|Home | About | Journals | Submit | Contact Us | Français|
The protozoan pathogen Neospora caninum is recognized as a leading cause of infectious abortions in cattle worldwide. To evaluate the impact of neosporosis on dairy and beef herd production, a retrospective, longitudinal study was performed to identify the impact of neosporosis alongside other causes of fetal abortion in British Columbia, Canada. Retrospective analysis of pathology records of bovine fetal submissions submitted to the Animal Health Centre, Abbotsford, British Columbia, a provincial veterinary diagnostic laboratory, from January 2007– July 2013 identified 182 abortion cases (passive surveillance). From July 2013–May 2014, an active surveillance program identified a further 54 abortion cases from dairy farmers in the Upper Fraser Valley, British Columbia. Of the total 236 fetal submissions analyzed, N. caninum was diagnosed in 18.2% of cases, making it the most commonly identified infectious agent associated with fetal loss. During active surveillance, N. caninum was associated with 41% of fetuses submitted compared to 13.3% during passive surveillance (P<0.001). Breed of dam was significantly associated with N. caninum diagnosis, with a higher prevalence in dairy versus beef breeds, and fetuses of 3–6 months gestational age had the highest prevalence of N. caninum. There was no significant association with dam parity. Neospora caninum was diagnosed in every year except 2009 and cases were geographically widespread throughout the province. Furthermore, the active surveillance program demonstrates that N. caninum is highly prevalent in the Upper Fraser Valley and is a major causal agent of production losses in this dairy intensive region.
Neospora caninum is recognized as a significant cause of bovine abortions worldwide (Dubey and Schares, 2011). N. caninum has a complex life cycle: the definitive canid host sheds highly infectious and stable oocysts in their feces (Dubey et al., 2007) which can be transmitted horizontally to infect a wide variety of intermediate hosts, including cattle, deer, sheep, buffalo, bison, rodents, birds and horses (Dubey and Schares, 2011). Here, the parasite expands asexually to produce a chronic infection harboring infectious cysts that are transmissible to canids. In cattle, it is also efficiently transmitted vertically by transplacental infection, with congenital transmission rates recorded to be 40–95% among calves born to seropositive mothers (Reichel et al. 2014).
Abortion and stillbirth result in significant production and economic loss in the beef and dairy industries (De Vries, 2006; Jonker, 2004). The extent to which transmission occurs by exogenous exposure to oocysts versus transplacental infection from a chronically infected seropositive dam is necessary information to manage the spread of this disease among cattle (Basso et al., 2010; Dubey et al., 2007). Dogs, coyotes, wolves and dingoes are the known definitive hosts of N. caninum, but experimentally infected canids do not appear to shed substantial numbers of oocysts (Dubey and Schares, 2011). Due to the apparent low fecundity of the N. caninum sexual cycle, it is thought that the primary driver of the high transmission rates in cattle herds is vertical transmission (Barr et al., 1990; Cardoso et al., 2012; Morales et al., 2001; Pare et al., 1996). However, recent studies have shown significant variation in transplacental infection rates (Haddad et al., 2005), and in some cases, there is evidence of exogenous transmission and post-natal seroconversion among herds where N. caninum circulates endemically (Basso et al., 2010; Dijkstra et al., 2001; Hall et al., 2005). Furthermore, the efficiency of vertical transmission decreases with parity (or the number of pregnancies post-infection) and the occurrence of sporadic, clustered outbreaks of abortions (epidemic pattern) suggest that exogenous exposure to the oocyst stage of the parasite (Anderson et al., 2000; Bartels et al., 2006) may be more significant than previously envisaged (Dubey and Schares, 2006; Thurmond and Hietala, 1997). Ultimately, to determine the exact transmission dynamics of N. caninum associated fetal loss, a robust assay and reliable baseline data in a region endemic with disease needs to be developed in order to evaluate the extent to which each transmission cycle contributes to the spread of disease. Such baseline data is important to begin to understand how the parasite circulates in nature, the impact it has economically, and the best control methods that could be adopted in order to limit disease (Reichel et al., 2014).
