies
    Infection Risk
        Age of cattle.
        Definitive hosts (dogs and coyotes).
        Other carnivores.
        Intermediate hosts other than cattle.
        Grazing, fodder, and drinking water.
        Feeding colostrum or milk.
        Calving management.
        Cattle stocking density and size of farmland.
        Herd size.
        Source of replacement heifers.
        Climate.
        Vegetation index.
        Human population density.
        Factors related to antibodies against other infectious agents.
        Breed.
        Type of housing.
    Abortion Risk
        Seropositivity of individual cattle.
        Seroprevalence in the herd.
        Factors related to infection risk.
            (i) Age.
            (ii) Farm dogs.
            (iii) Wild canids.
            (iv) Cats.
            (v) Other potential intermediate hosts such as poultry and horses.
            (vi) Fodder.
            (vii) Climate and season.
            (viii) Farm-raised replacement heifers.
            (ix) Proximity to a town or village.
            (x) Factors related to antibodies against other infectious agents.
            (xi) Housing.
        Factors associated with reproduction.
            (i) Previous abortions.
            (ii) Annual rate of cows returning to estrus postpregnancy.
            (iii) Retained afterbirths.
            (iv) Use of beef bull semen to inseminate dairy cattle.
            (v) Use of calving pens to hospitalize sick animals.
        Attendance at cattle shows.
PREVENTION AND CONTROL
    Economic Losses and Cost-Benefit Analyses
    Use of Diagnostic Tools in the Control of N. caninum
        Detection of the infection and infection-abortion relationship.
        Investigation of the route of transmission.
        Testing of replacements.
    Control Measures
        Farm biosecurity.
        (i) Quarantine and testing of replacement and purchased cattle.
        (ii) Prevention of transmission from dogs and other potential definitive hosts.
        (iii) Prevention of waterborne transmission.
        (iv) Rodent control.
        (v) Prevention of putative factors for disease recrudescence in congenitally infected cattle.
        Reproductive management.
        (i) Embryo transfer.
        (ii) Artificial insemination of seropositive dams with semen from beef bulls.
        Testing and culling.
        Chemotherapy.
        Vaccination.
        (i) Key points of vaccine design for bovine neosporosis.
        (ii) Live versus dead vaccines.
        (iii) Perspectives and recommendations.
ACKNOWLEDGMENTS
REFERENCES

   SUMMARY
Top
 Next
 References
 
Neospora caninum is a protozoan parasite of animals. Until 1988, it was misidentified as Toxoplasma gondii. Since its first recognition in dogs in 1984 and the description of the new genus and species Neospora caninum in 1988, neosporosis has emerged as a serious disease of cattle and dogs worldwide. Abortions and neonatal mortality are a major problem in livestock operations, and neosporosis is a major cause of abortion in cattle. Although antibodies to N. caninum have been reported, the parasite has not been detected in human tissues. Thus, the zoonotic potential is uncertain. This review is focused mainly on the epidemiology and control of neosporosis in cattle, but worldwide seroprevalences of N. caninum in animals and humans are tabulated. The role of wildlife in the life cycle of N. caninum and strategies for the control of neosporosis in cattle are discussed.


   INTRODUCTION
Top
 Previous
 Next
 References
 
Neospora caninum is a protozoan parasite of animals. Until 1988, it was misdiagnosed as Toxoplasma gondii (138). Since its first recognition in 1984 in dogs in Norway (52) and the description of the new genus and species Neospora caninum by Dubey et al. (138), neosporosis has emerged as a serious disease of cattle and dogs worldwide. Abortions and neonatal mortality are a major problem in livestock operations, and neosporosis is a major cause of abortion in cattle. We have previously reviewed the general biology of N. caninum (130) and the pathogenesis and diagnosis of neosporosis in cattle (128, 133, 135, 158, 328). Although antibodies to N. caninum have been reported (275, 440), the parasite has not been demonstrated in human tissues. Thus, the zoonotic potential is uncertain. This review is focused on the epidemiology and control of neosporosis in cattle.


   LIFE CYCLE
Top
 Previous
 Next
 References
 
N. caninum is a coccidian parasite with a wide host range. In general, it is very similar in structure and life cycle to T. gondii, with two important differences: (i) neosporosis is primarily a disease of cattle, and dogs and related canids are definitive hosts of N. caninum, whereas (ii) toxoplasmosis is primarily a disease of humans, sheep, and goats, and felids are the only definitive hosts of T. gondii.

The life cycle is typified by the three known infectious stages: tachyzoites, tissue cysts, and oocysts (Fig. 1 and 2). Tachyzoites and tissue cysts are the stages found in intermediate hosts, and they occur intracellularly (152). Tachyzoites are approximately 6 by 2 µm (Fig. 2). Tissue cysts are often round or oval in shape, up to 107 µm long, and are found primarily in the central nervous system. The tissue cyst wall is up to 4 µm thick, and the enclosed bradyzoites are 7 to 8 by 2 µm. Extraneural tissues, especially muscles, may contain tissue cysts (155, 348).


Figure 1
View larger version (49K):
[in this window]
[in a new window]

  		FIG. 1. Life cycle of Neospora caninum. (Reprinted from reference 128.)

 

Figure 2
View larger version (114K):
[in this window]
[in a new window]

  		FIG. 2. Life cycle stages of Neospora caninum. (A) Impression smear of the liver of an experimentally infected mouse depicting numerous tachyzoites (Giemsa stain). Notice that the tachyzoites vary in dimension, depending on the stage of division: (a) a slender tachyzoite, (b) a tachyzoite before division, and (c) three dividing tachyzoites compared with the size of a red blood cell (arrow). (B) Histological section of a tissue cyst inside a neuron in the spinal cord of a congenitally infected calf (hematoxylin and eosin stain). Note the thick cyst wall (opposing arrowheads) enclosing slender bradyzoites (open triangle). The host cell nucleus (arrow) is cut at an angle. (C) Unsporulated oocyst (arrow) with a central undivided mass in the feces of a dog (unstained). Bar, 10 µm. (D) Sporulated oocyst (arrow) with two internal sporocysts (unstained). Bar, 10 µm.

