Hematological, clinical and epidemiological aspects of Hepatozoon canis infection by parasitological detection in dogs from the rural area of Sousa, Paraíba, Brazil

ABSTRACT: Hepatozoonosis is caused by protozoa of the genus Hepatozoon. In dogs, the infection is caused mainly by Hepatozoon canis, and there are a few descriptions of the prevalence of this infection in the Northeast region of Brazil, especially in the semi-arid region. Therefore, we aimed to determine the prevalence of Hepatozoon canis infection in dogs in the rural area of Sousa, Paraíba, Brazil, as well as to determine the possible clinical and epidemiological aspects of this infection. Ninety-eight dogs in the rural zone of the municipality of Sousa that were at least 4 months old were evaluated, regardless of their breed or gender. Clinical examinations were carried out, and samples of systemic and peripheral blood were collected to determine the presence of the parasite in blood smears and carry out hemograms. In addition, epidemiological questionnaires about animal health and food management were completed. The prevalence of H. canis infections in dogs was 8.1% (8/98). There were three main changes in the hematological status: thrombocytopenia, anemia and hyperproteinemia, mainly related to percentage of leukocyte infection ≥5%, and also to the presence of clinical signs such as mucopurulent secretion, lymphadenomegaly, dry skin, pale mucous membranes, and lean or cachectic body score.


INTRODUCTION
The hepatozoonosis is a debilitating and an immunosuppressive disease with a difficult clinical diagnosis and its infection have been described in amphibians, reptiles, birds and mammals (O'DWYER, 2011).
The clinical presentation of hepatozoonosis varies according to the level of parasitemia, the immune status of the animal, and its association with other pathogens (MUNDIM et al. 2008;SPOLIDORIO et al., 2009). Animals may have edema, intermittent fever, weight loss, pale mucous membranes, pain, diarrhea, vomiting, and gait abnormalities (BERNARDINO et al., 2016). However, because these clinical findings are nonspecific, laboratory tests are needed for confirmation (O'DWYER, 2011).
Blood smear examination is the most frequent diagnostic method for H. canis infection used to detect gamonts within neutrophils or monocytes, even considered of low sensitivity (BANETH et al., 2003;ROTONDANO et al., 2015). The polymerase chain reaction (PCR) has provided an advanced diagnostic method considered of high sensitivity (LOPES et al., 2019); however, it is not widely used in the routine of veterinary clinics. Hemograms and biochemistry could assist on diagnosis of vector borne diseases, with similar results for several infections (MUNDIM et al., 2008;PAIZ et al., 2016).
The non-specific nature of the clinical signs in infected dogs, associated with the low sensitivity and specificity of the blood smear technique is a hindrance to the good understanding of the infection caused by this parasite. Due to the lack of studies on H. canis infection and its deleterious effects on the health of infected animals in the semi-arid region, we aimed to determine the prevalence of H. canis infection in dogs in the rural area of Sousa, Paraíba, and to describe the main epidemiological, clinical and hematological aspects of this infection.
From April to December 2017, 98 blood samples were collected from dogs that were at least four months old, regardless of their breed and gender, from the rural area of Sousa, Paraíba. The number of samples to be collected was calculated based on an expected prevalence of 9.3% H. canis infection, (BERNARDINO et al., 2016), confidence level of 95% and error of 5% (THRUSFIELD, 2005), resulting in 45 samples to be collected; however, 98 samples were collected for security. Blood samples (3.0 mL) were collected from the external jugular vein of each animal, placed into tubes containing 10% EDTA, stored in styrofoam with ice, and submitted to the Laboratório de Parasitologia Veterinária -LPV and to the Laboratório de Patologia Clínica -LPC, at the Hospital Veterinário Adílio Santos Azevedo -HV/ ASA, IFPB, Sousa, Paraíba. Blood counts were performed according to JAIN (1993). Blood smears were also performed at the collection site, using peripheral blood obtained with a lancet on the inner side of the ear tip, after trichotomy and asepsis with 70% alcohol.
Blood smears were made in order to categorically perform the differential leukometry and morphological evaluation of blood cells, as well as the search for blood parasites. The blood smears were stained with the Panótico Rápido ® commercial kit, from Laborclin-products for laboratories LLC, Pinhais-PR (JAIN, 1993).
Hemoparasites were investigated by optical microscopy (100x objective) in blood extensions of systemic and peripheral blood (tip of ear), stained with the Panótico Rápido ® kit (Laborclin-products for laboratories LLC, Pinhais-PR). In each slide, a total of 200 leukocytes were counted to obtain a reading pattern, and in order to determine the degree of parasitemia, the percentage of leukocyte infection (PLI) was determined by quantifying the leucocytes with gamont infection.
To determine the clinical profile of the infected animals, we collected information about the history and physical examination of each animal using a form. Simultaneously, investigative questionnaires were applied to all animal guardians, containing objective and subjective questions about sanitary management, feeding, and contact with other animal species.
During the clinical examination, ectoparasites were collected and placed in flasks containing 70% ethanol, individualized by animal. The identification of ectoparasites was performed under a stereoscopic microscope according to traditional taxonomic keys (ARAGÃO & FONSECA, 1961).

