Chronic Canine Ehrlichiosis (Ehrlichia canis): A Retrospective Study of 19 Natural Cases
Nineteen dogs from Greece with chronic ehrlichiosis were studied. The dogs exhibited bicytopenia or pancytopenia, bone marrow hypoplasia, seroreactivity to Ehrlichia canis (E. canis) antigens, and had no history of drug or radiation exposure. Anorexia, depression, severe bleeding tendencies, hypoalbuminemia, and increased serum alanine aminotransferase activity were also hallmarks of the disease. All these animals eventually died, irrespective of the treatment applied. Some dogs were also serologically positive for Rickettsia conorii, Leishmania infantum (L. infantum), and Bartonella vinsonii subspp. berkhoffii. Polymerase chain reaction testing of bone marrow samples revealed E. canis, Anaplasma phagocytophilia, Anaplasma platys, and L. infantum in some dogs. Concurrent infections did not appear to substantially influence the clinical course and final outcome of the chronic canine ehrlichiosis.
Introduction
Ehrlichiosis in the dog is caused by infection with one or more ehrlichial species, including Ehrlichia canis (E. canis), Ehrlichia chaffeensis (E. chaffeensis), Neorickettsia risticii, Ehrlichia ewingii (E. ewingii), Anaplasma phagocytophilia (A. phagocytophilia), and Anaplasma platys (A. platys).1 Historically, E. canis has been considered the primary cause of canine monocytic ehrlichiosis, a disease of worldwide distribution.2 In the natural disease, most affected dogs usually recover spontaneously or, after the acute infection, they enter into a subclinical phase that lasts as long as 5 years. An unknown percentage of subclinically infected dogs eventually develops a chronic disease phase that is characterized by hematological abnormalities such as thrombocytopenia, pancytopenia, and hyperglobulinemia.23 Bone marrow hypoplasia with an associated peripheral blood pancytopenia contributes to the terminal stage of chronic E. canis infections. Dogs may die from bacterial septicemia, severe bleeding, or both.2
Strain pathogenicity, concurrent infections, host immune status, or other, as yet-undetermined factors may affect the spectrum and severity of the clinical and pathological features of E. canis infection in dogs.4 Despite the considerable advances in understanding the pathogenesis of E. canis, as well as the improved recognition of the clinical and pathological features of the acute and subclinical phases of monocytic ehrlichiosis, knowledge of the chronic disease phase is still limited, in part because of the lack of a suitable experimental model.4
The purpose of this retrospective study was to thoroughly evaluate the clinicopathological findings and disease outcomes in 19 dogs that met clinical and clinicopathological criteria for the chronic phase of E. canis infection. Attempts were made to investigate the potential effects of concurrent infections with other vector-borne pathogens on the clinical features and progression of the disease.
Materials and Methods
Study Population
A total of 19 dogs admitted between April 1998 and January 2000 to the Clinic of Companion Animal Medicine, School of Veterinary Medicine, Aristotle’s University of Thessaloniki, Greece were included in this study. Each dog fulfilled all the following four criteria: 1) peripheral blood bicytopenia or pancytopenia (packed cell volume [PCV], <36%; white blood cell [WBC] count, <6.0 × 103/μL; platelet count, <175 × 103/μL); 2) seroreactivity to E. canis antigens via immunofluorescent antibody (IFA) assay; 3) bone marrow hypoplasia confirmed by cytopathology; and 4) no historical evidence of drug or radiation exposure for at least 6 months prior to admission. The medical records of the dogs were reviewed for signalment, seasonal distribution, other historical data, and presence of clinical and other clinicopathological abnormalities. Bone marrow and buffy-coat cytopathological smears were reviewed in an effort to visualize various vector-borne pathogens and to determine megathrombocyte percentages and megakaryocyte counts. The response to specific and supportive treatments and disease outcomes were also analyzed.
Laboratory Sample Collection
A complete blood count (CBC),a including a peripheral blood smear WBC differential was performed in all 19 dogs. An equal number (n=19) of serum samples were obtained for biochemical analysis and serological testing. Buffy-coat, lymph node, and iliac crest bone marrow aspiration smears were prepared and stained with Giemsa. Aliquots (0.5 mL) of bone marrow were collected in ethylenediaminetetraacetic acid (EDTA)-coated tubes and stored frozen (− 20°C) for future evaluation. Aliquots of serum (0.5 mL) and bone marrow samples (0.2 mL) from all 19 dogs were ice-packed and shipped with a transit time of <1 week, to the Intracellular Pathogens Laboratory, College of Veterinary Medicine, North Carolina State University (NCSU), in Raleigh, North Carolina.
Cytopathological Examinations
A maximum of 1000 oil-immersion fields (OIF) on the Giemsa-stained buffy-coat, lymph node, and bone marrow smears, were screened for morulae indicative of monocytic, granulocytic, or thrombocytic Ehrlichia species, Babesia canis (B. canis) merozoites, Hepatozoon canis (H. canis) gamonts, and Leishmania infantum (L. infantum) amastigotes. In three dogs, lung imprint smears that were made during postmortem examination were also examined for E. canis morulae. The percentage of megathrombocytes (size of platelets was ≥ 5 μm as measured by an eyepiece micrometer) present was estimated following serial evaluation of 100 platelets (× 1000 magnification) in buffy-coat smears. In bone marrow smears obtained from all 19 dogs, the cellularity of the erythroid, myeloid, and megakaryocytic series, as well as the total number of megakaryocytes per smear were also assessed. In the latter case, megakaryocytes were counted by screening the whole bone marrow smear (x100 magnification).
