Short Communication Seroprevalence and risk factors associated with canine visceral leishmaniasis in the State of Paraíba, Brazil

Introduction: The aim of this study was to determine the seroprevalence and risk factors associated with canine visceral leishmaniasis in Paraíba State. Methods: Blood samples were collected from 411 dogs in four municipalities of Paraíba State. The seroprevalence was assessed by using ELISA. Results: The seroprevalence ranged from 7.2% to 20%. The risk factors that were associated with the disease were the presence of ticks and contact with other animals. Conclusions: The seroprevalence of canine visceral leishmaniasis is high in Paraíba, indicating the need for health authorities to resume visceral leishmaniasis control activities. The ruralization of suburban areas in the region tends to homogenize the risk factors between suburban and rural areas.

Visceral leishmaniasis (VL) is a zoonotic disease that is endemic in many tropical and subtropical countries.In Brazil, the disease is caused by Leishmania infantum (syn.Leishmania chagasi).The main vector is the sandfly Lutzomyia longipalpis, and the dog is the main domestic reservoir 1 .
According to the Brazilian Information System for Notifiable Diseases [Sistema de Informação de Agravos de Notificação (SINAN)], 433 cases of human visceral leishmaniasis (HVL) were reported in the Paraíba State from 2001 to 2013.Several factors, which are still controversial, have been associated with CVL; these are the breed, fur length, age, the presence of other animals in the household or peridomestic environments, resting place, and the function of the dog [2][3][4][5][6] .
Little is known about the actual prevalence of canine visceral leishmaniasis (CVL) in Paraíba counties and the associated factors with CVL.The perception that there has been an increase in the prevalence of CVL is based on the following observations: a) there is an increased frequency of symptomatic dogs in veterinary clinics; b) there is an increased serological diagnosis of infection in animals referred for vaccination against leishmaniasis, and c) there is an increased frequency of stray dogs with suggestive lesions.Thus, the aim of this study was to determine the seroprevalence and the risk factors associated with CVL in pet dogs in counties of the Paraíba State.
The dogs were from the counties of Patos, Sousa, Cajazeiras, and Uiraúna, all of which are located in the semi-arid mesoregion of the Paraíba State in northeastern Brazil, which is endemic area for CVL (Figure 1).The climate of this region is semiarid, hot and dry.The region is subject to prolonged droughts, with an annual rainfall below 500mm and an annual average temperature exceeding 26°C 7 .
The sample size was calculated assuming a simple random sample would be obtained, with an expected prevalence of 50%, a minimum confidence of 95%, and a statistical error of 5% 8 .These parameters provided a sample size of 385 animals; however, blood samples were collected from 411 dogs in four counties.Dogs older than six months that were residing in the urban, peri-urban, and rural areas were included in this study.In the urban and peri-urban areas, sample collection was performed at collection points, one in each neighborhood, after an extensive campaign to increase the knowledge of the population through a local radio.In the rural area, sample collection was carried out at sites pre-selected by lottery; the sampling frame for the lottery comprised all houses and all dogs in the selected localities.Sampling was performed in 2011.A blood sample was obtained through a cephalic venipuncture using a 5-mL syringe and a 25 × 8mm needle.A total of 5mL of blood was immediately placed in a vacuum tube containing 4% sodium citrate as an anticoagulant.The samples were processed at the Blood samples were centrifuged at 2,000g for 10 minutes, and plasma was transferred to 1.5mL microcentrifuge tube and stored at -20°C until analysis.
The dogs' tutors answered questions in an epidemiological questionnaire that was developed with the objective of identifying whether standard management practices were used by the tutors and identifying other conditions that could be associated with the development of canine visceral leishmaniasis.The variables considered in the questionnaire were family income and education level; presence of other species of animals (domestic or wild) either in or around the house; environmental conditions of the rearing place; dog's characteristics (sex, age, breed, type of rearing, vaccination history, deworming history, and infestation with ticks).
An enzyme-linked immunosorbent assay (ELISA) kit (S7 ® , Biogene Indústria e Comércio Ltda., Recife-PE, Brazil) for CVL, registered at the Brazilian Ministry of Agriculture, Livestock and Supply, was used to diagnose the infection.The kit contains a recombinant antigen composed of the L. infatum heat shock protein (HSP70) carboxy-terminal moiety (the kit has a sensitivity of 96% and specificity of 90%).The protocol was performed according to the manufacturer´s instructions.
Statistical calculations were performed using Statistical Package for the Social Sciences (SPSS) 20.0 for Windows software at a 5% significance level.To identify the factors associated with seropositivity, univariate analysis was performed using either the Pearson´s chi-square or Fisher's exact test.Each independent variable was compared with the dependent variable (serological result), and independent variables that had a p ≤ 0.20 9 were selected for multivariate logistic regression analysis 10 .A descriptive statistical analysis of the relative and absolute frequencies of ELISA (sero) positive dogs was performed.
Epidemiological information was obtained using questionnaires.Mongrel dogs comprised 76.2% of the canine population, and 58.6% of the dogs were males.Dogs aged between 2 and 4 years had the highest (18.1%) prevalence of CVL, but there was no significant statistical difference with the other age groups.Most (70%) dogs were kept within the house.There was an increase in the seroprevalence that was inversely related to confinement: the prevalence in dogs strictly kept within houses (resident) was 11.8%; in semi-resident dogs, 15.1%; and in free-roaming dogs, 16.2%.However, there was no significant statistical difference between these categories.The tutor´s family income and educational level were not associated with CVL.
