Risk factors associated with Leishmania exposure among dogs in a rural area of Ilha Solteira, SP, Brazil

Abstract INTRODUCTION: We sought to determine risk factors (RFs) associated with the presence of antibodies against Leishmania in dogs from a rural area of Ilha Solteira, SP, Brazil. METHODS: Serum samples were collected from 250 dogs and tested using indirect enzyme-linked immunosorbent assay (ELISA) and indirect immunofluorescence antibody tests (IFATs). Data concerning dogs, their environment, and their owners’ knowledge of leishmaniasis were collected using a questionnaire. To determine RFs for contact with the parasite, univariate statistical analysis based on chi-squared and Fisher’s exact tests, followed by logistic regression, was used. RESULTS: It was found that 79/250 (31.6%) of the dogs were positive by IFAT, and 72/250 (28.8%) by ELISA. A total of 82/250 dogs (32.8%) were positive in at least one test. The RFs associated with occurrences of Leishmania exposure were large body size (OR = 2.25; 95% CI = 1.26-4.04; p = 0.003), presence of chickens (OR = 1.94; 95% CI = 1.05-3.65; p = 0.023), and lack of knowledge about Leishmania among dog owners (OR = 1.74; 95% CI = 0.96-3.21; p = 0.049). After multivariate analysis, the RFs for occurrence of Leishmania exposure in dogs that remained significantly associated were the dog’s size (large dogs) (OR = 1.2; 95% CI = 1.06-1.35; p = 0.003) and presence of chickens on the properties (small farms) (OR = 1.15; 95% CI = 1.02-1.30; p = 0.023). CONCLUSIONS: These results may be useful for improving preventive practices to reduce the incidence of Leishmania exposure among dogs in rural areas.


INTRODUCTION
Leishmaniases are zoonoses caused by protozoa belonging to the genus Leishmania 1 . The species that causes visceral leishmaniasis (VL) in countries in the Americas is Leishmania infantum (syn. L. chagasi) 2 .
The main means of transmission of the parasite to dogs and other mammalian hosts is through the bite of females of hematophagous dipterans of the family Psychodidae belonging to the genera Phlebotomus and Lutzomyia, in the Old and New World, respectively, which are infected with promastigote forms of Leishmania spp. 3,4 . The species Lutzomyia (Lutzomyia) longipalpis is considered to be the main transmitter of the parasite in Brazil. This vector species feeds on a wide variety of vertebrate hosts, such as birds, wild and domestic mammals, and humans 5 .
Although several wild hosts have been identified in urban areas, Canis familiaris is the domestic host, and is considered to be the main reservoir of infection for humans 6 . Clinical manifestations of visceral canine leishmaniasis (CanL) are characterized by dermatological symptoms: flaking and excessive depigmentation, which normally develop on the head, and which differ to other parts of the body, with itching, dry skin, hair loss and areas of hyperkeratosis and lignification, and onychocryphosis. They may also present ulcers and small intradermal nodules, digestive symptoms (intestinal hemorrhage), respiratory symptoms (runny nose), eye symptoms (conjunctivitis, blepharitis, corneal enlargement and opacity) and general symptoms (apathy, anorexia, anemia, limb edema, hyporexia, weight loss and lymphadenomegaly) 7 .
Many risk factors (RFs) for the occurrence of VL have been listed, thus indicating possible interactions between the links that make up the epidemiological chain, such as vectors, hosts and the environment 8,9,10,11,12 . Thus, knowledge of the distribution of the disease in endemic areas and possible associations between the disease and RFs can help in developing control strategies 13 . In this context, domestic dogs play an important role in the maintenance and spread of the disease. For this reason, factors that may be associated with the risk that these animals may become infected need to be well known 14 .
Some studies conducted over the last decade have identified certain RFs that are associated with VL in urban regions. These include poor housing conditions, especially with a lack of household waste collection and an irregular or absent sewage system 15 ; increased population density of phlebotomine sand-flies 15,16 ; breeding of birds in cages in the presence of the vector 15 ; and presence of other animals in the peridomestic area, particularly opossums 17 , chickens and pigs 18 .
Recently, a cross-sectional study carried out in endemic areas of Cuiabá, state of Mato Grosso, showed a CanL seroprevalence of 22.1%. Animals living in rural settings had a 1.9-fold higher risk of been infected than those in an urban environment. Factors relating to the habits of these animals, such as free access to the external environment and a watchdog function, along with the presence of agricultural activity were probably indicators that predicted Leishmania spp. exposure 19 .
Paulan et al. (2012) used geoprocessing techniques in association with satellite imaging to reveal that the estimated prevalence of CanL in Ilha Solteira, state of São Paulo was low to medium-high, ranging from 10% to 14.5%, depending on the neighborhood studied. The areas with the highest density of CanL cases were close to natural vegetation fragments (at a zoo) and near rural settlements, i.e. farther from the city center. Spada et al. (2014) studied the prevalence of Lu. longipalpis and CanL in the "Cinturão Verde" (green belt) area. They visited 12 properties over a 12-month period and collected biological samples from 32 dogs. Once a month, insects were caught using CDC (Centers for Disease Control and Prevention) traps. It was found that the vector was present on 100% of the properties, and that 31.25% of the dogs were positive for CanL.
The "Cinturão Verde" has a considerable human and canine population, which presents suitable conditions for vector maintenance, and is located near the urban perimeter of the city; this area represents an RF for maintenance of local disease and spread of this zoonosis to the urban area, if preventive measures are not implemented.
Thus, the objective of this study was to determine the RFs associated with Leishmania exposure among dogs in the "Cinturão Verde" of Ilha Solteira, SP, Brazil.