Statistical analysis
The data obtained from the epidemiological questionnaires were analyzed in two stages to determine the risk factors: univariate and multivariate analyses. Initially, the univariate analysis was performed with two groups of animals: positive and negative, and compared the analyzed variables between the two groups. The variables that presented a p-value ≤ 0.2 by the chi-square test or Fisher's exact test (ZAR, 1999) were selected for multivariate analysis using the multiple logistic regression (HOSMER & LEMESHOW, 2000). The level of significance adopted in the multivariate analysis was 5%, and all analyses were performed with the SPSS 20.0 software for Windows.

RESULTS
A prevalence of 8.1% (8/98) of H. canis infection in dogs was obtained, based on the diagnosis through the visualization of gamonts under optical microscopy. It was observed that 6.1% (6/98) were positive for Ehrlichia spp., with no comorbidity with H. canis.
No other hemoparasites were diagnosed in the blood smears. All H. canis positive animals were negative for anti-Leishmania infantum, anti-Ehrlichia canis, and anti-Brucella canis antibodies and for Canine Distemper Virus, Canine Parvovirus, Canine Coronavirus, and Dirofilaria immitis.
According to the geographical distribution of the infected animals, there was a higher prevalence in the Emiliano Zapata Settlement 19.4% (7/36), followed by São Gonçalo 5.2% (1/11). In the other evaluated locations, there were no positive animals.
Oliveira et al.
The univariate analysis for the assessment of factors predisposing to H. canis infection is described in table 1. Regarding the most relevant predisposing factors, among the positive animals, 87.5% (7/8) were raised free, and 75% (6/8) had infestations with R. sanguineus at the time of the clinical examination. Most of these H. canis infected animals had never received any vaccination (87.5%; 7/8). The variables presented in table 1 were selected (P≤0.20) for the multivariate analysis,;however, no risk factors were identified by multiple logistic regression.
Systemic blood PLI ranged from 1 to 22%, and that of peripheral blood from zero to 20%. Among the infected dogs, seven (87.5%) had anemia, six (75%) had thrombocytopenia, and five (62.5%) had hyperproteinemia, with a pattern of changes being observed mainly when the PLI was ≥5% (Table 2). Among the eight infected animals, three (37.5%) showed all the clinical signs described in table 3. The presence of at least two clinical signs was reported in 75% (6/8) of the animals, with the main associated signs being dry fur, pale mucous membranes, and lean or cachectic body score.