 
The environmentally resistant stage of the parasite, the oocyst, is excreted in the feces of dogs and coyotes in an unsporulated stage (188, 270, 294). Oocysts sporulate outside the host in as few as 24 h (270). Nothing is known about the survival of N. caninum oocysts in the environment. Because of its close relationship with T. gondii, it is assumed that the environmental resistance of N. caninum oocysts is similar to that of T. gondii oocysts (131).

All three infectious stages of N. caninum (tachyzoites, bradyzoites, and oocysts) are involved in the transmission of the parasite. Carnivores probably become infected by ingesting tissues containing bradyzoites, and herbivores probably become infected by the ingestion of food or drinking water contaminated by N. caninum sporulated oocysts. Transplacental infection can occur when tachyzoites are transmitted from an infected dam to her fetus during pregnancy.


   HOST RANGE AND GEOGRAPHIC DISTRIBUTION
Top
 Previous
 Next
 References
 
In order to understand the epidemiology of N. caninum, it is important to identify its host range and geographic distribution. Unlike T. gondii, viable N. caninum is difficult to isolate. Additionally, another species, Neospora hughesi, has been described as being isolated from horses (292). Therefore, we have made an attempt to identify different hosts of N. caninum.

Hosts Proven by Isolation of Viable N. caninum by Bioassays with Animals, Cell Culture, or Both

Viable N. caninum has been isolated from cattle, sheep, dogs, white-tailed deer, and water buffaloes (Table 1). Most of these isolates were from clinically affected animals and from neonatally infected animals, except for the isolates from buffaloes, sheep, and deer, which were from adult asymptomatic animals. Isolation of viable N. caninum has been achieved with a variety of cell cultures and by bioassays of immunosuppressed mice, gerbils, and dogs (135). Isolation in cell culture is limited by the necessity of having materials not contaminated with other microbes, and not all isolates can be adapted to grow in cell culture (457). Bioassays of immunosuppressed mice are expensive because outbred mice are not useful for propagating N. caninum. Isolation of N. caninum by feeding infected tissues to dogs and then examining canine feces for oocysts has the advantage that larger volumes of material can be fed to dogs than can ever be tested with cell culture or rodents. However, the identification of N. caninum in the feces of dogs should be based on the recovery of viable tachyzoites in cell culture or rodents inoculated with oocysts because of the existence of other N. caninum-like parasites in canine feces (403).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Intermediate and definitive host ranges and distributions of N. caninum or N. hughesi proven by isolation of the parasite

 
Hosts with N. caninum-like Parasites Demonstrated by Immunohistochemical (IHC) Staining of Parasites by Specific Antibodies, by N. caninum DNA, or by Both but Not by Isolation of Viable Parasites

N. caninum was demonstrated histologically in a few clinically affected deer, a raccoon, a rhinoceros, and goats, and DNA was found in a few animals (Table 2). We stress that finding DNA is not synonymous with finding viable N. caninum. Attempts to isolate viable N. caninum from rodent tissues that had demonstrable DNA were unsuccessful (235).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Host range and distribution of N. caninum demonstrated by IHC or DNA but not by isolation in noncanine, nonbovine domestic animals

 
Serologic Prevalence of N. caninum Antibodies in Animals and Humans

Worldwide seroprevalences of N. caninum in dogs (Table 3), dairy cattle (Table 4), beef cattle (Table 5), other domestic animals (Table 6), wildlife and zoo animals (Table 7), and humans (Table 8) are summarized. Although these results are not comparable because of different serologic methods and different cutoff values used, they do provide evidence that many species of mammals have been exposed to this parasite. Many data summarized in Tables 3 to 8 are based on convenience samples obtained for other purposes. Also, the clinical status of the subjects surveyed was not stated, and in many of the reports, the prevalence of N. caninum was consistently higher in rural than in city dogs or pets (Table 3). In a well designed study, seroprevalences were compared in dairy and beef cattle from Germany, The Netherlands, Spain, and Sweden by use of randomized samples and enzyme-linked immunosorbent assays (ELISAs) that had been previously standardized among laboratories (39, 460). In this study, the seroprevalence in cattle in Sweden was much lower than in neighboring countries and prevalences in beef cattle were lower than in dairy cattle (Tables 4 and 5). As yet, there is no evidence that avian species are natural hosts for N. caninum (183).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Prevalence of N. caninum antibodies in dogs

 

View this table:
[in this window]
[in a new window]

  		TABLE 4. Serologic prevalence of N. caninum antibodies in dairy cattle

 

View this table:
[in this window]
[in a new window]

  		TABLE 5. Serologic prevalence of N. caninum antibodies in beef cattle

 

View this table:
[in this window]
[in a new window]

  		TABLE 6. Prevalence of antibodies to N. caninum in noncanine, nonbovine domestic animals

 

View this table:
[in this window]
[in a new window]

  		TABLE 7. Seroprevalence of Neospora caninum antibodies in wildlife

 

View this table:
[in this window]
[in a new window]

  		TABLE 8. Seroprevalence of N. caninum in humans

 
None of the serologic tests used to detect N. caninum antibodies have been validated based on recovery of the viable parasite in any host. Therefore, the cutoff values used for serologic diagnosis of N. caninum are presumptive. Because N. caninum is structurally and molecularly related to T. gondii, these parasites are antigenically different and serologic cross-reactivity, if present, is considered minor. It is noteworthy that about 80% of black bears in the United States were found to be infected with T. gondii, but none had antibodies to N. caninum (136, 156).

Zoonotic Aspects of N. caninum

Because two rhesus monkeys (Macaca mulatta) have been successfully infected with N. caninum (35), there is concern about the zoonotic potential of N. caninum. However, at present there is no firm evidence that N. caninum successfully infects humans, because only low levels of antibodies have been reported (Table 8), and neither N. caninum DNA nor the parasite has been demonstrated in human tissues. As yet, no accidental N. caninum infections in persons handling viable organisms have been reported, and thus there are no reference sera with which to compare the results reported in Table 8.


