Visceral leishmaniasis (VL) is an emerging zoonosis of great significance to public health; its geographic distribution is restricted to tropical and temperate regions [1]. About 90 % of the individuals infected with VL in Latin America are in Brazil. Zoonotic VL is a potentially fatal disease and domestic dogs are considered the main epidemiologic reservoir of L. infantum (syn. L. chagasi) due to intense skin and viscera parasitism. It is believed that some infected dogs do not develop clinical signs and remain asymptomatic for inconsistent periods of time [1, 2]. To control the spread of the disease, the Brazilian Ministry of Health created the Visceral Leishmaniasis Control and Surveillance Program (VLCSP), which is in charge of measures throughout the country, such as early diagnosis and treatment of human VL, identification and euthanasia of seropositive dogs, control of insect vectors, and health education [2]. Diagnosis of canine leishmaniasis (CL) is usually carried out using serologic enzyme-linked immunosorbent assay (ELISA), immunochromatographic tests, indirect fluorescent antibody test (IFAT), or molecular tests [1, 2]. Ehrlichia canis, a rickettsia microorganism transmitted by ticks, is one of the major vector-borne diseases in dogs with a worldwide distribution and is directly involved in the infection and spread of CL [3].

Despite the measures adopted by the VLCSP to control VL [2], it is suspected that the disease is spreading fast throughout the western region of São Paulo state [4–6]. From 1997, when Lutzomyia longipalpis was demonstrated in the municipality of Araçatuba in the neighboring region of Presidente Prudente, up until 2014, 93.3 % of cases of canine or human transmission of VL in São Paulo occurred in the western region [7]. Our aim was to determine the prevalence of CL using different methods and the prevalence of E. canis infection in Presidente Prudente as well as the spatial dispersion of the disease in western São Paulo state, Brazil.

In the first phase of the study, 4547 dogs were enrolled from different areas of Presidente Prudente and Montalvão district, either by active or passive search. Clinical examinations were performed and of the serum samples submitted to the ELISA test for CL, 509 (11.2 %) were positive (Fig. 1): 204 samples were positive by IFAT, with a seroprevalence rate of 4.5 % (Fig. 1).

In the follow-up study of 170 animals, 56 were positive for the CL ELISA test and 20 were positive for the DPP test, with prevalence rates of 32.9 and 11.8 %, respectively, and 26 remained positive by IFAT with a prevalence rate of 15.3 %. With a seroprevalence rate of 66.5 %, 113 animals were found to be positive for E. canis infection (Fig. 1, follow-up). Forty-eight (28.2 %) of the animals showed leishmaniasis/ehrlichiosis co-infection by the ELISA/IFAT test and 10 (5.9 %) by the DPP/IFAT test. There was a significant positive association between Leishmania and E. canis infection (P < 0.0001). In the clinical examination of the follow-up cohort, 113 of 170 dogs (66.5 %) were classified as asymptomatic; 57 of 170 (33.5 %) presented one or more clinical signs connected with CL. In the follow-up, the kappa statistics for the results between ELISA and DPP resulted in fair agreement (κ = 0.222, 95 % CI, 0.04–0.2).

Figure 2 shows the outcome for the IFAT serological confirmatory test according to the Brazilian Ministry of Health [2]. Seropositive animals were plotted according to the dog owner’s address at the time of collection. Based on the Gaussian kernel, it was possible to detect a concentration of positive tests in the downtown area, and the southwest and west regions of Presidente Prudente city and Montalvão district. The spatial distribution of vectors and CL throughout the study period in São Paulo state is shown in Fig. 3. L. longipalpis was found in 148 counties of São Paulo state (Fig. 3b); 32 of 45 counties (71.1 %) were located in Presidente Prudente’s mesoregion, of which eight are under investigation. CL was found in 102 counties (Fig. 3c); 15 of 45 counties (33.3 %) were located in Presidente Prudente’s mesoregion (Fig. 3d).

In this study carried out in Presidente Prudente city and in the district of Montalvão, the seroprevalence of CL was 11.2 and 4.5 % by ELISA and IFAT, respectively. When a follow-up was conducted, 32.9 % were positive on ELISA, 15.38 % on IFAT, and 11.8 % in the DPP test. As far as we know, this is the first description of autochthonous CL in this region. Furthermore, the disease is spreading to the border of Paraná state in the southern region of Brazil.

Until 1997, VL was known in the state of São Paulo only by imported cases. The first case of autochthonous CL was described in 1998 in Araçatuba, an area located about 160 km from Presidente Prudente, in the northwest region of the state [5, 6]. Thereafter, the infection spread throughout the administrative areas of Bauru, Marília, and São José do Rio Preto. Leishmaniasis has been detected in humans and dogs in the western region of São Paulo state, the first in the municipality of Dracena in 2005, considered by Brazil Health Ministry to be an area with a high transmission rate [4–6]. In São Paulo state, the vector and canine and human leishmaniasis are controlled by the Health Secretary of São Paulo state according to the Manual of Surveillance and Control of American Visceral Leishmaniasis [8]. Despite the measures adopted to control CL, the disease has spread quickly throughout the municipalities of the Dracena microregion. In recent years, like a ghost projecting its shadows, the disease has reached Presidente Prudente County, where canine VL was found in 2009 [5, 7].