Despite the significant contribution of N. caninum to reproductive losses, a definitive diagnosis of N. caninum infection remains a major challenge requiring multiple diagnostic approaches (Dubey and Schares, 2006). Historically, immunohistochemistry (IHC) has been utilized because it is very specific, however, it is an insensitive method of identification so many positive diagnoses are missed (Jenkins et al., 2002). Recently, the sensitivity and specificity of PCR methods have improved diagnosis (Ortega-Mora et al., 2006), but because transplacental infection does not always result in fetal loss, no diagnosis can be made on congenitally infected but clinically normal calves (Dubey and Schares, 2011; Sager et al., 2001).
In Canada, seroprevalence ranges from 5–25% among beef and dairy cattle (Haddad et al., 2005; Pruvot et al., 2014). However, few studies have investigated the extent to which N. caninum infection causes fetal abortion in Canada. Two studies that utilized pathology and/or IHC methodologies reported an abortion rate of 2–5% (Alves et al., 1996; Khodakaram-Tafti and Ikede, 2005). In British Columbia, Canada, little is known about the prevalence and impact of this important production pathogen (Haddad et al., 2005). Our aim was to 1) establish a baseline dataset that measures the relative impact of N. caninum associated fetal loss compared to other agents associated with infection, and 2) determine the risk factors associated with infectious loss. To do this, we evaluated a suite of veterinary diagnostic and pathologic tools including fetal pathological examination, PCR, IHC and serology in combination with epidemiologic data to identify etiologies associated with fetal submissions to the Animal Health Centre over a 7 year period (Abbotsford, British Columbia).
The Animal Health Centre, Abbotsford, BC is the only veterinary laboratory in the province performing gross post-mortems. It is the only American Association of Veterinary Laboratory Diagnosticians (AAVLD)-accredited veterinary diagnostic laboratory in the province, and it accepts the majority of case material from veterinarians and producers in BC. Pathology records of bovine fetuses submitted from January 1, 2007 – June 1, 2014 were retrieved through a database search (Vetstar Animal Disease Diagnostic System (VADDS); Advanced Technology Corporation, Ramsey, New Jersey, USA) with “bovine fetus” and “abortion” as search criteria. Data extracted from the pathology reports was entered into Microsoft Access and Excel (Microsoft Corporation, Redmond, Washington, USA). In addition to routine submissions, as part of Agriculture and Agri-food Canada’s Growing Forward-2 Initiative, dairy farmers in the Upper Fraser Valley Region were actively recruited from July 1, 2013 to May 16, 2014 to submit all of their aborted fetuses, placenta, milk and serum from the dam (active surveillance) for diagnostic work-up. The active surveillance program was initiated by personal communications with farms by neighboring veterinary practices, clinic newsletters, email, and seminar/client meetings.
Upon presentation, fetuses were examined for external abnormalities. The sex, weight and crown-to-rump length (CRL) were recorded, and a standard suite of tissues was harvested for histologic analysis and ancillary diagnostic studies (Kirkbride, 2012). If specific lesions or inflammatory infiltrate was identified by the pathologist consistent with an infectious process, immunohistochemistry (IHC) and histochemical stains were used to demonstrate or confirm the presence of the following pathogens, including N. caninum (mouse anti-N. caninum monoclonal antibody (VMRD)), bovine herpesvirus I (BHV-1), and intralesional bacteria and fungi. Aerobic bacterial culture of lung, placenta, and stomach content was part of the standard fetal work protocol. Additional tests were conducted as directed by the assigned pathologist, including aerobic and anaerobic culture and pathogen identification by polymerase chain reaction (PCR), tissue culture (TC), serum neutralization test (SN) or fluorescent antibody tests (FAT). Specialized diagnostics for other viruses, protozoa or bacteria were completed upon request of the pathologist. The majority of neosporosis cases possessed myocardial lesions and epicarditis, and lesions were typically observed in the heart, brain, and liver. Molecular diagnosis of neosporosis was done using the N. caninum-specific primer pair Np2l/Np6 (Yamage et al., 1996) for cases submitted before July 2013. After this period, ApiITS1 primers that target a region within the 110-copy ribosomal SSU rDNA gene array followed by DNA sequencing of the positive amplicon (Gibson et al., 2011) was performed due to the increased sensitivity and specificity of this method. A diagnosis of neosporosis was made if any tissue (brain, diaphragm, heart, liver, lung, muscle, placenta, pooled tissues, or tongue) tested PCR positive for N. caninum. A commercially available competitive inhibition ELISA test (N. caninum Antibody Test Kit, cELISA, VMRD Inc., Pullman, WA, USA) was used to test dams for N. caninum antibodies, but this test was not routinely requested.
