Risk factor analysis associated with Neospora caninum in dairy cattle in Western Brazilian Amazon

rabelo.vet@hotmail.com Abstract Neospora caninum is considered to be one of the main causes of abortion among cattle. The present survey was conducted in the municipality of Rolim de Moura, Rondônia State, Brazil. A questionnaire that investigates the epidemiological aspects of neosporosis was used in the analysis of risk factors associated with the animal-level and herd-level prevalence in dairy cattle. A total of 416 bovine blood samples were collected from 30 farms, and N. caninum antibody levels were measured by Indirect Fluorescent Antibody Test (IFAT). Analysis of dairy cattle serum samples revealed the presence of anti- N. caninum antibodies to be 47.36% (n = 197). Risk factors associated with N. caninum infection were the management system and access locations of dogs. The results of the present survey indicated that infection of dairy cattle with N. caninum is widespread in the studied region of Western Amazon, which has implications for prevention and control of neosporosis in this region. Therefore, integrated control strategies and measures are recommended to prevent and control N. caninum infection in dairy cattle. In addition, direct contact between dairy cattle, dogs and wild animals, which can influence the epidemiology of neosporosis, should be investigated


Introduction
Neosporosis is a parasitic disease caused by Neospora caninum, a protozoan misdiagnosed because of their morphological similarities (Dubey et al., 1988). Neospora caninum is considered to be one of the main causes of abortion among cattle (Almería et al., 2017;Dubey & Schares, 2011;Maldonado Rivera et al., 2020;Reichel et al., 2013). Although this parasite has not been demonstrated as a zoonosis, antibodies to N. caninum have been reported in humans (Duarte et al., 2020;Lobato et al., 2006).
In pregnant cows, infection with N. caninum can lead to several outcomes, including early foetal death and reabsorption; abortion, stillbirth or parturition of a deformed calf; and birth of clinically normal but infected offspring (Dubey et al., 2006;Favero et al., 2017). Neosporosis is of great economic importance, resulting in huge economic losses for farmers.
The State of Rondônia is located in the Western Amazon and has the second largest cattle herd stock in the North Region, which accounts for the third largest herd in Brazil, with 32 million animals (IBGE, 2016). In Brazil, the presence of N. caninum antibodies in cattle has been studied in several areas (Cerqueira-Cézar et al., 2017;Gennari, 2004), but only five of these reports was related to the Amazon region (Aguiar et al., 2006;Benetti et al., 2009;Boas et al., 2015;Silva et al., 2017;Minervino et al., 2008). Of the studies conducted in cattle in the Amazon region, only two were conducted in Rondônia (Aguiar et al., 2006;Boas et al., 2015).
The study of Aguiar et al. (2006) took place in the rural area of Monte Negro, Rondônia, and a N. caninum prevalence of 10.4% and 12.6% was detected in cattle and farm dogs, respectively. Boas et al. (2015) confirmed the presence of anti-N. caninum antibodies (10.6%) in cattle from Rondônia State, suggesting the necessity to further investigate the epidemiology of N. caninum in the Amazon region. Cañón-Franco et al. (2003) reported a seroprevalence of 8.3% for N. caninum in the urban dog population of Monte Negro municipality, state of Rondônia, suggesting the presence of the agent infecting cattle and other species in the region.
The possibility of transmission of N. caninum among wild and domestic animals has been widely discussed. The existence of a so-called wild cycle of neosporosis that may occur between canids and wild herbivores could influence the epidemiology of the disease in domestic livestock (Gondim et al., 2004;Rosypal & Lindsay, 2005).
In the Legal Amazon, the raising of livestock is the activity most strongly correlated with deforestation. Currently, livestock is present both at the consolidated border and in the expansion zones of forest occupation, popularly known as capoeira areas (Rivero et al., 2009). In Rondônia, this intimate contact between dairy cattle, dogs and wild animals could influence the epidemiology of neosporosis.
In this context, the present study aimed to evaluate the current status of the prevalence of N. caninum in dairy cattle from the state of Rondônia, Brazil and correlate possible variables associated with seropositivity for these protozoan.

Ethics statement
All procedures using animals complied with the Ethical Principles in Animal Research adopted by the College of Animal Experimentation and were approved (protocol number PP010/2014) by the Ethical Committee for Animal Welfare, UNIR, Rolim de Moura, Rondônia, Brazil.