   OOCYST SHEDDING BY DOGS AND OTHER DEFINITIVE HOSTS
Top
 Previous
 Next
 References
 
Oocysts are the key in the epidemiology of neosporosis, but little is known of the biology of N. caninum oocysts. Dogs shed oocysts 5 days or more after ingesting tissues of experimentally or naturally infected animals (Table 9). The total duration of oocyst shedding after primary infection varied from 1 to several days. The total number of oocysts shed, prepatent periods, and duration of oocyst shedding varied tremendously (Table 9). Factors affecting oocyst shedding are largely unknown and difficult to investigate because of the costs involved in housing dogs in a secure facility and the low numbers of oocysts shed and because oocyst shedding is erratic (Table 9). Apparently dogs shed more oocysts after ingesting bovine tissues than when fed murine tissues (187), and pups shed more oocysts than adult dogs (Table 9). Some of the dogs that had been given corticosteroids shed more than 100,000 oocysts after being fed with murine brains, suggesting that immunosuppressed dogs may shed more oocysts than immunocompetent dogs (270, 273). Schares et al. (403) found the highest number of oocysts from a naturally infected dog. This dog was splenectomized. Nothing is known about the effect of different breeds of dogs on oocyst shedding. In most experiments, hounds were used to collect oocysts (Table 9).


View this table:
[in this window]
[in a new window]

  		TABLE 9. Details of N. caninum oocyst shedding by dogs

 
Oocyst Shedding by Naturally Infected Dogs

N. caninum-like oocysts have been identified in only a few dogs worldwide. Because N. caninum oocysts structurally resemble another coccidian in dog feces, Hammondia heydorni (403, 416, 419), it is epidemiologically important to properly identify N. caninum oocysts. Available information on oocyst shedding by naturally infected dogs is reviewed. To our knowledge, there are only a few reports of N. caninum oocyst shedding by naturally infected dogs (44, 299, 300, 403, 416). Basso et al. (44) found a few N. caninum oocysts in the feces of a 45-day old Rottweiler from La Plata, Argentina. Viable N. caninum was recovered from the gerbils that were fed these oocysts, and the strain was successfully cultured in vitro.

Slapeta et al. (416) found 1 million oocysts in a 1-year-old German shepherd from the Czech Republic. The oocysts were considered N. caninum based on PCR, and bioassay was not reported.

McGarry et al. (299) examined a total of 15 fecal samples from two foxhound kennels in the United Kingdom (10 from one kennel of 80 and 5 from the second kennel of 60 dogs) and found N. caninum oocysts in two samples. One of these samples (from the pack of 60 foxhounds) was identified as N. caninum based on PCR; there were approximately 84 oocysts per gram of feces. A second fecal sample from this dog taken 4 months later revealed a few oocysts that were identified as N. caninum based on PCR.

McInnes et al. (300) detected N. caninum DNA in the feces of a dog in New Zealand 2.5 years after they had isolated viable N. caninum from the skin of the dog.

A comprehensive survey of N. caninum infection in the feces of dogs from Germany was reported by Schares et al. (403). N. caninum-like oocysts were found in 47 of 24,089 fecal samples. Twenty-eight of these fecal samples were bioassayed in gerbils. Based on seroconversion in bioassayed gerbils, seven samples were considered to be N. caninum. Five samples were definitively identified as N. caninum, based on successful in vitro cultivation. Among the other isolates, 12 were considered to be H. heydorni, 2 T. gondii, and 2 Hammondia hammondi. T. gondii and H. hammondi are pseudoparasites in dog feces and result from the ingestion of cat feces by dogs. This investigation highlights the difficulties of identification of N. caninum oocysts in canine feces. The number of N. caninum oocysts in naturally infected dog feces varied from a few to 114,000 per gram (in a 13-year-old dog that had been splenectomized). The infected dogs were 2 months to 13 years of age and were of seven different breeds (403).

Coyotes and Other Definitive Hosts of N. caninum

One of four captive-raised coyotes shed a few N. caninum oocysts after ingesting experimentally infected bovine tissues (188). N. caninum DNA was found in the feces of 2 of 85 coyotes and 2 of 271 foxes from Canada (471).


   STRAIN VARIATION AND PATHOGENICITY
Top
 Previous
 Next
 References
 
It is now well established that N. caninum can cause serious illness in cattle and dogs. Isolates of N. caninum from various hosts are genetically similar, although each strain has its own signature (365). Little is known of the strain variation with respect to pathogenicity. There are no suitable animal models for testing strain variation. In limited studies, some N. caninum strains were more pathogenic to mice than others (21, 264, 268, 300). Abortion or fetal infections have been induced in cattle by using a variety of isolates in different laboratories (158), but a meaningful comparison with pregnant cattle would be economically prohibitive. There is the additional complication of the stage of the parasite used and the source of the parasite. Most N. caninum strains are maintained in cell culture, and prolonged passage in culture can alter the pathogenicity and other characteristics of the parasite (42, 346). Additionally, data obtained from rodents may not be applicable to cattle.


   TRANSMISSION
Top
 Previous
 Next
 References
 

Transmission in All Hosts

N. caninum can be transmitted postnatally (horizontally, laterally) by ingestion of tissues infected with tachyzoites or tissue cysts or by ingestion of food or drinking water contaminated by sporulated oocysts, or it can be transmitted transplacentally (vertically, congenitally) from an infected dam to her fetus during pregnancy. Recently, the terms "exogenous transplacental transmission" and "endogenous transplacental transmission" have been proposed to describe more precisely the origin of the transplacental infection of the fetus (442). Exogenous transplacental transmission occurs after a primary, oocyst-derived, infection of a pregnant dam, while endogenous transplacental transmission occurs in a persistently infected dam after reactivation (recrudescence) of the infection during pregnancy. Mice were infected successfully by oral inoculation of tachyzoites or bradyzoites (264). These results are of interest because tachyzoites treated with acidic pepsin were rendered noninfective for cell cultures, whereas bradyzoites survived the acidic pepsin (264). Tissue cysts and bradyzoites can survive up to 2 weeks at refrigeration temperature (4°C) but are killed by freezing (155, 267). Oocysts were orally infective to cattle (111, 190, 443), goats and sheep (397), and rodents such as mice, gerbils (Meriones unguiculatus), and guinea pigs (Cavia porcellanus) (134, 294, 397). Transplacental transmission has been induced experimentally in cattle, dogs, sheep, goats, monkeys, cats, and mice and occurs naturally in many hosts (133). Transplacental transmission occurs when tachyzoites from the dam cross the placenta. The ingestion of oocysts is the only demonstrated mode for postnatal (horizontal) transmission in herbivores. Because of the epidemiological importance, we will discuss the modes of transmission of N. caninum in dogs and cattle separately.