We analyzed data from all fetal submissions associated with fetal loss, including abortions (a fetus which was less than 260 days old or of unknown age, n = 212) and stillborn cases (260 days old to full term, n = 22). Fetal age of submissions ranged from 3 months to near term abortions and stillbirths. Age was based on crown-to-rump length (CRL) and owner submitted information when CRL measurements were not available (Forar et al., 1996; McGeady et al., 2013). CRL was maximized to agreement between submitted gestation length and measured CRL with fetal age in days correlated to the CRL (range in cm) as follows; 60 (0–15), 90 (15–27), 120 (27–40), 150 (40–52), 180 (52–70), 210 (70–87), 260 (>87).
The cause of fetal death was classified based on the pathologists’ interpretation, which reflected gross and histologic examinations and review of results from ancillary diagnostics tests. We diagnosed N. caninum based on histopathology and an additional positive ancillary diagnostic test (IHC, PCR, or serology) or PCR testing alone. Fetal loss was assigned to one of the following five categories: confirmed infectious (an etiological agent was diagnosed), inflammatory (lesions indicative of infectious processes were observed but no infectious agent was diagnosed), nutritional (histologic lesions were observed indicative of iodine, vitamin A or vitamin E/Se deficiency), developmental (gross malformations or developmental lesions were observed and no pathogens were identified) or idiopathic (no gross or histologic lesions or significant pathogens were identified).
Based on available data for each case, we categorized fetal submissions by breed (Beef: Charolais, Hereford, Angus, Brahma vs. Dairy: Holstein, Jersey, Ayershire, Brown Swiss, Crossbred), dam parity (heifers: first pregnancy vs. multiparous dams: cows that had been calved at least once before), and gestational age (3–6 months vs. greater than 6 months). N caninum diagnoses were compared between these separate groups by chi-squared tests. All statistical analyses were performed in Prism (Version 6, GraphPAD Software, San Diego, CA). Additionally, we compared the frequency of histological lesions observed across different tissue sites (adrenal, brain, kidney, liver, muscle, heart, and other tissue sites combined) for Neospora caninum cases (n = 43). We also analyzed the number of N. caninum cases diagnosed by month and year. To examine the spatial distribution of submissions and N. caninum occurrence, fetal specimens were linked to a farm location and farming region as presented in the BC Holstein directory 2013–2014 (British Columbia Holstein-News, 2013/2014) or openly available on Google Maps. Distribution maps were constructed in R 3.01 (R Core Development Team) using the package RgoogleMaps (2015) with a 5% randomization added to farm location GPS coordinates. Occurrence data for beef farms was not presented due to limited data availability.
Between April 16, 2007 and May 16, 2014, 237 bovine fetuses were submitted to the AHC from across British Columbia, Canada. One case was excluded from the study due to missing observations. In total, 146 (61.9%) cases of fetal loss were attributable to an infectious process (Table 1). Of these, we classified 55 (23.3%) as inflammatory, where histopathological evidence suggested an infectious process but no causative pathogen was identified. Ninety-three (39.4%) were diagnosed as a confirmed infectious case where the infectious agent was identified. Among these cases, N. caninum (n = 43) was the primary agent diagnosed, representing 46.2% of infectious cases and 18.2% of all cases. Other pathogens identified included bacterial (n = 36), viral (n = 10), and fungal (n = 2) agents (Table 1). Fetal loss due to nutrition (n = 6) or reproductive development (n = 4) was limited. We classified 80 (33.9%) cases as idiopathic, as there was no pathologic findings or pathogens identified. Focused active surveillance and recruitment from dairy farmers in the Upper Fraser Valley (UFV) resulted in 54 fetal submissions in 2013 and 2014, and approximately 41% (n = 22) of these were diagnosed as N. caninum compared to 13.3% (n=4 out of 30 total submissions from the UFV) during the routine surveillance period, a significant increase in diagnoses (χ2 = 6.778, p=0.009).