Study area
This study was conducted in a rural area referred to as the "Cinturão Verde" (Green Belt), which belongs to the municipality of Ilha Solteira (51°06'35" W and 20°38'44" S). The Cinturão Verde occupies an area of 880.46 hectares (ha) and is divided into agricultural production areas (563.29 ha); reforestation areas (317.68 ha); talvegues (lines connecting the lowest points of a river bed) (45.65 ha); area used for hydroelectric construction (227.39 ha); and legally enriched reserves (area with native vegetation cover) (44.12 ha). The entire extent of the Cinturão Verde is surrounded by 77 areas of dry land (non-irrigated) and 14 areas of irrigated land that are distributed among approximately 200 families. These families carry out various functional activities, such as growing vegetables and raising small animals, such as poultry and pigs.

Ethics Committee
The present study was approved by the Ethics Committee for Animal Use (CEUA) of the School of Engineering School of Ilha Solteira (part of São Paulo State University, UNESP). It formed part of a research project entitled "Distribution of the Phlebotomine Entomophase (Diptera: Psychodidae) and Canine Visceral Leishmaniasis Area of the "Cinturão Verde" of Ilha Solteira, State of São Paulo". Approval was granted at an ordinary meeting of CEUA held on May 9, 2011, under protocol no. 002/2011/CEUA. Procedures were performed based on current standards for research involving animal use according to the National Council for Animal Experiment Control (CONCEA).

Study design and dog samples
A cross-sectional study on Leishmania exposure in dogs was conducted between February 2012 and February 2013. The sample size was established considering a population of 400 dogs (2 dogs/ family) in the study area. Thus, the size of the sample, based on an arbitrary random method and with finite population adjustments of less than 200 dogs and a sampling error of 5%, was estimated to be approximately 250 dogs 22 . To ensure representativeness of the sample size, it was defined that the methodology for collecting the material should not involve any concentration of samples in any single region of the total area, but rather that the collection of material should cover the entire perimeter of the area. With the aid of a map provided by city authorities and a number of local visits to the study area, land areas and ownership were determined. In total, 104 families were visited, and biological material was collected from all dogs belonging to each family, irrespective of the numbers of dogs.

