Ehrlichia and Anaplasma belong to the subgroup α-Proteobacteria, Rickettsiales order, Anaplasmataceae family (1), are obligate intracellular gram-negative bacteria, replicate in cytoplasmic vacuoles derived from the host cell of mammalian vertebrates mainly in leukocytes and platelets (2) and arthropods vectors such as ticks (3). These bacteria produce emerging zoonotic diseases that infect mainly dogs and cattle, horses, human beings (2), also it has been reported in felines (4). They are multisystemic and characterized by nonspecific clinical signs and hematological alterations in dogs (5) and in human beings (6). They are diagnosed by different laboratory techniques that include indirect (IFI, ELISA) and direct tests (eg Blood Frotis and Polymerase Chain Reaction-PCR) (5,7).

These vector-borne diseases (VBD) have increased markedly in dogs and humans beings and are widely distributed worldwide (8), frequently present in tropical and subtropical areas (9), where proliferate ticks as Rhipicephalus sanguineous and Ixodes spp. which have been identified in South America (10) and whose reservoirs could be wild animals, cattle and pets (8).

The occurrence of VBD can be influenced by several factors such as climate change, and anthropogenic factors such as urbanization, deforestation, demographic changes, economic crisis, displacement of people and animals (8,11). The exploitation of natural resources and loss of biodiversity affect the transmission cycle of ticks, which leads to a variation in vectorial competition (8,11).

Because of its geographical location, the Colombian Caribbean brings together all the characteristics and factors mentioned above that favor the incidence of these infections. This raises a challenge for public health, human and veterinary medicine, differential diagnosis with other microorganisms and the occurrence of coinfections that aggravate the clinical profile (12,13,14), especially in coastal areas such as Barranquilla, where higher prevalence of Ehrlichia sp. has been reported (9,15) in comparison to other regions of the country (5,16,17,18,19,20)

Therefore, it is essential to monitor the spread and increase of these agents in endemic areas, to contrast prospective and retrospective studies, and characterize epidemiological, clinical and laboratory aspects in all regions of the country. Thus, the objective of this study was to determine the prevalence of Ehrlichia canis and Anaplasma sp. and its correlation with epidemiological and laboratory factors in dogs attended in veterinary clinics. Barranquilla, Colombia

Type of study. Descriptive, retrospective of cross-sectional study.

Collection of information. The data were collected between January and August 2015, from secondary sources of information, clinical profiles and data from veterinary laboratories of nine veterinary clinics in the city of Barranquilla, Atlántico, to which 184 dogs were attended, clinically compatible with infection by Ehrlichia and Anaplasma; the data of results of positivity of the techniques extended peripheral blood for the observation inclusions in the cytoplasma for Ehrlichia sp. and Anaplasma sp., and a commercial kit inmunocromatographic for the qualitative detection of antibodies to Ehrlichia canis, epidemiological data such as sex, age, race, and blood count data obtained by conventional techniques such as microhematocrit and plaque counts were also collected.

Analysis of data. Data were analyzed using SPSS 21® and Statgraphics Centurion Plus® to perform Two-way ANOVA analysis with a p value ≤0.05. A descriptive analysis was utilized, determining central tendency measures and dispersion measures, in the case of the continuous variables; for categorical variables, proportions were shown. To measure the correlation between epidemiological and laboratory variables with the positivity probability of the studied infections, with a confidence level of 95% and a p value ≤0.05, correlation coefficients were generated by Pearson.

Ethical aspects. The project was evaluated by the Simon Bolivar University Microbiology program ethics and research committee, defining the study as risk free according to article 87 of resolution 8430/93 of the Ministry of Health of Colombia and law 84 of 1993 according to the Congress of Colombia, because, the data were taken from the stories clinics of the dogs.