Of the 43 N. caninum cases, the pathologist noted lesions indicative of N. caninum infection alongside the positive PCR diagnostic result for 37 cases; 6 diagnoses were made on PCR results alone. For the N. caninum cases, lesions were most commonly observed in the heart (68.9%, n=31), liver (60.0%, n=27), and brain (46.7%, n=21) (data not shown). More specifically, myocarditis was the most commonly associated lesion type presenting in 29/31 cases with lesions in the heart. For all other sites (adrenal, kidney, lung, muscle, placenta, thymus, and spleen), lesions were observed in less than 25% of N. caninum cases. For 14 cases, lesions indicative of N. caninum infection were identified but N. caninum was not detected by an ancillary diagnostic test (n=12) or no ancillary diagnostics were done (n=2).
Breed data was submitted for 218 cases, with 180 identified as dairy breeds (Holstein = 173, other = 7) and 38 identified as beef breeds. N caninum was diagnosed more commonly in dairy fetuses (20.0 %) compared to beef fetuses (5.3%) (χ2 = 4.73, p = 0.03). Gestational age was also significantly associated with N. caninum infection, with positive diagnosis occurring in 26.9% of fetuses between the ages of 3–6 months versus 14.2% of fetuses older than 6 months (χ2 = 5.21, p = 0.02). The parity of the dam was known in 55 cases, and categorized as born from heifers or multiparous cows. N caninum was diagnosed in 28.6% of fetuses born from heifers versus 33.3% of fetuses born from multiparous cows, and the difference was not significant (χ2 = 0.024, p = 0.88).
The proportion of N. caninum diagnoses peaked in the autumn months, (September – November; 53%, n=23) and represented the majority of cases associated with fetal loss (χ2 = 11.26, p= <0.001) compared to all other months combined (December – August; 47%, n=20). No N. caninum diagnoses were made in May or August. N caninum cases occurred every year except 2009. Submissions were collected from six regions in British Columbia (Central, Lower, and Upper Fraser Valley, Richmond Delta, North Okanagan, and Vancouver Island). The majority of diagnoses were in the Fraser River Valley where the most intensive production occurs, but in general, the occurrence of N. caninum infection was widespread throughout the province (Figure 1).
In this largely retrospective, longitudinal study, N. caninum was identified as the leading infectious agent associated with fetal loss, with 18.2% of cases diagnosed as primary N. caninum infection. This is significantly higher than previously reported for Canada, although this is likely due to differences in the sensitivity of the diagnostic techniques utilized as previous studies primarily used IHC, a diagnostic method with significantly lower sensitivity than PCR (Alves et al., 1996; Khodakaram-Tafti and Ikede, 2005). Studies in California, New Zealand, and the Netherlands have shown similar rates of N. caninum associated abortions as we found in this study, with ~20% of bovine fetal abortion due to neosporosis (Anderson et al., 2000). Importantly, the active surveillance program dramatically improved diagnostic success, with 41% of cases diagnosed as neosporosis. We presume that the immediate responses from farmers after they detected a fetal abortion led to higher-quality submissions (i.e. fetuses that were submitted immediately after abortions or were stored properly) for detection by the PCR methodology used.
In our estimate, we included cases where PCR testing alone was positive for N. caninum since a definitive diagnosis of N. caninum infection is a major challenge (Jenkins et al., 2002). However, it is also important to note that N. caninum DNA has previously been identified in congenitally infected but clinically normal calves, which indicates that the detection of N. caninum DNA in an aborted fetus does not necessarily indicate causality. However, the lack of identification of other pathogens associated with fetal loss in the PCR alone positive cases strongly argues for the N. caninum infection as the primary or contributing factor that led to the fetal loss.