Transmission of N. caninum in Dogs

How dogs become infected with N. caninum in nature is not fully understood. Historically, vertical transmission of neosporosis was first recognized in dogs (52, 140). Three successive litters from a bitch in Norway were found to have neosporosis (52). In a retrospective study, the most severe neosporosis was discovered in four German Shepherds from one owner in 1957 from Ohio (140), and there was evidence that a congenitally infected bitch transmitted the infection to her progeny (140). Transplacental transmission in experimentally infected dogs has been demonstrated (82, 132). In most cases of neonatal neosporosis, clinical signs are not apparent until 5 to 7 weeks after birth (133). These data suggest that N. caninum is transmitted from the dam to the neonates toward the terminal stages of gestation or postnatally via milk. According to Barber and Trees (27), vertical transmission of N. caninum in dogs is considered highly variable and not likely to persist in the absence of horizontal infection. In a prospective study, only 3% (4 of 118) of pups from 17 seropositive bitches were seropositive. Overall, 80% of pups born to seropositive bitches were considered to be uninfected with N. caninum (133). These results are supported by a recent study in which 3 of 11 pups in the first litter and only 1 of 7 pups in the second litter were infected with N. caninum (157). These results obtained with dogs are dramatically different from those obtained with cattle.

Age-related prevalence data indicate that the majority of dogs become infected after birth. Higher prevalences have been documented in older than in younger dogs (15, 45, 73, 117, 119, 290, 334a, 489).

In one report, 51% of 300 foxhounds fed bovine carcasses were found to have N. caninum antibodies (441). While consumption of aborted bovine fetuses does not appear to be an important source of N. caninum infection in dogs (48, 123), the consumption of bovine fetal membranes may be a source of N. caninum for dogs. The parasite has been found in naturally infected placentas (49, 172, 412), and dogs fed placentas from freshly calved seropositive cows may shed N. caninum oocysts (120). That dogs can become infected by ingesting infected tissues has been amply demonstrated (Table 9), but whether they can be infected by the ingestion of oocysts is unknown.

Transmission of N. caninum in Cattle

Transplacental (vertical) transmission. N. caninum is one of the most efficiently transplacentally transmitted parasites among all known microbes in cattle. In certain herds, virtually all calves are born infected but asymptomatic. Evidence for this efficient transplacental transmission comes from several sources: familial, comparison of antibody status in cows and their progeny, infection status of progeny, and experimental.

Björkman et al. (54) traced the familial history of N. caninum-seropositive dairy cows in a herd in Sweden and found that all infected animals were the progeny of two cows that were bought when the herd was established 16 years earlier. Insemination records suggested that venereal transmission was not a factor. Similar results were obtained in studies performed in Germany (391), Canada (47), Australia (201), and Sweden (176). A strong evidence for transplacental transmission of N. caninum has been obtained by comparison of seroprevalence in dams and their progeny. In cattle and other ruminants, there is no transfer of antibodies from the dam to the fetus, not even through a placenta that has been damaged by an infectious process (137). Therefore, detection of specific antibodies in precolostral serum indicates in utero synthesis of antibodies by the fetus. However, a finding of no antibody in the fetus is not conclusive of the absence of infection, because the fetus might have been infected late in gestation, leaving insufficient time for antibody synthesis. Rarely, it is possible for a seronegative dam to give birth to a seropositive calf; this may be because the cow has been infected for some time and the level of antibodies has declined to an undetectable level (85, 176, 281, 382).

Results obtained from studies with dam and progeny are summarized in Table 10. In this respect, precolostral data are noteworthy (Table 10). Up to 95% of calves were born infected. The actual congenital transmission rate was likely to be higher because, as stated above, a few positive calves are likely to be born from seronegative dams. The data from cow-calf pairs obtained after birth are not absolute, because mismatches are possible.


View this table:
[in this window]
[in a new window]

  		TABLE 10. Asymptomatic congenital transmission of N. caninum in cattle

 
Anderson et al. (11) provided convincing evidence that chronic persistent infection can be passed to progeny via endogenous transplacental transmission. In their study, 25 seronegative heifers were housed with 25 seropositive heifers beginning at birth, and their progeny were evaluated for N. caninum infection. The seronegative heifers remained seronegative and gave birth to calves not infected with N. caninum. The seropositive heifers remained clinically normal but gave birth to congenitally infected calves. Seven of these congenitally infected calves were necropsied; all had histologic evidence of N. caninum infection, and four were recumbent (11). Presumably, cows remain infected for life and transmit N. caninum infection to their offspring in several consecutive pregnancies (173) or intermittently (58, 197, 486). The rate of endogenous transplacental infection may decrease in subsequent pregnancies, indicating immunity (10, 125, 375).

Although exogenous transplacental N. caninum infection and abortion have been induced in cows experimentally infected with tachyzoites or oocysts by several research groups using many strains (158), little is known of the distribution and persistence of N. caninum in tissues of postnatally infected adult cattle.

Mathematical models of N. caninum infections within dairy herds (175) indicate that even low levels of horizontal transmission may be important in the maintenance of the infection within herds, because transmission by endogenous transplacental infection is below 100% and thus would lead to a continuous decrease in infection prevalence in the infected herds.

Post-natal (horizontal) transmission. The ingestion of sporulated N. caninum oocysts from the environment is the only demonstrated natural mode of infection in cattle after birth (111, 190, 443). To date, cow-to-cow transmission of N. caninum has not been observed. At present there is no evidence that live N. caninum is present in excretions or secretions of adult asymptomatic cows. Neonatal calves may become infected after ingestion of milk contaminated with tachyzoites (110, 446), and N. caninum-DNA in milk, including colostrum, has been demonstrated (316, 317). However, there is no conclusive evidence that lactogenic transmission of N. caninum occurs in nature (120).

