Neospora caninum antibodies in dairy cows and domestic dogs from Vojvodina, Serbia

Neospora caninum, the causative agent of neosporosis, is a protozoan parasite responsible for high rate of abortion in cattle worldwide. In dogs, consequences of infection vary from severe neuromuscular disorders to asymptomatic infection and shedding of environmentally resistant oocysts. In this study, we determined the occurrence of N. caninum antibodies in dairy cattle and dogs in Vojvodina (Northern Province of Serbia) and possible risk factors. N. caninum antibodies were found in 15.4% (55/356, CI 95%:12.0–19.6) of cows and 17.2% (17/99, CI 95%: 10.8–26.2) of dogs. Cows from smallholdings showed significantly greater odds (OR = 5.28, CI 95%: 2.0–13.6, p = 0.0006) of being seropositive in comparison to the farm cows. Epidemiological importance of results is discussed.


Introduction
The protozoan parasite Neospora caninum causes considerable economic losses to cattle industry worldwide [1].
One of the definitive hosts of N. caninum, which is responsible for shedding of environmentally resistant oocysts, is the domestic dog [6]. In dogs, neosporosis can cause severe clinical manifestations, including myositis-polyradiculoneuritis, encephalomyelitis and dermatitis [7,12].
Worldwide reports of clinical and subclinical infections in all hosts were summarized by Dubey et al. [1] and Dubey and Schares [2], but there was no mention of neosporosis in Serbia. Recently, occurrence of N. caninum specific antibodies in both cattle and dogs has been confirmed in Serbia as well [3,9,13,17] and published in local journals. The aim of this study was to investigate current serological status of dairy cattle and dogs in Vojvodina (Northern Province of Serbia), with regard to the possible risk factors.

Studied area
Vojvodina is a northern (45°15 0 N 19°50 0 E) province of the Republic of Serbia which occupies 21,506 km 2 of the state territory [5]. The major part of the province's territory consists of fertile plains with the Danube, Tisa and Sava rivers dividing it in to three regions: Bačka, Banat and Srem. The climate of Vojvodina is moderately continental, characterised by hot, dry summers, cold winters and relatively low rainfall.

Animals and sample collection
Blood samples were collected from 356 dairy cows from both commercial farms (109 cows) and smallholdings (247 cows) in Srem, Banat and Bačka region (Figure 1), during the 2009-2013 period. Samples from 271 cows were obtained on the farm by jugular venipuncture and 85 samples were collected at the abattoir. Among these were 74 samples from cows with a history of various reproductive disorders and 197 samples from reproductively healthy cows. Medical history could not be obtained for 85 abattoir samples. Prior to sampling, minimum recommended size of the sample was calculated, using Win Episcope 2.0 software [16]. Announcement of the Statistical Office of the Republic of Serbia on the number of cattle [14] and expected prevalence of 17.3% [13] that would ensure a 95% confidence interval and produce an error of 5% were used as input data for this calculation. Minimum recommended size of 220 animals was obtained. Therefore, a sample of 356 animals would not only provide an unbiased estimation of the prevalence of N. caninum-specific antibodies in the dairy cows of Vojvodina but would also reduce the error to 3.92%.
Blood samples were obtained from 99 dogs during the 2008-2013 period, from various sources on several locations in Srem, Banat and Bačka ( Figure 2). None of the sampled dogs, including six farm dogs, belonged to the same farms as cows from this study. All canine blood samples were collected from cephalic vein. The dogs were clinically examined prior to sampling and for each dog a record sheet, which included information on breed, gender, age, previous health problem and diet, was filled. For stray dogs, only gender and age (estimated by competent veterinarian) could be obtained.
Sample preparation procedure was identical for both cows and dogs: after being left to clot, blood was centrifuged (3000 rpm for 10 min) and separated sera were stored at À20°C until examination.

Serological testing
Two types of serological tests were used to detect N. caninum antibodies in cow sera. Initially 100 sera were assayed using the commercial competitive ELISA test kit (cELISA, VMRD Inc., Pullman, USA). The remainder 256 cow and all dog sera were examined with an indirect fluorescent antibody test (IFAT) using reagents marketed by VMRD. Both tests were performed according to the manufacturer's instructions. Cow sera, examined with ELISA, that presented inhibition percentages equal to or higher than 30% were considered positive. For IFAT, a recommended cut-off of 1:200 for cows and 1:50 for dogs was used. Dog sera that exhibited positive reaction at 1:50 were serially diluted until negativity was reached.