Overall, the majority of cases diagnosed as N. caninum had histologic lesions that were characteristic of protozoal infection, with lesions noted in the heart, brain, and liver, with myocarditis especially characteristic of diagnosed cases. These lesions and lesion sites have been shown previously to be highly associated with N. caninum infection (Anderson et al., 1990; Collantes-Fernández et al., 2006; Ortega-Mora et al., 2006; Otter et al., 1997). Twelve cases that had inflammatory lesions suggestive of N. caninum infection were identified, but these cases were negative on PCR testing. False negatives are not uncommon for a variety of reasons, including PCR test performance as a result of the quality and type of tissues submitted, or whether the tissue sampled contained parasite DNA (Ortega-Mora et al., 2006). While these cases are likely due to N. caninum, whether other closely related protozoa were causal agents, such as Sarcocystis spp. or Toxoplasma gondii, both of which have similar patterns of pathogenesis and cause fetal loss (Jenkins et al., 2002), was not systematically addressed in this study.
The occurrence of fetal loss due to neosporosis was geographically widespread, and the observed clustering of cases in the Upper Fraser Valley was likely due to the high number of submissions in this region after the active surveillance program was initiated. Other factors significantly associated with N. caninum diagnosis were 1) time of year, with more infections occurring in the Fall; 2) breed, with dairy fetuses more frequently diagnosed with fetal abortion due to N. caninum than beef breeds; and, 3) gestational age, with the greatest proportion of N. caninum fetuses aborted between 3–6 months. These results are in line with most studies of N. caninum abortion risk, in which higher rates have previously been found in dairy versus beef herds globally (Reichel et al., 2013) and the majority of fetuses aborting were between 5–7 months of gestation (Dubey and Schares, 2006). In our study the majority of cases submitted were from dairy breeds, and in general data on N. caninum prevalence in beef operations are scarce compared to dairy cattle globally (Reichel et al., 2013). Parity was not significantly associated with N. caninum diagnosis, and different studies have identified increasing dam age/parity as both a risk factor and a protective factor for N. caninum abortions (Thurmond and Hietala, 1997; Wouda et al., 1999) potentially due to a difference in epidemiologic patterns of N. caninum-associated abortion where increasing dam age is associated with increased epidemic abortions but decreased endemic abortions (Collantes-Fernández et al., 2006; Dubey et al., 2007). Whether environmental contamination of water or silage with oocysts shed from canid definitive hosts has contributed to the high incidence of fetal loss as a result of N. caninum infection could not be ascertained from this study. Certainly, recent studies have shown that abortion risk is associated with the presence or proximity of dogs (Dubey and Schares, 2011; Pruvot et al., 2014). Future work should be done to examine the incidence rate of Neospora infection in wild and domestic canids that co-associate with cattle in the Upper Fraser Valley, in addition to developing a serological test that identifies dams that are acutely infected by the presence of IgM and low-avidity IgG antibodies.
In our study, N. caninum was the most commonly diagnosed infectious etiology associated with bovine fetal loss, representing a significant cause of economic loss to the dairy industry in British Columbia, Canada. Furthermore, we found that N. caninum was geographically and temporally widespread, suggesting that this pathogen circulates endemically in this region. Focused active surveillance significantly increased the success rate of making a definitive diagnosis and it identified the Upper Fraser Valley as a region with high occurrence of fetal loss due to N. caninum infection. Future targeted surveillance of dairy farms in this region, including determining whether infection is the result of endogenous or exogenous transmission, and the identification of risk factors associated with N. caninum infection and associated fetal loss will provide critical insight into how this pathogen is being maintained in cattle populations and will help to elucidate optimal control strategies to manage this high-impact production pathogen.
The data in this paper were collected with the support of the Agriculture and Agri-food Canada’s Growing Forward-2 initiative and the Intramural Research Program of the NIH and NIAID (MEG). MEG is a scholar of the Canadian Institute for Advanced Research (CIFAR) program for Integrated Microbial Diversity. The authors acknowledge the professional and technical staff from the AHC, BC and NIH, MD, along with the veterinarians and staff of Greenbelt Veterinary Services, and the Fraser Valley Dairy producers for their dedicated contributions to the project.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.