Venereal transmission may be possible, but unlikely, as evidenced recently in heifers experimentally infected by intrauterine inoculation of semen contaminated with tachyzoites (408), and a dose response has been observed in a titration experiment with seroconversion and maintained antibody levels in heifers inoculated with semen contaminated with 5 x 104 tachyzoites (410). Although N. caninum DNA has been found in the semen of naturally exposed bulls (65, 166, 327), results suggest that viable organisms, if present, are few and infrequent. Additionally, cows inseminated with frozen and thawed semen contaminated with N. caninum tachyzoites failed to acquire infection (70).


   RISK FACTORS FOR BOVINE NEOSPOROSIS
Top
 Previous
 Next
 References
 
The knowledge of risk factors for herds to acquire N. caninum infection and N. caninum-associated abortion is important for the development and implementation of measures to control bovine neosporosis. Our knowledge of risk or protective factors with respect to bovine neosporosis is based largely on retrospective cross-sectional or case-control studies. Retrospective assessment generally allows the identification of putative risk or protective factors, but conclusive data can be obtained only by prospective cohort or experimental studies. However, the repeated identification of the same risk or protective factor in several independent retrospective cross-sectional or case-control studies increases the evidence that this factor is a "true" risk or protective factor for an infection or for a disease.

The serologic prevalences of N. caninum summarized in Tables 4 and 5 indicate that there are considerable differences among countries, within countries, between regions, and between beef and dairy cattle (39, 112, 250, 311, 359). However, caution should be used in evaluating these results because of differences in serologic techniques, study design, and sample size used. Data reported by Bartels et al. (39) are noteworthy because the sera were tested by standardized serological techniques (460) and similar study designs. From the data it is evident that the seroprevalence of N. caninum is lowest in Sweden, compared with prevalences in other European countries. Results suggest that there are differences in the infection risk among different regions, within a particular region, and among different management systems. Therefore, caution should be used when transferring the results of a risk factor analysis obtained in a particular region or management system to another. One example is that in a multivariate spatial regression analysis, the factors "abundance of coyotes" and "abundance of gray foxes" are both able to explain the differences between ecological regions regarding the N. caninum seroprevalence in beef calves (32). The possible importance of the factor "abundance of coyotes" was corroborated when coyotes were proven to be definitive hosts of N. caninum (188). However, this risk factor is definitively not relevant in European countries because there are no wild living coyotes in Europe.

Epidemic and Endemic N. caninum-Associated Abortion

N. caninum-associated abortion in bovine herds may have an epidemic or an endemic pattern. There are reports that in the years after an epidemic abortion outbreak, the affected herd may experience endemic abortions (56, 309, 352). Abortion outbreaks have been defined as epidemic if the abortion outbreak is temporary and if 15% of the cows at risk abort within 4 weeks, 12.5% of the cows abort within 8 weeks, and 10% of the cows abort within 6 weeks (309, 399, 488). In contrast, an abortion problem is regarded as endemic if it persists in the herd for several months or years. It is likely that these two patterns of N. caninum-associated abortion are related to two routes by which N. caninum infections can cause abortion (Fig. 3) (442).


Figure 3
View larger version (40K):
[in this window]
[in a new window]

  		FIG. 3. Overview of potential risk or protective factors influencing the horizontal or vertical transmission of Neospora caninum and the occurrence of exogenous or endogenous N. caninum-associated abortion. In this diagram, naïve cattle are gray, postnatally infected cattle are orange, and vertically infected cattle are red.

 
Epidemic abortions are thought to be due to a primary infection of naïve dams with N. caninum, probably due to ingestion of feed or water contaminated with oocysts (296, 297). Because pregnant dams may be exposed to contamination with oocysts almost at one time (point source exposure), exogenous transplacental fetal infection and the resulting abortions occur within a short period of time. The finding of low-avidity immunoglobulin G (IgG) responses, suggesting a recent infection (56, 57) in herds with epidemic abortion, supports this hypothesis (233, 296, 383, 399). Recrudescence of a latent infection in the dam during gestation (resulting in endogenous transplacental fetal infection) may cause abortion (197, 338, 422, 474).

Latent infection in dams may have been acquired vertically (11) or postnatally (309). The mechanism of reactivation of latent N. caninum infection is unknown. Whether immune suppression induced by ingestion of toxic feeds or other concurrent infections can cause reactivation has been debated but not supported by data (37, 352, 488). Recently it was shown that progesterone supplementation during midgestation increases the risk of abortion in Neospora-infected dairy cows with high antibody titers (46).

Irrespective of the origin of infection (exogenous or endogenous), not all congenitally infected fetuses die or become sick. In abortion epidemics, up to 57% of aborting dams have been reported (399, 488). However, in The Netherlands, high rates of seroconversion together with low-avidity responses were observed in a dairy herd, suggesting a recent exposure of this herd to N. caninum, though no increased abortion incidence was observed in this herd (122). If epidemic abortion is caused by an exposure to oocyst-contaminated feed or water, the observed variability regarding abortion risk may be explained by factors such as the infection dose (190), the pathogenicity of the parasite strain by which the animals became infected, and by the susceptibility of the dams (e.g., immune status, state of gestation) (190). However, nothing is known of the differences in pathogenicity of N. caninum isolates in cattle. Transplacental infection has been induced in cattle inoculated with N. caninum isolates from different sources (158).

In many cattle herds with endemic abortion due to neosporosis, there is often a positive association between the serostatus of mothers and their progeny; i.e., there is evidence that the major route of transmission in these herds is vertical (47, 54, 56, 121, 201, 391, 399, 436, 486). Several studies demonstrate that chronically infected seropositive cows can have more than a twofold-increased risk of abortion compared to seronegative dams (281, 338, 486). There are indications that the risk of endogenous abortion is influenced by the parity of the dams (284, 434). Thurmond and Hietala (434) observed a markedly increased abortion risk in congenitally infected heifers during their first gestation but not in later gestations, compared to the abortion risk in seronegative controls.