Blood collection
Blood samples from the dogs were taken directly from the cephalic vein or the external jugular vein, using vacuum flasks without anticoagulant, to obtain serum samples. The whole blood was centrifuged at 900 × g for 10 minutes to separate the serum and was then kept at -20 °C until further use.

Clinical Characterization
At the time of blood collection, the animals were examined clinically by means of general physical examination, and classified according to the clinical signs evident for CanL with one or more clinical signs and without clinical signs. Among the findings on physical examination, cachexia, hyperthermia, hyporexia, dermatological changes such as alopecia, ulcerative skin lesions, flaking, crusts, lymphadenomegaly, periocular lesions, uveitis, conjunctivitis, pale mucous membranes, and onychogryphosis were noted.

Detection of anti-Leishmania antibodies
Anti-Leishmania antibodies were assayed using indirect immunofluorescence antibody tests (IFATs) and indirect enzymelinked immunosorbent assays (ELISAs) as described by Oliveira et al. (2008). Positive control serum was obtained from confirmed CanL cases that had been detected using direct methods. For negative controls, serum from healthy dogs was used.
To perform the ELISA test, the soluble antigen of L. infantum was used at a concentration of 5 µg/mL, diluted in 0.05 M sodium carbonate bicarbonate buffer, pH 9.6. An anti-dog conjugate, rabbit anti-dog IgG coupled to alkaline phosphatase (Sigma Chemical Co, San Luis, Missouri, EUA) was diluted 1: 4000 in phosphate buffered saline (PBS), 0.01 M, pH 7.2 with 0.05% Tween-20 (PBS-Tween). As a substrate, paranitrophenylphosphate diluted to 1 mg/mL in diethanolamine buffer, pH 9.8 was used. The plates were read in an ELISA reader (Dynex Technologies, Chantilly, Virgínia, USA) at 405 nm. The cut-off point for the ELISA test corresponded to two and a half times the average value of the mean optical density (OD) of the negative reference sera.
To perform IFATs, antigenic substrate was obtained from L. infantum promastigotes grown in RPMI -1640 medium, at 25°C. Serial dilutions of each serum were performed commencing with a 1:40 dilution. The conjugate was dog anti-IgG linked to fluorescein isothiocyanate (KPL, Milford, Massachusetts, USA) diluted according to the manufacturer's recommendations. Sera were considered as positive when the parasites exhibited fluorescent color throughout the periphery, with a cut-off point of ≥ 1:40.

Questionnaires
To determine RFs, standardized questionnaires were used. At the time of collection of blood from the dogs, the questionnaire was applied to the owners or caregivers of the animals. The information sought through the questionnaires included the identification and characteristics of the dogs, the environment in which the dogs lived, and the degree of knowledge regarding VL among the owners.

Definition
Positivity for Leishmania exposure among the dogs was defined as positive detection of antibodies by means of IFAT or ELISA.

Geographic location of the animals
The locations were georeferenced using the Global Positioning System (GPS). These data were imported into a geographic information system (GIS) using the QGIS version 2.18.10 software package (Free Software Foundation, Boston, Massachusetts, EUA) with Open Layers plugin, to visualize the spatial distribution of the data. Finally, dots representing data points were projected into an image layer obtained from the Google Earth database (Figure 1).

Statistical analysis
The association between potential RFs and Leishmania exposure in dogs was assessed by means of univariate analysis using the chisquared test, or Fisher's exact test, when required, and by means of multivariate logistic regression analysis. Odds ratios (OR) and 95% confidence intervals (CI) were calculated, and p values < 0.05 were considered statistically significant. All analyses were performed using R software (The R Foundation, Vienna, Austria), version 2.15.3 24 .