Study region
The survey was conducted in Rolim de Moura (11º48'13"S, 61º48'12"W), which extends over 402 km 2 in Rondônia State, North Brazil (IBGE, 2016). The properties evaluated were classified according to the size: into small (<10 hectares), medium (11-100 hectares) and large (> 100 hectares). The climate is Aw of the Köppen classification (Peel et al., 2007), which is characterized as equatorial with changes to the hot and humid tropical with a well-defined dry season from March to September, minimum temperature of 24 ºC and a maximum of 32 ºC, with precipitation between high and moderately high (2000 to 2250 mm) and 85% relative air humidity.

Epidemiological survey
A comprehensive questionnaire investigating the epidemiological aspects of neosporosis was used in the analysis of risk factors associated with the animal-level and herd-level prevalence. The analysed variables and respective categories were as follows: sex, age, animal category, management system (intensive, semi-intensive and extensive), herd size (small: < 100 animals, medium: 100-150 animals and large: > 150 animals), technical monitoring (yes or no), stock up facility for food (yes or no), surface type of milking shed (wooden clapboard or cemented), farm hygienic status (not clean or clean), water source (dam or drinking fountain), feeder (wooden or cemented), presence of wild animals (yes or no), presence of dogs (yes or no), access locations of dogs (pasture or barn), abortion in the last 12 months and stage of gestation that abortions occurred (first, second or third trimester). The main classification criteria for the management system was animal stocking rate (SR: animal unit ha -1 , in which an animal unit = 450 kg of body weight): extensive: with pasture support capacity of up to 1.9 AU/ha; semi-intensive: with support capacity between 2 and 3.5 AU/ha and intensive: with support capacity between 3.6 and 7 AU/ha. Furthermore, in the extensive system the animals to lived predominantly pasture feeding management and in the semi-intensive, food management based on grazing, salt mineral and feed supplementation. Intensive system was characterized by a significant investment in buildings, machinery (such as milking parlour with automatic milking facilities) and private pastures (which are cultured and fenced).

Serum samples
The number of samples was calculated assuming that the prevalence of N. caninum is approximately 50% to maximize the sample size, obtain a minimal confidence interval (CI) of 95%, and maintain the statistical error The georeferencing (Figure 1) of the visited properties was carried out with a GPS device. The data were then transported to the QGIS Desktop (3.12.3) program (QGIS Development Team, 2020), where they were placed on the Rolim de Moura/RO map. The database modelling and chart plotting stages were carried out at the Institute of Agrarian Sciences, Federal University of the Jequitinhonha and Mucuri Valleys. under 1%. Calculations were executed using an EpiInfo program (CDC, version 7.2.0.1), resulting in a sample size of 416 bovines collected from the 30 farms (Thrusfield, 2005). Cattle from each farm were selected randomly using a table of random digits. In the case of cattle, approximately ten percent of animals from each farm were sampled. The sampled animals were crossbred cows, usually derived from the crossing of animals of a purebred European origin (Holstein) with animals of one of the zebu breeds (Gyr) in various degrees of blood. All the animals sampled were clinically healthy. A total of 416 crossbred Holstein cattle blood samples (10 mL) were collected by venocentesis, using identified vacuum tubes, and centrifuged at 1000g for 10 min. The serum was separated and stored at −20 °C until analysis.

Serological tests
Neospora caninum antibody levels were measured by Indirect Fluorescent Antibody Test (IFAT). using rabbit antibovine IgG conjugate labelled with fluorescein isothiocyanate (Sigma, St. Louis, MO, USA) according to Dubey et al. (1988), with a cut-off value of 1:200 (Benetti et al., 2009;Gennari, 2004). Neospora caninum tachyzoites (NC-1) maintained in Vero cell cultures were used as antigens. Positive and negative control sera were added to each slide, and all the positive samples were retested using a two-fold serial dilution.

Determination of prevalence and statistical analyses
Statistical tests were performed using STATA, version 14.1, software (Stata Corp LP, College Station, Texas, USA). The research variables were presented as percentages and 95% CIs for proportion. For inferential statistics, the seroprevalences of N. caninum infection were considered as the dependent variables and other factors were considered as the explanatory variables.
Pearson's Chi square (χ2) or Fisher exact tests was executed to evaluate the differences between groups. To investigate the independent risk factors of each explanatory variable, all variables that showed a p value ≤ 0.25 in the univariate analysis were offered to the multivariate logistic regression model as suggested by (Bendel & Afifi, 1977;Bursac et al., 2008). It is advised to use an initial screening p value cut-off point of 0.25, as more traditional levels such as 0.05 can fail to recognize variables known to be important. The occurrence probability ratio [odds ratio (OR)] and the corresponding 95% CI were calculated by using univariate and multiple logistic regression. A p value < 0.05 was considered as the level of statistical significance for all tests.

