Analysis of risk factors associated with seropositivity to Neospora caninum in dogs
L. Kramer,L. D. Risio,V. Tranquillo,S. Magnino,Claudio Genchi
DOI: https://doi.org/10.1136/vr.154.22.692
2004-05-29
Abstract:Neospora caninum is an apicomplexan protozoan that has been recognised as a leading cause of abortion in cows and which also causes neuromuscular disease in the dog, a definitive host (McAllister and others 1998, Dubey 1999). Epidemiological studies carried out on dogs have shown that the parasite is cosmopolitan (Trees and Williams 2000) and transmission is both vertical, with transplacental transfer of tachyzoites to the fetus, and horizontal, through the ingestion of cysts contained in intermediate host tissue (Dijkstra and others 2001). Using a non-probabilistic sampling procedure of convenience, blood samples were taken from 282 dogs which visited the Veterinary Teaching Hospital of the University of Parma Veterinary School during 2000 to 2001. The dogs' cinical status was not considered in the sampling. Owners were asked to complete a questionnaire regarding their dog's sex, age, breed, habitat (urban/rural), aptitude (companion, hunting or guard dog), access to cattle farms and feeding with raw bovine meat. Blood samples were centrifuged and serum was frozen at -20°C until processing. N caninum serology was carried out by the indirect fluorescent antibody test, using the NC-1 strain of the protozoan (Magnino and others 2000). A titre of at least 1:50 was considered positive (Bjorkman and Uggla 1999). Statistical analysis included univariate analysis of single risk factors associated with seropositivity and multivariate analysis of significant variables. All independent variables were analysed separately by 2 x 2 contingency tables and expressed as an odds ratio (OR) (how many times the risk of N caninum seropositivity is higher or lower in the presence of a given variable). The OR can be 1 (no association), more than 1 in the case of a positive association (that is, risk factor) or less than 1 in the case of a negative association (protective factor). Statistical significance for OR values was calculated using a chi-squared test with 95 per cent confidence intervals (cIs). For multivariate analysis, all variables with P<0IO in the univariate analysis were included in a model of logistic regression. A backward building model was followed based on the likelihood-ratio test as described by Hosmer and Lemeshow (1989). All analyses were carried out in STATA 7.0 software (StataCorp 2000). Fifty-one of the 282 dogs were seropositive to N caninum (18-1 per cent; 95 per cent CI: 13-5 to 22-6). Table 1 shows the results of the multivariate logistic regression. Values were significantly higher in dogs over three years of age (OR 2.80) and in dogs that had had contact with cattle (OR 2-96). In addition, dogs that were fed raw bovine meat (OR 2.97) were at greater risk of infection. There was no significant interaction among the single factors. In this cross-sectional study, dogs were chosen at random from a hospital population and, as such, are likely to have been influenced by some sampling bias. However, prevalence values for serum antibodies to N caninum were within the values reported elsewhere in Europe (Trees and Williams 2000). This study confirmed several factors previously observed by others as being significantly associated with N caninum seropositivity: older dogs are at greater risk, suggesting postnatal infection, and dogs from rural settings and dogs which have had contact with cattle tend to show higher prevalence values. The feeding of raw meat (specified in the present study as raw bovine meat) has been mentioned as a possible risk factor (Basso and others 2001, LeJeune and Hancock 2001), but to the authors' knowledge this is the first epidemiological study to include this factor in a statistical analysis. Some of the literature (Nurminen 1995, Frankena and Thrusfield 1997, Thrusfield and others 2001) mentions the possibility of using prevalence ratios instead of the classical approach of using odds ratios, as used in this study (Table 1). The use of prevalence ratios yielded virtually the same conclusions as those shown in Table 1. Peters and others (2001) reported the presence of N caninum tissue cysts in the skeletal muscle of infected cattle, and high seroprevalence values have been reported in a group of foxhounds which were fed cattle carcases (Trees and Williams 2000). Raw bovine meat should therefore be considered a possible transmission route for N caninum in dogs and feeding such meat should be avoided in the control and prevention of disease in this species.