RESULTS
A total of 250 dogs were analyzed. Anti-Leishmania antibodies were found in 79/250 (31.6%) of the dogs using IFAT, and 72/250 (28.8%) using ELISA. In total, 82/250 (32.8%) of the dogs were positive. Regarding the information obtained from the questionnaire, seven variables relating to dog characteristics were analyzed: hometown,    Among the 250 dogs sampled in this study, there were 145 males and 105 females. There was greater occurrence of Leishmania exposure in the male population (54/82; 65.9%), but this difference was not significant (OR = 1.63; 95% CI = 0.94-2.82; p = 0.079). Regarding the age of the animals, 200 (80.0%) were adults (more than one year of age), while 50 (20.0%) were less than one year old. Although the positivity rate was higher among adult animals (69/82; 84.2%), i.e. those over one year of age, the difference was not statistically significant (OR = 0.66; 95% CI = 0.33-1.33; p = 0.250). The small number of young dogs observed in this study suggests that there was little rotation or replacement of dogs in the area studied.

5/9
Regarding the habits of these animals, there was no significance (OR = 2.07; 95% CI = 0.15-2.64; p = 0.599). However, it was observed that during the night, among the 82 positive animals, two dogs slept indoors and 50 outside in the yard, and also that among these 82 positive dogs, 37 were restrained and 45 were loose on the properties.
Regarding the environment (Table 4), of the 82 animals that were positive for Leishmania exposure, all lived with other animals (60 with poultry, 29 with cats, 27 with pigs, 24 with horses and 13 with cattle). In the present study, the presence of poultry (hens) cohabiting with dogs was shown to be another RF that influenced the presence or maintenance of infected dogs in the rural regions, as shown by univariate ( observed that of the 82 seropositive dogs, 60 lived on properties on which chickens were also kept. It was found that 20.8% of the dogs sampled were on properties that reported dogs having been put down because of Leishmania exposure (OR = 0.53; 95% CI = 0.26-1.10; p = 0.055). This indicated that the disease was present in this rural environment and that greater attention to it needs to be paid by health authorities. None of the properties reported that any of their dogs had been treated for VL. Among the 82 seropositive dogs, 20 lived on properties that did not have any selective garbage collection.
Finally, regarding the owners' knowledge concerning VL ( Table 5), it was found that 27.2% of the dogs were under the ownership of people who did not know about the disease and how it is transmitted (72.8%); or about its severity and lethality, not only in relation to dogs but also to humans (60.8%). This lack of knowledge concerning the disease among rural populations was a RF for the disease by univariate analyses (OR = 1.74; 95% CI = 0.96-3.21; p = 0.049), whereas this factor ceased to be significant after multivariate analysis. However, it should be noted here that a large proportion (72.8%) of the rural population interviewed reported having a lack of knowledge about the role of the insect vector in relation to transmission of leishmaniasis, and regarding dogs being the main domestic reservoir (64.4%). In addition, 49.2% of the interviewees answered that they were totally unaware of the presence of the vector on their properties, thus corroborating the hypothesis that knowledge concerning CanL among the human population is a crucial factor regarding its prevention.