DISCUSSION
The prevalence of H. canis infection reported in the present study was 8.1%, which is similar to the prevalence of 9.3% reported by BERNARDINO et al. (2016) in the municipality of Areia, Brejo Paraibano, but was higher than that reported by RAMOS et al. was a greater ease of transmission of the parasite among the contact of dogs and the rural environment, due to the free access to other farms, a fact suggested by the high prevalence in one location: the Emiliano Zapata Settlement (19, 4%; 7/36). Rhipicephalus sanguineus was the only species reported to infest positive animals, reinforcing the assertion that canine Hepatozoonosis is frequent in areas where this tick is endemic (SPOLIDORIO et al., 2009).
When comparing the positivity for H. canis in relation to the gender, breed and age of the animals, no statistically significant difference was observed (p> 0.05). Probably, this is because the transmission of Hepatozoonosis depends on the ingestion of infected ticks, and the greatest epidemiological importance is in the presence of tick infestation, in which there is also no predisposition for the variables mentioned (LABRUNA & PEREIRA, 2001).
Contrary to what is reported in the literature, that hepatozoonosis is a disease that develops with hematological changes without a pattern of association to the clinical signs (EIRAS et al., 2007;BERNARDINO et al., 2016), in the present study, there were simultaneously high levels of PLI in some animals with hematological changes and clinical symptomatology.
There were three main changes in the hematological status: thrombocytopenia, anemia, and hyperproteinemia, mainly related to PLI ≥5%, and also to the presence of clinical signs, such as secretions, lymphadenomegaly, dry skin, pale mucous membranes, and lean or cachectic body score.
The most common hematological abnormalities in H. canis-infected dogs have been anemia, leukocytosis, and neutrophilia or lymphopenia (MUNDIM et al., 2008). Although, Clinical hepatozoonosis was observed in 6 animals (80%), related to PLI >5%, presenting ocular and nasal mucopurulent secretion, lymphadenomegaly, dry fur, pale mucous membranes, and/or lean or cachectic body score, considered to be nonspecific. In two animals (20%), when the PLI was <5%, the infection occurred subclinically. In addition, no coinfections were found in animals positive for H. canis. The degree of parasitemia was also associated with the clinical manifestation as reported by KARAGENC et al., (2006), who stated that dogs with parasitemia up to 3% developed mild disease, up to 18% moderate disease, and up to 39% severe disease. However, it differs from the literature that characterizes the disease as intercurrent with other immunosuppressive diseases (BANETH et al., 2003;SPOLIDORIO et al., 2009;MEGID et al., 2016;AGUIAR et al., 2019).
Direct blood smear research has been described as highly specific, but with low  sensitivity (RAMOS et al., 2010;O'DWYER, 2011;ROTONDANO et al., 2015;PERLES et al., 2019). In general, when the parasitemia is high, blood smear is extremely efficient, however, at low levels of infection, detection becomes difficult. It is worth mentioning that low levels of parasitemia do not indicate absence of infection, and a false negative diagnosis is common in these cases (EIRAS et al., 2007). The diagnosis of H. canis can be done with serological and molecular tests for higher sensitivity, such as Indirect Fluorescence Antibody Test (IFAT), ELISA and Polymerase Chain Reaction (KARAGENC et al., 2006;O'DWYER et al., 2011). During the present study, it was noted that the technique was effective not only for diagnosing the infection, but also for determining the percentages of the infection. It was noted that the observation of 200 leukocytes per slide in two slides, one of peripheral blood and the other of systemic blood, influenced the greater effectiveness of the technique.
In six (75%) of the positive animals, a higher PLI was observed in peripheral blood smears than in systemic blood. This occurs, according to VINCENT-JOHNSON et al. (1997), because the gametocytes quickly leave the leukocytes after blood sampling, mainly through contact between the blood and the anticoagulant, which may make it difficult to search for gametocytes in blood smears.
Most positive animals belonged to families that lived in precarious socio-economic conditions, lacking knowledge about the correct handling of animals, where basic health and food care was neglected. Moreover, as they come from rural areas, many animals were confined during the day and released at night to protect the homes. These factors, linked to the environmental conditions that favored the tick cycle, made the animals more predisposed to the infection, favoring the high prevalence observed and the occurrence of animals with a chronic infection status.
For the control of H. canis infections, infestation of animals by ticks should be avoided. Considering the high prevalence found, it is advisable that veterinarians carry out the search for hemoparasites, with slides of peripheral and systemic blood, observing at least 200 leukocytes per slide, especially when there are cases of hematological changes and non-specific clinical signs. This study demonstrated the need for continuous epidemiological studies, for the detection of endemic areas for hepatozoonosis, since it is a disease with a difficult clinical diagnosis.
It was concluded that the prevalence of H. canis infection in dogs in the rural area of Sousa, Paraíba, was high, with the majority of positive animals showing simultaneously hematological changes, clinical symptomatology and a high degree of parasitemia.