E. canis and Anaplasma sp. overall prevalence was 34% (63/184 dogs), the seropositivity for E. canis was 28% (52 dogs), by means of the commercial kit inmunocromatographic (Table 1). Obtained results from the extended peripheral blood demonstrated a prevalence of 6% (11 dogs) for Anaplasma sp. (Table 1) and in 2% (4 dogs) a coinfection with these agents was evidenced from both techniques

In this study we evaluated the infection by E. canis and Anaplasma sp, in Barranquilla, the overall infection was detected in 34% (63 dogs), Higher prevalence in contrast to reported in Medellín, Antioquia 29.7% (52/175). However, the results found in this study were lower, contrasting with those reported in Cartagena, 80% (80/100 dogs), in Barranquilla, 8% (182/223 dogs) positive for E. canis and A. phagocytophilum (9) and 81.6% (178/218) canines positive for E. canis and A. Platys (15); All by detection of antibodies by serological tests.

Differences in sample size and studied canine populations were important factors in determining the prevalence of these infections. In the present study, data were obtained from dogs with owners who visited veterinary clinics, while in the other studies data and samples were obtained from dogs that inhabit streets, shelters and that belong to the military service who present a higher prevalence for these infections in comparison with dogs with owners (9,15).

On the other hand, the prevalence found of E. canis 28%, differs with that reported in other regions; Showed to be higher than reported in Medellín 25% (9) and 24.8% (22); Similar to what was found in Ibagué, Tolima 31.6% (17); And lower than in other cities such as Bucaramanga, Norte de Santander 89.7%, Villavicencio, Meta 83.8%, Bogota, Cundinamarca 66% (18), Cali, Valle del Cauca 40.5% (19), and finally in the Caribbean region, Cartagena 79%, Barranquilla 82% (9) and 74.4% (15). As for the prevalence of Ehrlichia sp., found in this study was higher in contrast to Medellin 11.2% (5) and Florencia, Caquetá 22.4% (16), considering that these reports may contain any species, including E. canis, so the comparison with these studies was important.

The notable differences in frequencies reported in each region can be explained by two factors, among others a) influence of the climatic conditions of the regions studied, it has been demonstrated a higher prevalence in tropical wetlands and coastal mainly Barranquilla and Cartagena, conditions that favor the proliferation and the cycle of transmission of the vector (9) and b) the different techniques used for the confirmation of the infection, in Colombia the veterinary clinics have generally used ELISA, Immunochromatography and blood smear (21), although the highest prevalence rates have been determined by serological tests (15,18).

The wide panorama in the use of techniques for the E. canis diagnosis has its own limitations.Serological tests are only useful in the subclinical and/or chronic phase of the infection, which is unknown in this study; while the blood smear is useful in the acute phase, often asymptomatic, the low sensitivity in both natural (23) and experimental infections (24), and the low bacteraemia, make microscopic identification difficult (25); In contrast the technique of polymerase chain reaction (PCR), although more sensitive also has the limitation of throwing false negatives in the subclinical phase of the disease (18). Taken this into account and the different phases and variety of clinical manifestations, it entails to define several aspects in the diagnosis of this disease; History of tick infestation, endemicity of the region with the presence of other related VBD, clinical signs and hematological and biochemical alterations with the respective confirmation by molecular methods (7).

With respect to Anaplasma, few reports have been found in Colombia, most studies have been directed towards livestock animals. The prevalence of Anaplasma sp. 11%, was higher in contrast to what was reported in Bucaramanga 1.1% canines (23). Although the prevalence is low, the use of a microscopic technique that has been shown to be less sensitive, yields valuable data, taking into account that the prevalence reports in the country for commercial serological tests, Barranquilla 53.2% canine with A. platys, (15) and 40% canines with A. phagocytophilum (9), Cartagena 51% and Medellin 11% canines with A. phagocytophilum (9), could be limited to the low specificity or cross reactivity of some kits between Several species of Anaplasma and Ehrlichia (26).

E. canis and Anaplasma sp. coinfections have been reported mainly in the Caribbean region, between 40-49% of dogs studied in Cartagena and Barranquilla (9,15); these results were higher than those found in this study 6% (4/63). It has been suggested that these agents simultaneous presence may lead to more serious immunological and hematological effects (13,14).