Risk Factor Studies

There are a number of risk factor studies assessing the risk of individual cattle or herds either becoming infected with N. caninum or experiencing N. caninum-associated abortions. We believe that these risks (infection risk and the abortion risk) are positively associated with each other but are influenced differently (Fig. 3). After exogenous transplacental transmission, the abortion risk might be influenced by, e.g., the number of oocysts ingested by the dam and the gestational stage (190), whereas the occurrence of abortions in endogenous transplacental transmission might be influenced by as-yet-unknown factors, e.g., the immune status of the dam.

Several studies have examined N. caninum infection risk at the herd level or animal level with the serostatus of herds or individual cattle (dams, calves) as dependent variables, i.e., as the target or outcome variable (Table 11). The results of these studies have been influenced by the sensitivity and specificity of the serological tests used. Fluctuations in the antibody levels of individual cattle during gestation, the gestational stage, or the gestation number could be a cause of variation (103, 173, 197, 236, 338, 360, 422). The use of seropositivity to identify infected cattle is simple but does not provide information on the viability of infection. Furthermore, rarely, an animal may be infected but seronegative, or a seropositive animal may not have a viable infection. In addition, seropositivity also provides no information on the route of infection (horizontal or vertical) or how recently the infection occurred. To partially overcome the latter problem, some risk factor studies have focused on herds with epidemic abortion (37, 124, 488).


View this table:
[in this window]
[in a new window]

  		TABLE 11. Putative risk and protective factors for N. caninum infections and abortions identified in epidemiological studies of dairy and beef cattle

 
Infection Risk

In the following, we summarize the results of studies that have assessed risk factors for infection on either the animal or herd level.

Age of cattle. The risk of being seropositive may increase with age or gestation number in beef and dairy cattle (160, 236, 371, 386), suggesting that horizontal transmission of N. caninum is of particular importance in some herds. Waldner et al. (465) reported a negative age effect on the prevalence of seropositive animals in dairy cattle in Canada. In the same study it was observed that the risk of being culled was significantly greater in seropositive than in seronegative cows, suggesting that selective culling could be a possible reason for the age effect. In a recent European study it was observed that the age effect on seropositivity in dairy cattle may vary in different study areas. In Spain, for instance, the risk of being seropositive increased with age, while in Sweden the situation was the opposite (39). It was hypothesized that the age effect might be influenced by variations in the probability of horizontal transmission (e.g., by the risk of ingesting oocysts), by regional differences regarding replacement rate (influencing the time cattle may be exposed to horizontal transmission), and by management practices such as selective culling of seropositive animals (39). Nonselective culling of animals in a herd with a high seroprevalence could result in a positive relationship between age and prevalence, if the population from which successive external replacement heifers are purchased has a lower seroprevalence than the herd itself. This effect is further strengthened by the fact that the proportion of vertical transmission is often much lower than 100% (106).

A British study of cattle in dairy herds with N. caninum-associated problems revealed a significantly lower seroprevalence in 13- to 24-month-old animals than in cattle 7 to 12 months old and cattle older than 24 months (107). It was hypothesized that some of the 13- to 24-month-old animals (most likely heifers) were congenitally infected with N. caninum, although they were seronegative. Recrudescence during gestation may have caused an elevated seroprevalence in older age groups (107).

Definitive hosts (dogs and coyotes). In most epidemiological studies of dairy herds, the presence of farm dogs, either currently or within the past 10 years (339, 461), or the number of farm dogs (93, 289, 339, 402, 461) was a risk factor for seropositivity in cattle. This is not surprising, as dogs are definitive hosts of N. caninum. Furthermore, the putative ways by which dogs may pose an infection risk to dairy cattle have been studied (123). Defecation by farm dogs on feeding alleys and on stored grass or corn silage was reported more often by farmers of herds with evidence of postnatal bovine infection than by those of herds with no such evidence (123). Interestingly, in a study of herds with evidence of recent postnatal infection, seropositivity to N. caninum was more often associated with common housing than with common feeding of the seropositive age group (124). Based on these results, it may be justified to assume that contaminations of the feeding area are more closely related to infection than are contaminations of fodder during storage.

Farmers of herds with evidence of postnatal infection more often observed dogs feeding on bovine placenta, uterine discharge, and colostrum or milk than did farmers of control herds (123). This suggests that these materials may pose an infection risk to dogs; i.e., these materials may facilitate dogs becoming infected with N. caninum. In an experimental study, placenta, but not colostrum, has been confirmed as an infection source for dogs (120). Interestingly, feeding on aborted fetuses was not identified as a potential risk factor in herds with evidence of recent postnatal infection (123), and no oocyst shedding was observed when aborted fetuses or brains of fetuses were fed to dogs experimentally (48). However, these results were most likely influenced by the stage of autolysis in the fetus, killing the parasite along with the host cells. Most N. caninum organisms in aborted fetuses die with the host cells, and it is rare to find intact tachyzoites in such tissues (158). Conrad et al. (86) were able to isolate viable N. caninum parasites from only 2 of 49 histologically confirmed fetuses. Dogs have shed oocysts after ingesting a variety of tissues, including neural, muscular, visceral, and fetal membranes (Table 9).

There is some evidence that recently introduced dogs pose a higher risk of transmission of N. caninum than do resident dogs (124). This could be explained by analogy to T. gondii, for which it is well known that naïve definitive hosts are crucial for the life cycle (105). In N. caninum, the situation seems to be similar, as dogs shed no or only few oocysts after being fed repeatedly with infectious material (120, 191, 397). Additionally, higher oocyst numbers are shed by young dogs (10 to 14 weeks old) than by older dogs (2 to 3 years old) (191).