The variables that were significant in the univariate analysis were contact with wild animals, contact with other domestic species, kennel cleaning frequency, and tick infestation (Table 1).The results of the multiple logistic regression analysis confirmed tick infestation [odds ratio (OR) = 3.89] and contact with other domestic species (OR = 3.44) as factors associated with CVL (Table 2).Contact with other animals in the nearby environment was reported by 79.6% (327) of tutors, and 31 dogs had contact with other animal species; the risk of infection of CVL was increased by 3.44 times in such dogs.
The seroprevalences found in this study are similar to those previously reported for endemic counties and regions 11,12 .The prevalence varies greatly between different locations in the same county; generally, the prevalence is higher in rural than in urban areas 2,11 .As noted by Belo et al. 5 and Azevedo et al. 12 , prevalence estimates can be influenced by many factors, such as the study area, the diagnostic method, and the sampling method.The prevalence estimate can be safely used to assess the risk of CVL in the region and may help public health authorities to plan for control programs against human visceral leishmaniasis.
There is no consensus among experts with regard to the factors associated with infection (as it with regard the prevalence); there factors vary between the Brazilian regions and between countries.All animals were domiciled, and both female and male dogs had similar roaming patterns.There was no significant statistical difference between genders with respect to the prevalence.Moreover, Curi et al. 6 , in a study carried out in preserved areas of the Atlantic Forest, have suggested that sedentary animals and those that remain in restricted spaces have a greater risk of infection because they are easier targets for the sandfly.Based on this suggestion, we speculate that the lack of correlation between the prevalence and the mobility of the animals (although all animals were domiciled) points to the role of other risk elements in the epidemiological chain.
According to Cesse et al. 13 , the ruralization of the urban outskirts in semi-arid landscapes causes a significant increase in the number of potential breeding sites for Lutzomyia longipalpis; this process (ruralization) has been observed repeatedly in the counties studied in Paraiba and may constitute the most relevant risk factor for both CVL and human visceral leishmaniasis.
Both low family income and low education level are considered to be risk factors for human visceral leishmaniasis 5,13 .In relation to CVL, Coura-Vital et al. 4 also pointed out that low income, as well as limited tutor´s knowledge about the vector, is a risk factor for CVL.The difference between the results cited above and those obtained in our study may arise from the difference in the distribution of social strata in large cities compared to that in the Paraíba counties; further work is required to confirm this hypothesis The seropositivity index was 3.89 times higher in dogs infested with R. sanguineus.There have been reports of infection of ticks with L. infantum and L. donovani, but there is still controversy whether these ticks transmit the disease.Coutinho et al. 14 demonstrated that it is possible to infect hamsters orally by feeding them with infected macerated R. sanguineus, and Dantas-Torres et al. 15 demonstrated experimental transovarian transmission of L. infantum in infected R. sanguineus.However, the latter authors were unable to show any association between seropositivity and dog infestation with Amblyomma ovale and R. sanguineus.Taking into account that the environment around the house is an important factor and that the presence of unpaved yards and livestock offer breeding sites and blood meals, respectively, for both ticks and sand flies, we advance the hypothesis that ticks are not directly involved, but they serve as an indicator of the environmental conditions that favor the presence of sand flies and of breeding sites.
As described by Azevedo et al. 12 , other animals (hens, swine, and horses) serve as an additional food source to phlebotomines; hence, the former can facilitate maintenance of the latter and allow development of high vector densities around houses.Moreira Jr. et al. 3 have also reported that other animals, especially pigs, but not chickens, significantly increase the risk of CVL.In our study, pig rearing was not associated with CVL, but other livestock may be involved, as suggested by Cesse et al. 13 .Although other vertebrates can serve as food for the sandfly and favor its maintenance in areas around houses, the former can also reduce the number of infectious bites for the dogs.
The current seroprevalence of canine visceral leishmaniasis in semi-arid Paraíba hinterland is high, suggesting the need for more effective control measures.
Although the risk factors for CVL in Brazil are controversial, our study suggested that the environment around the house is an important element with respect to the risk of disease.
The ruralization of suburban areas in the region tends to homogenize the risk factors between suburban and rural areas.

Ethical considerations
The project was approved by the Ethics Committee on Animal Use of the Universidade Federal de Campina Grande (Protocol CEP nº 59-2011).

FIGURE 1 :
FIGURE 1: Map of the State of Paraíba, Brazil.The studied area comprised the counties of Patos, Sousa, Uiraúna, and Cajazeiras.

TABLE 1 :
Univariate analysis of possible risk factors of canine visceral leishmaniasis in the studied area.
MW: minimum wages.*Variables used in the multiple logistic regression.
Silva RBS et al. -Canine visceral leishmaniasis epidemiology

TABLE 2 :
Risk factors of canine visceral leishmaniasis in the studied area, as determined using multiple regression analysis.