As for the epidemiological characteristics, the mean age reported in this study (52 ± 41, range, 1-192 months) is similar to that found in Medellín (62.2 ± 42.5, range, 2 -216 months) (22), the age group with the highest dogs frequent with E. canis and Anaplasma sp. infection, 13-84 months, as reported in Florencia 89.8% (16), Medellín 55% (22) and Cali 30.7% (19). Significant statistical associations between seropositivity and age have been identified in other studies (16,27). As reported, infection can be mainly associated in dogs older than one year in comparison with younger dogs, due to their immunological status, greater exposure to vectors and hosts and less care by their owners (16,28).

As for sex, there were discrepancies in the results, while males were the most susceptible to E. canis and females to Anaplasma sp. infections; similar findings were reported in Cali 34.7% (19) and Florencia 56% (16) with higher number of male dogs with E. canis and Erhlichia sp. respectively; in contrast, in Ibagué (17) the frequency of females with E. canis was higher 34.3% and results with equal proportions for males 25.9% and females 24% were reported in Medellín (22); thus there is currently no evidence of clinical susceptibility of dog sex-related infections in the country.

Likewise, the prevalence of dogs with Erhlichia sp. was higher in medium and small breeds, mainly French Pooodle, Labrador and YorkShire Terrier as reported in other studies (16,17,19); however, in Medellín there was no infection predisposition in the different breeds (22).

On the other hand, the most significant laboratory data (P=<0.05) were anemia, thrombocytopenia and leukopenia with lymphopenia, results which were similar to another study in dogs with experimental infections (24) and contrary to results reported in Ibagué in which no laboratory data was statistically significant (17).

Infection with E. canis showed more notorious haematological changes such as anemia and thrombocytopenia with respect to infection by Anaplasma sp; mean values ​​of platelets in dogs with E. canis were below the normal value (194 ± 46 103 xmm3), as reported in other studies in experimental infections (14,24). Thrombocytopenia in E. canis infections has been associated with the development of antiplatelet antibodies, whereas in infections by Anaplasma platys, it occurs by direct destruction of platelets (14,24), therefore, it is presumed in this study, that E. canis prevalence is higher, by means of serological tests that detect IgG antibodies production in the disease subclinical or chronic phase, this prolongation of immune response would favor thrombocytopenia in these dogs, whereas the prevalence of Anaplasma sp was lower through blood smear, which probably indicates acute infection.

It is important to mention that the moderate decrease in platelets found in this study is not confirmed for true thrombocytopenia, a pseudothrombocytopenia in vitro is not dismissed (7), partly because of the destruction and satellitism of blood samples, due to the delay with which they often come to the laboratory to be processed. Therefore, it is essential to correctly evaluate platelet values ​​in these infections.

Pancytopenia was also present mainly in dogs with E. canis 11.5%, as reported in Cebu, Philippines 5.7% (29), this could aggravate the clinical profile in the chronic phase, due to cellular production failure in the bone marrow (30). Also, in dogs that presented co-infection, even when it was low, one dog developed pancytopenia, and reason why, it could not be ruled out the possibility of occurring in the other dogs, more complex hematological and clinical alterations as has been reported in other studies (14,23).

Although, the rest of variable were not significant, in the descriptive analysis, the most important epidemiological and laboratory characteristics, such as anemia, thrombocytopenia and leucopenia, were more frequent in dogs from 13 to 84 months and medium breeds. However, the lowest mean values ​​for hematocrit (37.6 ± 8.3%), leukocytes (5.3 ± 3.0 103x mm3) and platelets (184 ± 43 103x mm3) were present in <12 months dogs.

In conclusion, this study findings corroborate the high prevalence and endemicity of these agents in the region and other areas of Colombia, this poses a challenge for human and veterinary public health and zoonoses centers, complicating the differential diagnosis and the probabilities of transmission to humans beings, mainly of occupational type and people in frequent contact with dogs.

The authors thank the participating veterinary clinics, the Laboratory of Veterinary Diagnosis of the Caribbean (LABVECA, Barranquilla, Colombia) and Lizeth Paola Mejía Mendoza, for their support and collaboration in collecting data and valuable contributions to this study.