DISCUSSION
In other studies carried out in Ilha Solteira, the prevalence of CanL in urban areas ranged from 10% to 14% 20,25 , while it was 37.7% rurally 26 and 89% at animal shelters of the Association for Animal Protection 27 . Here, there were seropositive dogs with and without clinical signs for both serological tests performed, but with a greater number of positive dogs with clinical manifestations (51.7% and 56.8%) than positive dogs without clinical signs: 21.9% and 23.9% for ELISA and IFAT, respectively. According to other authors, a large majority of dogs without clinical signs are negative in the different tests routinely used for CanL detection 28,29,30 . However, dogs without clinical signs but positive for CanL may be infectious for sand-flies in a proportion similar to those with clinical signs, and are equally important in the epidemiological chain of the disease 31,32 .
Some studies have shown that dogs without clinical signs are generally seronegative, or have antibody levels that are difficult to detect in serological tests. On the other hand, dogs with clinical signs generally exhibit high levels of antibodies 33  During an epidemiological evaluation of the southeastern and southern regions of Spain, it was observed that the seroprevalence of CanL gradually increases with the size of the animal. This characteristic was found to be an RF for the disease 37 , which 7/9 supports our results. Besides the fact that all dogs are susceptible to Leishmania exposure, Feitosa et. al. (2000) also correlated greater frequency of CanL cases with larger dogs. Since these serve as guard dogs, they are kept in peridomestic areas, and thus are probably more exposed to the vector. Penaforte et al. (2013) also studied this association and suggested that large dogs suffer more sand-fly bites because they are used as guard dogs, living outside houses. found that dogs younger than three years of age and older than seven were at higher risk of contracting CanL, and that the first of these groups was more susceptible than the second.
Amorá et al. (2006) observed that the greatest number of seropositive dogs had semi-domestic habits, which they explained by suggesting that these animals are more exposed to vector action. Some research has shown that dogs from endemic areas exposed at night can be stung by hundreds of sand-flies. This continuous exposure may favor seroconversion and development of the disease, since the parasite is continuously introduced into the skin of these animals 43 .
Although other vertebrate animals can serve as food sources for sand-flies, and favor their maintenance in areas close to homes, some, such as chickens, can also reduce the number of infectious bites in dogs 44 . However, Azevedo et al. (2008), verified that association of the prevalence of seropositivity in dogs is related to cohabitation with other species: chickens were the most frequent cohabitees among the positive dogs, followed by pigs and horses. Borges et al. (2009) demonstrated that poultry have immense potential for attracting sand-flies, and noted that chickens deserve special attention because of their higher frequency among households, as well as their potential for generating a favorable environment for procreation of sand-flies because of the organic waste that they produce. In another study, Barboza et al. (2009) observed the co-presence of chickens, pigs and horses, but found that cats were the most frequent cohabitees with seropositive dogs.
Recently, a cross-sectional study conducted in endemic areas of Cuiabá (MT) has shown that dogs living in rural settings were 1.9 times more likely to acquire the infection than were those in urban environments 19 . Factors relating to the dogs' habits, such as free access to the streets and serving a guard function, as well as the presence of agricultural activity, were considered to be indicators that predicted infection by Leishmania spp. In addition,  and Moreno et al. (2005) observed that poor housing conditions, open sewage ditches, lack of household waste collection, and irregular or absent disposal of sewage were RFs for Leishmania infection in urban areas. Moreno et al. (2005) reported in a study conducted in the metropolitan region of Belo Horizonte that the likelihood that a population would be affected by CanL was six times higher for people who did not know about the vector than for those who were aware of it.
Our data have reinforced the hypothesis that many people still have poor knowledge regarding leishmaniasis and how it is transmitted. This corroborates a recent report by Paulan et al. (2016), who found that rural families established in the "Estrela da Ilha" rural settlement in Ilha Solteira, SP, presented fragmented knowledge concerning the disease, thus resulting in inefficient practices of prophylactic measures against leishmaniasis among humans and dogs in this rural area.
Changes in attitudes in populations is a goal to be achieved over time, since this involves cultural changes, which seem to be a crucial factor regarding the difficulty in attaining control over this zoonosis. According to Borges et al. (2008), knowledge of the forms of VL transmission and vector recognition decreases the risk of contracting leishmaniasis by a factor of 0.79, while lack of knowledge about the disease increases the risk by a factor of 2.57.

CONCLUSIONS
The RFs associated with occurrence of Leishmania exposure in domestic dogs on properties of the "Cinturão Verde" in Ilha Solteira, SP, were large body size among the dogs, the presence of chickens, and lack of knowledge regarding Leishmania among dog owners. After adjustment through multivariate analysis, only dog size and the presence of chickens were related to the presence of Leishmania exposure among the dogs. However, we must emphasize that this probably happened due to the fact that most of the local population had no knowledge about the disease.