In addition to farm dogs, dogs kept in the neighborhood of farms may pose an infection risk. In a German cross-sectional study, dog densities in districts, cities, or municipalities were predictors of the prevalence of bulk-milk-positive herds (400) or were identified as risk factors for herd seropositivity (402, 461). Recently, coyotes were found to be additional definitive hosts of N. caninum. This was suspected after epidemiological studies of beef calves had shown that the abundance of coyotes or gray foxes in different ecological zones of Texas was associated with the seroprevalence of N. caninum in beef calves (32). Whether gray foxes are also definitive hosts of N. caninum remains to be determined. Although one experimental study indicates that the red fox is not a definitive host for N. caninum (398), there is an ongoing discussion as to whether red foxes or wolves could be important as sources of postnatal infections with N. caninum, and N. caninum-like oocysts in the feces of naturally infected foxes from Canada were reported (471). Recently, it was hypothesized that wolves, because of their close phylogenetic relationship to dogs, may be another potential definitive host of N. caninum (188). The sylvatic (deer-canid) cycle may be important in maintaining the domestic (cattle-dog) cycle of the parasite (189).

For beef cattle, there is as yet no evidence that farm dogs or dogs kept in the surroundings of farms pose an infection risk (461). A possible explanation for this is that on the less intensively managed beef farms, there is in general no close contact between the excretions of farm dogs and beef cattle (33, 332, 386). Moreover, Barling et al. (33) observed that the presence of farm dogs on beef farms was a putative protective factor. That study was conducted in Texas, i.e., in the same region where it was demonstrated that the abundance of wild canids could explain the seroprevalences in beef calves. Possibly the presence of dogs was inversely related to the presence of wild canids on farm land, as suggested by Hobson et al. (218).

Other carnivores. In experimental studies, cats failed to serve as definitive hosts for N. caninum (295). Interestingly, there is one epidemiological study of dairy cattle that observed a protective effect for the presence of cats on a farm (333). It is possible that this factor is a confounder related to the absence of dogs. However, another possible explanation for the protective effect of the factor "presence of cats" is that cats are predators of putative intermediate hosts of N. caninum (e.g., mice), which could reduce the frequency by which definitive hosts of N. caninum have access to the tissues of infected intermediate hosts.

Intermediate hosts other than cattle. Not only cattle but also other intermediate hosts of N. caninum may present a source of infection for dogs and other canids. The presence of N. caninum DNA in naturally infected mice and rats suggests that these animals may be important sources of infection for carnivore hosts of N. caninum (Table 2). One study from France reported the presence of rabbits and/or ducks as a putative risk factor for seropositivity in dairy cattle (333). In a study from northern Italy, the risk of seropositivity in individual cattle increased with the number of farm dogs when poultry were present on the farm (332). Bartels et al. (37) also found the presence of poultry on the farm to be a risk factor for the occurrence of N. caninum-associated abortion and discussed their possible role as a vector of canine oocysts. These results warrant further examination of the susceptibility of rabbits, ducks, and other poultry to N. caninum and whether these potential intermediate hosts pose an infection risk to definitive hosts.

Grazing, fodder, and drinking water. Oocyst-contaminated pastures, fodder, and drinking water are regarded as potential sources for postnatal infection of cattle. Therefore, it is important to know which feeding practices pose an increased infection risk.

In the northwestern United States and Italy, grazing of cattle on rangeland during summer seems to be a protective factor (332, 386). Although wild canids and dogs have free access to rangeland, oocyst contaminations caused by definitive hosts may be too low to pose a significant infection risk or oocysts may not survive during the summer months if they are very hot and dry. Unfortunately, information on the climatic conditions under which N. caninum oocysts are able to survive in the environment is rare.

In beef herds, the use of a hay ring appeared to be a putative risk factor for seropositivity (33). This factor was explained by the observation that cows often calve, abort, or expel placentas near hay feeders. Because these feeders are seldom moved, it was hypothesized that fecal contaminations by definitive hosts that have fed on placentas may be concentrated close to the feeders (33). In the same study, a procedure implemented to avoid the contamination of fodder, i.e., the use of a self-contained feeder for cow supplements, was identified as a probable protective factor (33). Related to this is the observation that ranches with wildlife access to the weaning supplement had an increased risk of calves being N. caninum positive (33).

In a study conducted in France, the use of ponds rather than the use of a well or public water supply for drinking water was found to be a risk factor for N. caninum infection in dairy cattle (333). Seroprevalence data from feral marine mammals suggests that N. caninum oocysts may contaminate surface water and subsequently contaminate seawater (131, 154). Outbreaks of toxoplasmosis in humans have been linked epidemiologically to contaminated drinking water, and T. gondii has been isolated from municipal waters (60, 116).

Feeding colostrum or milk. Experimental studies have demonstrated that neonatal calves may become infected by the ingestion of milk containing tachyzoites (110, 446). However, cross-suckling of calves born to seronegative mothers on seropositive cows has not led to an infection (110). Because N. caninum DNA was found in bovine milk (316, 317), there is an ongoing debate regarding whether or not the lactogenic transmission of N. caninum is possible. With respect to this, it is interesting that one study in dairy cattle has suggested that feeding of pooled colostrum is a putative risk factor for seropositivity (93).

Calving management. In one risk factor analysis of beef calves in Texas, the effect of seasonal calving during spring was profound; i.e., the risk of calves of being seropositive was higher than it was on ranches with a fall calving season (33). No explanation for this observation was offered. Possibly, there are seasonal effects in these beef herds on the risk for calves to become infected, either by transplacental or by horizontal (postnatal) transmission. This seasonality may be biologically linked to the whelping season of the putative definitive hosts in Texas, coyotes and gray foxes. Since, naïve or young dogs are more submissive definitive hosts for N. caninum than are older or immune dogs (120, 191, 397), the same may also be true for young coyotes and gray foxes. Further studies are needed to explain the observations with Texas beef calves. Interestingly, in a French study, prolonged herd calving periods of 3 to 6 or 6 to 12 months reduced the risk of herd seropositivity compared to herd calving periods of up to only 3 months (333). There was no explanation for this observation.