In 2012, L. longipalpis and CL were found in some of the border counties of Paraná state, a region with few cases of VL [14]. With a large road network linking different regions, small towns and increasing levels of circulating people, animals and goods, the border area of both states is changing rapidly. Livestock breeding is being replaced by extensive agricultural mechanization of sugar cane plantations with aerial spraying of insecticides and herbicides. This new scenario may be slowing the burden of the disease in Paraná state. This scenario is not shared in Dracena and Presidente Prudente regions; they are notably less mechanized with predominantly areas of cattle pasture. There is a strong hypothesis that the route of expansion and dissemination of VL in the west and northwest region of São Paulo state originated in Bolivia [15]. The vector and parasite reached Brazil through Corumbá, on the western border of Mato Grosso do Sul state, moving on to Campo Grande, in the central region and many years after the disease was found in Três Lagoas, on the border of Andradina and Dracena, western São Paulo state. Later, it reached Bauru following the northwest railroad, Marechal Rondon highway, and recently, throughout the construction of the Bolivia-Brazil gas pipeline [6, 15].

The reasons why CL is spreading fast in the western region of São Paulo and in the direction of Paraná state are not well understood. With a tropical climate, dry winters and wet summers, the region consists of dozens of small towns and villages, and is historically one of the poorest regions in São Paulo state. It neighbors, Mato Grosso do Sul state and Araçatuba County, are well-known endemic foci of VL in Brazil. Certainly the lack of control of the canine population is one probable mechanism; packs of domestic and homeless dogs wander freely in urban areas, most of them without a zoonosis/vectors control service [4]. An overwhelming factor is the overlapping possibility of VL and cutaneous leishmaniasis throughout the region. Since 1950, outbreaks have occurred, mainly in the region of Pontal of Paranapanema [5]. In the present study, surprisingly, a significant number of infected dogs were found in the business district. Empty areas on the map do not mean the absence of infected animals; on the contrary, it means those places were not surveyed. Montalvão is a higher district of Presidente Prudente County. Linked by a highway, there is a daily flow of people for shopping, health care, work, and education services sharing the same risk factors for CL as Presidente Prudente.

The CL seroprevalence of 4.5 % is about 5-fold lower than the average rates of 23.8 % found in the counties of Dracena’s microregion, varying from 4.9 % in Monte Castelo to 29.2 and 30.0 % in Dracena and Santa Mercedes, respectively [4]. We suggest that low levels of parasites were circulating in Presidente Prudente city when the samples were obtained from dogs during the 2010–2011 period, only 1 year after the vector and CL were reported. Possibly, this is also the reason why high rates of discordance between the ELISA and IFAT methods were found. In endemic areas, with increasing rates of circulating parasites, the results between ELISA and IFAT tend to be closer [16–18]. Serological tests for leishmaniasis diagnosis, particularly in epidemiologic surveys of the active search, lack specificity and sensitivity because they are not completely purified and cross reactions may occur, mainly with canine blood-borne agents such as E. canis [17, 18]. In the city, the dispersion routes of CL occurred via allochthones dogs coming from surrounding endemic counties. Worldwide, in urban areas, the domestic dog is the main source of infection for the vector. They exhibit intense cutaneous parasitism, which allows easy infection from sandflies, and the maintenance of the epidemiologic cycle of transmission [17].

With an interval of 64.5 ± 2.05 days, 170 ELISA-positive dogs in the baseline study were re-tested and 56 (32.9 %) were positive in ELISA, 26 (15.3 %) were confirmed in IFAT, and 20 (11.8 %) were positive in the DPP test. An array of factors may be involved in the 3-fold reduction in the rate of ELISA-positive dogs in the follow-up compared with the baseline study. In a scenario of low endemicity, few parasites are in circulation, reducing the burden of infection; healthy animals may clear the parasite via a cell-mediated and humoral immune response resulting in low antibody concentration with negative or borderline titres at the follow-up. E. canis/IFAT showed a seroprevalence of 66.5 and 28.2 % in the animals with leishmaniasis/ehrlichiosis co-infection. Furthermore, significant levels of dogs infected with E. canis were found compared with CL, confirming the suspicions of the veterinarian team of the CZC in Presidente Prudente. Occurring in a wide geographic distribution, especially in tropical and subtropical areas, E. canis infection has a role in increasing susceptibility and making the clearance of L. infantum (syn. L. chagasi) difficult [3]. However, there are conflicting results on the role of E. canis in the cross-reaction of the serologic diagnosis of CL due to lack of specificity [3, 19, 20]. Our results demonstrate that canine ehrlichiosis was an important differential diagnosis of CL because of the prevalence of co-infected dogs. In areas of low CL endemicity, this aspect must be considered by clinicians; symptomatic dogs can be falsely diagnosed with CL and euthanized [19]. No data on the seroprevalence of ehrlichiosis were found in the region.

The association between DPP positive results and the presence of active clinical disease in both humans and dogs is well known [21]. In acute VL, the host may produce specific antibodies against replicating Leishmania. DPP antigens occur predominantly on amastigotes (the replicating form on human or dog hosts) and not in promastigotes [22]. In Brazil, until 2012, CL was diagnosed using IFAT to confirm positive cases detected by ELISA. After 2012, the DPP CL rapid test was introduced for screening with ELISA as a confirmatory test [23].

In the follow-up, 33.5 % of the animals were classified as symptomatic after clinical examination. These dogs did not originate from the passive or active search and they are not a representative sample for estimating CL.

The data presented here allow us to conclude that low CL, suggesting low circulation of parasites and high prevalence rates of ehrlichiosis, were found in Presidente Prudente city. A cross-reaction between L. infantum (syn. L. chagasi) and E. canis antibodies may have occurred. CL is moving through the western counties of São Paulo state toward the border of Paraná state. Hence, implementation of integrated actions to fight the vector and the parasite, as well as the early diagnosis of dogs and humans, are needed for effective monitoring of the disease; in addition, strategies for epidemiologic control must be implemented.