Many authors consider male infertility and mistimed breeding as the major causes of reproductive failure (RF) in bitches. However, once these causes are excluded, the role of infectious diseases in pregnancy losses should be addressed. The impact of infectious diseases on the normal progression of pregnancy can be associated with the direct action of the infectious agents or their toxins, inducing placentitis that leads to fetal death (^{Givens and Marley, 2008}). In dogs, the viral action of the canid herpesvirus 1 (CaHV-1) and the bacterium Brucella canis have been found to be the main causes of reproductive problems. In Brazil, leishmaniasis, parvoviruses, and tick-borne diseases (Anaplasma platys and Ehrlichia canis) are endemic, particularly in the North-East region, and all these pathogens can cause anemia. Apart from these agents, previous studies on reproductive disorders in dogs have neglected other pathogens such as Neospora caninum (^{Greene, 2012}). Due to the lack of research concerning potential causes of RF, this study was designed to investigate the association between infectious agents, unexplained RF (abortion, stillbirth, or fetal death), and anemia in bitches.

Samples from 36 pregnant dogs (20 with RF and 16 without RF) treated at the Veterinary Hospital Jerônimo Dix-Huit Rosado Maia-HOVET in the Federal Rural University of the Semi-Arid (UFERSA) in Mossoró/RN, North-East Brazil, were obtained according to the Ethics Committee on Animal Use (CEUA) under license number 23091.006326/2014-88 (date of approval March 3^rd, 2015). Depending on the clinical condition: RF (unhealthy) or without RF (healthy), tissue samples were collected through ovariosalpingohisterectomy, Caesarean, parturition, and through the post-mortem examination of the uterus, placenta, and fetus. Blood samples were collected after the RF episodes or after parturition in healthy bitches, and a complete blood count (CBC) was performed in both groups. In both groups, using a STRATEC Molecular kit (Invitek), DNA was extracted from the female dogs (blood/serum for all bitches, and uterus/placenta in unhealthy group) and from fetus (amniotic fluid, fetal abdominal fluid, and pool of liver/spleen and kidney/lung) of the bitches with RF. The samples were subjected to molecular/serological immunodiagnostic tests, both for detection of pathogens related to RF and for anemia (hemoglobin, erythrocytes, and hematocrit below the reference values, Table 1). A Chi-square test was used to determine if there was a relation between the presence/absence of the infectious pathogens and RF. In addition, an analysis of variance (ANOVA) was conducted to determine if RF was related to the blood biochemistry parameters measured (anemia).

None of the animals tested positive for the main infectious causes of RF in bitches in either group (Table 2). Canine parvovirus (CPV) was detected in the maternal blood without hematological alterations, and fetal tissues had signs of autolysis, which probably occurred after the infection (Table 3). For leishmaniasis, five females (C2, C4, C6, C9, C16; RF group) were serologically positive (Table 2). Excepting C16, the other four animals also had co-infection with A. platys; C6, C9 and C16 also had anemia with CBC with characteristics of infection (leukocytosis with neutrophilia), and C9 was positive for leishmaniasis in PCR (Table 2). Some bitches (C1, C3, C7, C13, Tab.2) presented co-infection of E. canis/A. platys, demonstrating anemia and thrombocytopenia (except female C3 which showed only anemia and CBC with characteristics of infection).

The other females with ehrlichiosis/anaplasmosis had at least one alteration in the CBC and alterations in the fetuses that suggest an infectious pathogen (Table 3). The presence or absence of the pathogen did not significantly explain the occurrence of RF (χ2 = 0.02, P>0.05, n = 20). Regarding the hematological data, a relation between anemia and RF in the animals was observed when considering the erythrocytes (F 32.2, DF= 1.33, P<0.05, Figure 1a), hemoglobin levels (F = 32.8, DF = 1.33, P<0.05, Figure 1b), and the haematocrit (F = 35, DF = 1.33, P<0.05, Figure 1c). This indicates a significant difference between animals who suffered RF (unhealthy group), and those who did not experience RF (healthy group). About the data in Tab. 2 and 3, CaHV-1, brucellosis, and neosporosis may cause abortion of fetuses in late pregnancy and foetal autolysis (^{Greene, 2012}). A previous work also did not find B. canis or CaHV-1 in maternal tissues of bitches with RF (^{Mir et al., 2013}). Since dogs are vaccinated for CPV, RF associated with this disease is unlikely to occur due to mass population immunity (^{Givens and Marley, 2008}). Leishmaniasis is zoonotic and endemic in North-Eastern Brazil (^{Costa et al., 2014}); it is a systemic disease, which can lead to anemia and placentitis (^{Oliveira et al., 2015}).

Figure 1 Reproductive failures (RF) in bitches due to anemia. Blood analysis comparing females that suffered of RF (unhealthy) and females that did not suffer of RF (healthy). A, Quantification of erythrocytes (mean ± standard error; in 1x106/mm3); B, Number for hemoglobin (mean ± standard error; in g/dL); C, Quantification of haematocrit (mean ± standard error; in %). 

The other four animals also had co-infection with A. platys, so we cannot infer which disease caused the anemia. Their features in fetuses (Table 3) suggest a pathogenesis of infectious origin. Ehrlichiosis and anaplasmosis are related as emerging zoonotic diseases (Arraga-Alvarado et al., 2014), which are not associated with RF in animals. These pathogens may have caused weakness and consequently RF. Thrombocytopenia in dogs with leishmaniasis (C6), ehrlichiosis (C1, C7, C13), and anaplasmosis (C1, C7, C13, C19) is expected; however, RF was not previously described as a clinical sign (Borin et al., 2009). The presence of these pathogens induces anemia, which can increase the likelihood of RF. Anemia has been associated with animals chronically infected with multisystemic diseases (anaplasmosis, ehrlichiosis, or leishmaniasis), and clinical signs vary with the severity of the disease and the presence of co-infections (Borin et al., 2009; ^{Ferreira et al., 2014}). Genetic diversity of E. canis strains can express different forms of hematological and clinical manifestations of the disease. However, previous studies on clinical signs of anaplasmosis, ehrlichiosis, or leishmaniasis were conducted in nonpregnant animals. The co-infection of E. canis and A. platys in nonpregnant dogs resulted in more pronounced anemia and thrombocytopenia (^{Gaunt et al., 2010}). Maternal and perinatal mortality is significantly higher in anemic women (^{Bencaiova et al., 2012}). Maternal systemic diseases can result in fever, anemia, or endotoxemia, which in turn can result in RF (^{Givens and Marley, 2008}). The endotheliochorial placenta of dogs allows high blood supply to the fetus. Maternal infections by intracellular parasites can lower oxygen levels in the blood transported to the fetus, leading to RF. The principal reproductive pathogens were not found in this study, which suggests that the systemic disease diagnosed in the maternal blood is related to RF. Hence, bitches with anemia caused by vector-borne diseases are significantly more likely to experience RF than those with normal hematological values (Figure 1).