Statistical analysis
Seroprevalences and their confidence intervals, for both cows and dogs, were calculated using Quantitative Parasitology 3.0. [11]. For the statistical analysis of the possible effects of different factors (origin, farm type and history of reproductive disorders in cows and utilisation, breed, gender, age, origin and feeding habits in dogs) on the occurrence of anti-N. caninum antibodies the chi-squared test was used at a significance level of 95% (p 0.05). Where appropriate chi-squared test was replaced by Fisher's exact test or unconditional exact test. All these tests were computed using the above mentioned statistical software. Odds ratio (OR) was calculated using Win Episcope version 2.0 software [16]. Positive OR (OR > 1), with 95% confidence interval (CI) that does not overlap with the null value of 1 and with p 0.05, was considered statistically significant [15]. Confidence interval and p-value for OR were calculated using online MedCalc software version 12.6 [8].
Most (92 of 99) dogs were clinically normal and seven dogs had clinical signs which could not be related to neosporosis.

Discussion
The results of this study corroborate previous findings [3,9,13,17] of N. caninum antibodies in both cattle and dogs from the territory of Vojvodina.
The prevalence of antibodies in dairy cattle from our study (15.4%) was higher than the one found by Gavrilović et al. [3]. Similar to our study, their sample consisted of both aborting and randomly sampled cows, from both commercial and smallholding farms. Nevertheless, they found only 4.6% (23/500) of cows to be seropositive, which could be partly due to the small proportion of aborting cows in the total sample and the fact that they restricted their research to the south district of Banat. Different commercial tests were utilized in these studies (Gavrilović et al. used ELISA manufactured by IDEXX and Svanova, while we used IFAT and ELISA manufactured by VMRD) which may have also influenced the results.
Vidić et al. [17] and Savović et al. [13] reported occurrence of N. caninum antibodies in 3.7% (5/132) and 17.3% (9/52) of dairy cows, respectively. Although small size of the study samples and the fact they consisted entirely of cows with reproductive disorders make these prevalences difficult to compare with, we referred to them in the interest of data transparency.
With regard to the risk factors, significant difference was observed between seroprevalences in cattle from commercial and from small farms in our study. Probable reason for these observations is the fact that biosecurity measures on commercial farms are generally superior to those on small family farms which are more likely to have dogs, the potential source of infection. Furthermore, cattle from smallholdings are usually associated with extensive production and often allowed to graze, which is, according to some authors [4,10], one of the risk factors associated with seropositivity in cattle. Gavrilović et al. [3] also observed that all cows from seropositive commercial farms in their study had access to the outdoor paddocks which resulted in higher herd prevalence in commercial farms (50.0%), in comparison to smallholdings (22.64%). On individual level, however, they found no significant difference between commercial (4.67%) and small (4.50%) farms.
Prevalence of N. caninum antibodies in dogs varies based on many factors, including age, origin and diet [1]. Pavičić et al. [9] were first to detect N. caninum antibodies in four out of 31 (12.9%) dogs from Vojvodina. Examined dogs were mostly hounds, originating from Srem and Banat region. In the present study, we found antibodies in 17 out of 99 dogs (17.2%). Samples were collected from all three regions of Vojvodina and higher prevalence was observed in dogs from Bačka (26.1%) in comparison to dogs from Banat (7%). Other factors, such as breed, age, diet or utilisation did not appear significant. However, since the sample size was relatively small and samples originated from various sources and were collected over several years, no definitive conclusion can be made. Nevertheless, we recorded findings for the benefit of future studies. Presently, there are no reports of confirmed clinical neosporosis in any host in Serbia, including cattle and dogs.
In conclusion, our study confirmed the presence of N. caninum antibodies in general population of dairy cows in Vojvodina, with a higher prevalence detected in animals originating from smallholding farms. Specific antibodies against N. caninum were also found in dogs from the same province, with slightly higher prevalence observed in dogs from Bačka.