Results
Analysis of crossbred Holstein cattle serum samples revealed the prevalence percentages of anti-N. caninum antibodies to be 47.36%. Herd-level seroprevalence was of the order of 93.33%. Seroprevalence results of N. caninum are summarized by herds in Table 1. There was no statistically significant association between groups and subgroups of the animal categories (p > 0.05) All the evaluated herds showed the presence of wild animals, domestic dogs and abortion in the last 12 months; therefore, these explanatory variables were not analysed statistically.
Regarding the sex, age, stock up facility for food, water source and stage of gestation that abortions occurred, there were no significant differences (p > 0.05) in the seropositivity for neosporosis and no were included in the logistic regression model ( Table 3).
The data from the univariate analysis is summarized in Table 3, and we can see that the statistically significant (p < 0.05) explanatory variables were the management system (OR: 1.68, p = 0.01), farm hygienic status (OR: 1.84, p = 0.01) and access locations of dogs (OR: 1.92, p < 0.01).
The multivariable logistic regression model (Table 3) of the predictors of N. caninum seroprevalence was performed in all variables (management system, herd size, technical monitoring, surface type of milking shed, farm hygienic status, feeder and access locations of dogs) with p value ≤ 0.25 in the univariate analysis.

Discussion
In the present study, the animal-level and herd-level seroprevalence of N. caninum in dairy cattle was 47.36% and 93.33%, respectively. In Brazil, the seroprevalence of anti-N. caninum antibodies in cattle varies from 6.7% to 97.2% (Cerqueira-Cézar et al., 2017). Regarding seroepidemiological studies in dairy cattle carried out in the Amazon region, the prevalence found in this research was higher than the results found by Minervino et al. (2008) in Pará (17.5%; cut-off 1:50); Aguiar et al. (2006) in Rondônia (11.2%; cut-off 1:25); Boas et al. (2015) in Rondônia (10.2%; cut-off 1:100); and similar to those described by Benetti et al. (2009) in Mato Grosso (53.5%; cut-off 1:200). However, these values are not entirely comparable since different cut-offs were applied and IFAT reading is partially subjective (Álvarez-García et al., 2003;Campero et al., 2018;Von Blumröder et al., 2004;Wapenaar et al., 2007) We observed that N. caninum is widely spread in the studied region, with only two farms obtaining negative results; however, we could not exclude the presence of this protozoan in these properties due to the number of samples collected.
The high infection at the herd-level and the similarity with the seroprevalence reported by Cerqueira-Cézar et al. (2017) could indicate that N. caninum is endemic in the country. This prevalence may be explained by the hot and humid climate of the Amazon region and high levels of precipitation that average 2000 mm annually, which may contribute to higher viability among oocysts in the environment (Dubey et al., 2007;Moloney et al., 2017).
Milk production in Rondônia (8th place in the ranking of the largest national producers) is an economic activity composed of approximately 32,000 families, the majority of which are small producers (IBGE, 2019). Dairy farming has great social relevance, being one of the rural activities that employs most of the labour among the productive chains in family farming.
Higher N. caninum prevalence was detected in bulls in relation to cows. This difference could be related with the constant renewal of the cow herds, while the bulls remained in the herd for a longer period of time than the females. These results are in agreement with those reported by Guimarães et al. (2004).
In this study, we found a prevalence of 39.06% for calves, which can be explained by vertical and horizontal transmissions. For heifers, the prevalence was 53.06%, which could have occurred due to the maintenance of horizontal infection.
Vertical transmission is considered effective in bovines, since it is responsible for maintaining the infection in the herd (Dubey et al., 2006). However, the infection could not be sustained in the herd only with vertical transmission,  since its efficiency is not 100%. Therefore, horizontal transmission also plays a key role in perpetuating the disease in the herds. Without it, the infection levels would tend to reduce over time (French et al., 1999).