Cattle stocking density and size of farmland. In two studies of beef calves in Texas, a high stocking density was identified as a potential risk factor for seropositivity (32, 33). A similar effect was observed for the stocking density of beef cows during winter in the northwestern United States (Idaho, Montana, Oregon, Washington, and Wyoming) (386). This effect was explained by the observation that ranches with a high density of cattle are more likely to use supplemental feeding practices (32, 33). Places on farms were supplemental feed is stored or fed to cattle may attract rodents that are potential prey for definitive hosts of N. caninum. This could cause these places to have an increased risk of being contaminated with the feces of definitive hosts, thus increasing the risk of postnatal infection (32).

In a study of dairy cattle in southern Brazil, it was observed that with increasing size of farmland, the seroprevalence in herds decreased. However, this protective effect was not linked to the stocking density (93). It was hypothesized that on small farms it is easier for farm dogs to have access to bovine carcasses, aborted fetuses, placenta, and uterine discharge than on larger farms.

Herd size. In a study from Italy, the risk of individual cattle becoming seropositive increased with the size of the herd. When the analysis was restricted to data from northern Italy, the number of dogs per farm interacted significantly with herd size; i.e., the risk of being seropositive increased in larger herds with an increasing number of dogs per farm (332). In a study conducted in Germany, larger herds had an increased risk of being bulk milk positive (402). Possible explanations are that with increasing size of the herd there is an increasing chance of acquiring N. caninum infection by, for instance, the purchase of external replacement heifers. Another explanation for herd size as a risk factor could be that hygienic measures to prevent dogs from feeding on placentas or other infectious material are more difficult to follow with large herds than with small herds (402).

Source of replacement heifers. The vertical transmission of N. caninum is very efficient. Thus, the rearing of replacement heifers on the farm rather than purchasing them from outside sources supports the contention that an existing prevalence in a herd may persist for many years (176, 423). If the seroprevalence is higher in the recipient herd than in the population from which the replacement heifers were obtained, the purchase of replacement heifers should reduce infection in the recipient herd. This could explain why, in one of the risk factor studies of beef cattle, "rearing of own replacement heifers" was identified as a potential risk factor for a high seroprevalence in calves (33).

Climate. In two European studies that analyzed climate effects on the risk of seropositivity in herds or individual cattle, the factors "mean temperature in spring in a buffer zone around farm location" and "mean temperature in July in the municipality where the herd is localized" were identified as putative risk factors (371, 402). These observations can be explained by the effects of climate on sporulation or survival of oocysts. For example, a higher temperature (up to not-yet-defined limits) may favor a faster sporulation of oocysts in fodder or in the environment surrounding the cattle.

Vegetation index. An Italian study observed that the risk of seropositivity in individual cattle decreased with increasing summer normalized difference vegetation index (NDVI) values determined for 3-km buffer zones around the farm location (371). A high summer NDVI is indicative of forests or broadleaved trees. It was assumed that cattle from the respective farms were not pastured and thus had a smaller chance of ingesting N. caninum oocysts. However, this interpretation is not supported by the finding of another Italian study, in which "no grazing" was identified as a risk factor for seropositivity in individual cattle (332).

Human population density. In Germany, human population density was correlated positively with dog density and could, like dog density, be used to predict the prevalence of bulk-milk-positive herds in districts and cities (400). Because dog density was identified as a putative risk factor for infection, it is not surprising that human population density seems to have the same effect.

Factors related to antibodies against other infectious agents. Björkman et al. (55) observed in Swedish cows a statistically significant association between antibodies against N. caninum and bovine viral diarrhea virus (BVDV). From this result it was assumed that risk factors supporting the introduction and spread of BVDV in cattle, such as high cattle density and frequent purchase of animals, also increase the risk of N. caninum infection. In an Italian study, a positive association between antibodies against bovine herpesvirus 1 (BHV-1) and antibodies against N. caninum was demonstrated (372). The possibility of whether BHV-1-induced immunosuppression after natural infection or vaccination could increase the susceptibility of cattle to secondary infection with N. caninum was discussed. However, to prove this hypothesis, experimental or follow-up studies after infection or vaccination are necessary (372). In a Canadian study of 78 dairy herds in Ontario, no significant association between antibodies against N. caninum and serostatus to Leptospira interrogans serovar Hardjo, Icterohaemorrhagiae, or Pomona was observed (343).

Breed. There are indications from several countries that N. caninum seroprevalences differ according to the cattle breed (39). However, these results must be interpreted with caution, because the differences observed might have been caused by differences in the production systems used for the different breeds and not by differences in breed-related susceptibility to infection. For example, native Spanish breeds were less likely to be seropositive than Holstein Friesian, Rubia Gallega, or mixed breeds. This was explained by differences in the intensity of management (39): in contrast to Holstein Friesian and Rubia Gallega cattle, which in Spain are more intensively managed, native breeds are predominately located on highland pastures with very low stocking densities. In the same study, breed-associated differences from Sweden were reported.

Type of housing. In a French study, tethered dairy cattle had a higher risk of being seropositive than did dairy cattle kept untethered indoors (333). No explanation for this effect was offered.

Abortion Risk

Factors having an effect on the occurrence of epidemic abortion outbreaks may completely differ from those influencing the risk of endemic abortions. Risk factor analyses often have the disadvantage that there is no information regarding the context (epidemic or endemic) in which the abortions occurred. Consequently, it is not possible to assign the risk or protective factors identified in epidemiological studies to the occurrence of epidemic or endemic abortions. Some risk factor analyses are based on case-control studies limited to herds with epidemic outbreaks (37, 488); therefore, the risk factors identified in such studies can be related only to the occurrence of epidemic abortions.

Seropositivity of individual cattle. Seropositive cows are more likely to abort than are seronegative cows, as demonstrated in a large number of studies, including retrospective and prospective cohort studies (10, 92, 107, 109, 180, 206, 213, 236, 249, 281, 282, 289, 309, 312, 315, 338, 391, 393, 394, 399, 423, 436, 447, 464, 474).

The strength of the association between seropositivity and abortion in a single group of animals may vary considerably if different serological assays are used or if for the same assay different cutoffs values are applied (392, 465). Consequently the estimates for odds ratios or relative risks may vary in relation to the serological test applied.

The abortion risk increases with increasing levels of N. caninum-specific antibodies in individual animals (239, 285, 293, 360, 393, 394, 423, 464,