Non-Reagent n (%) OR a (95% CI † ) P-values AOR b (95% CI † ) P-values
All the herds evaluated showed the presence of wild animals and domestic dogs, and these explanatory variables were not analysed statistically. In agreement with several researchers (Cerqueira-Cézar et al., 2017;Corbellini et al., 2006;Hobson et al., 2005;Macchi et al., 2020;Otranto et al., 2003;Ribeiro et al., 2019;Schares et al., 2003Schares et al., , 2004a, an association with the number of dogs, there was no report of differences between the presence and absence of this species. The prevalence increment association with rise in herd size could be linked to compromised hygienic measures in maintaining adequate waste disposal practices, increased likelihood of more dogs in the vicinity of the farm and/ or a tendency towards raising replacement heifers from their own infected stock (Dubey et al., 2007).
The differences in prevalence could be influenced by many factors, such as the method and cut-off used; the number of cows examined; and their health status, age, gender and breed (Bártová et al., 2015). However, it can be influenced by the type of trough, barn floor, cleaning of facilities, and the presence of dogs on agricultural property (Cerqueira-Cézar et al., 2017).
The Western Amazon of Brazil has prevalence values within the range found in other important regions of cattle production in Brazil. In the present study, the management system (OR 1.68, 95% CI: 1.11-2.54), farm hygienic status (OR: 1.84, 95% CI: 1.13-3.00) and access locations of dogs (OR 1.92, 95% CI: 1.25-2.94) were identified as risk factors associated (p < 0.05) with the prevalence of N. caninum infection in the univariable analysis.
In our study, the hygienic status of the farms was assessed according to floor condition, ventilation, drainage, and waste disposal routines. All these hygienic factors may contribute to maintaining the N. caninum cycle by providing canids access to tissue cysts through feeding with infected meat or organs and providing cattle access to oocysts through contamination of feed and water with dog faeces (Dubey et al., 2007;Schares et al., 2004b).
Applying multivariate logistic regression analysis ( Table 3), tests of the association between the seroprevalence of N. caninum infection and potential predictors revealed a strong association with the management system (OR 2.05, 95% CI: 1.19-3.53) and access locations of dogs (OR 2.24, 95% CI: 1.43-3.49).
Farms with an extensive management system (Table 3) were more than 2.05 (95% CI: 1.19-3.53) at risk of acquiring the infection compared to the farms with a semi-intensive management system. This can be explained by the greater chances of contact with oocysts in the environment when the extensive management system is used. Animals with access to pasture have greater opportunity to infect themselves horizontally, when compared to those raised in the semi-intensive system, that the animals remain part of the day free in pastures and part of the day confined. Thus, an effective management mode is of extreme importance to control the spread of disease. These results are similar to those cited by Macchi et al. (2020), Ragozo et al. (2003), Ribeiro et al. (2019), Silva et al. (2008).
The access of dogs to the pasture presented a significant risk factor (OR: 2.24, 95% CI: 1.43-3.49), influencing the high seroprevalence. This result agree with Nematollahi et al. (2011) who found that the presence of domestic dogs or dingoes can directly reflect the high rates of neosporosis in these locations. Aguiar et al. (2006) showed contact with forest areas and the presence of dogs were not associated with the coccidian infection.
The presence of dogs on cattle farms has been considered an important risk for cattle infection and abortion by N. caninum in different countries Corbellini et al., 2006;Mainar-Jaime et al., 1999;Martins et al., 2012;Paré et al., 1998;Wouda et al., 1999). In addition, Barling et al. (2000) found a significant association between N. caninum infection in cattle and the presence of coyotes and grey foxes (Urocyon cinereoargenteus) in farms in the Unites States.
Thus, observation of a higher seroprevalence associated with the access of wild carnivores to cattle ranches in Rondônia suggests that they may be involved in maintaining a wild cycle in this scenario. The apparently higher risk of exposure in the herds with wild carnivore access corroborates reports from Brazil (Martins et al., 2012).
A possible way of infection could be the consumption of water contaminated with N. caninum oocysts from the faeces of infected wild or domestic canids, because they may come to the water source region to drink and at the same time defecate in or near it (Sun et al., 2015). Thus, a strict detection system for the drinking water of cattle should be implemented in the study regions.
It is concluded that neosporosis can be one of the possible causes of abortion in dairy cattle in Rolim de Moura, Rondônia. Regarding the distribution in dogs as definitive hosts for the parasite, further studies in dog and cattle are recommended. Neosporosis is of great economic importance, resulting in huge economic losses for farmers.

Conclusion
The results of the present survey indicate that infection of dairy cattle with N. caninum is widespread in Rolim de Moura, Western Amazon, which has implications for the prevention and control of neosporosis in this region. Therefore, integrated control strategies and measures are recommended to prevent and control N. caninum infections in dairy cattle. In Rondônia, this intimate contact between dairy cattle, dogs and wild animals could influence the epidemiology of neosporosis.