Serology
All dogs were tested using the IFA technique for seroreactivity to E. canis (Florida strain) antigens, Bartonella vinsonii (B. vinsonii) subspp. berkhoffii antigens, and Rickettsia conorii (R. conorii) (Israeli-2 strain) antigens.5–7 Reciprocal titers of ≥64 were considered indicative of prior exposure or active infection. Eleven dogs were also screened for antibodies against L. infantum with an ELISA test kit.b To confirm IFA specificity, the serum from 13 of the 19 dogs was further tested for E. canis antibodies by Western immunoblotting.5
Polymerase Chain Reaction Assays
Extraction of deoxyribonucleic acid (DNA) from 200-μL EDTA-treated bone marrow samples was performed according to the manufacturers’ instructions.c An Ehrlichia-genus polymerase chain reaction (PCR) assay was performed on each sample. For those samples in which an Ehrlichia amplicon was obtained, species-specific PCR primers were used to amplify E. canis, E. chaffeensis, E. ewingii, A. phagocytophilia, or A. platys.8 Rickettsia, Babesia, and Leishmania-genus PCR amplifications were also performed in the Intracellular Pathogens Laboratory, NCSU, as previously described.89 Nested PCR amplification of E. canis 16S rDNA, using a previously described PCR assay, was also carried out in EDTA-treated bone marrow aliquots in the Laboratory of Microbiology and Infectious Diseases, School of Veterinary Medicine, Aristotle’s University of Thessaloniki in Greece.310
Statistical Analysis
Pearson’s chi-square test was used to compare the proportion of German shepherd dogs in the study population with that in the general hospital population during the study period. This test was also applied to compare the incidence of the disease between the warm (April through October) and cold (November through March) months of the year. All tests were done with a commercially available statistical software program,d and results were evaluated at the 5% level of significance.
Results
Clinical Data
Signalment, historical data, clinical abnormalities, treatment, and outcome are summarized in Table 1. The study included 12 male and seven female dogs, with body weights ranging from 3.7 to 50 kg (median, 20 kg). The dogs ranged in age from 5 months to 10 years (median, 1.5 years). The study population consisted of German shepherd dogs (n=8), mixed-breed dogs (n=7), collies (n=2), a Brittany spaniel (n=1), and a Great Dane (n=1). Forty-two percent (8/19) of the affected dogs were German shepherd dogs compared to 16.7% (270/1650) of the general hospital population during the same time period (P=0.003). Fifty-eight percent (11/19) of the dogs were admitted to the hospital during the warm months of the year (April through October) compared to 42% (8/19) admitted during the cold months (November through March); there was no significant seasonal difference in regard to the months of admission (P=0.18). Historical abnormalities, ranging in duration from 1 day to 4 months (median, 1 week), included depression (n=16), anorexia (n=15), bleeding tendencies (n=10; e.g., bilateral epistaxis in six dogs), weight loss (n=6), anterior uveitis (n=3), polyuria/polydipsia (n=2), hind-limb edema (n=1), and vomiting (n=1).
Physical examination abnormalities are summarized in the Figure and Table 1. Bleeding diatheses were present in all 19 dogs and were characterized by cutaneous and mucosal petechiae/ecchymoses (n=16), moderate to severe bilateral epistaxis (n=4), melena (n=3), subcutaneous (n=2) and sublingual (n=2) hematomas, gingival bleeding (n=1), hematochezia (n=1), hematuria (n=1), and prolonged bleeding from venipuncture sites (n=1).
Ten dogs were treated with doxycycline hyclatee (5 mg/kg q 12 hours, per os [PO], for 4 weeks) and imidocarb dipropionatef (two doses of 5 mg/kg subcutaneously [SC], 14 days apart) [Table 1]. Supportive measures were instituted in eight dogs with severe anemia (whole blood transfusions, anabolic steroids, iron supplements), and/or severe neutropenia (bactericidal antibiotics) to combat the potential of life-threatening sepsis. Because of the historically poor prognosis associated with E. canis-induced bone marrow hypoplasia, euthanasia was elected for three dogs. The owners of six dogs declined treatment. Four of the untreated dogs were lost to follow-up, whereas two were reported dead within 3 days of admission. All remaining 10 dogs succumbed to disease complications within 3 weeks, despite the initiation of specific and supportive treatments [Table 1].
Laboratory Abnormalities
Hematological abnormalities included anemia (PCV, <37%) in 19/19 (100%) dogs; thrombocytopenia (<175 × 103/μL) in 19/19 (100%) cases; lymphopenia (<1.0 × 103/μL) in 18/19 (95%) dogs; eosinopenia (<0.1 × 103/μL) in 18/19 (95%) dogs; neutropenia (<3.0 × 103/μL) in 17/19 (90%) dogs; pancytopenia in 17/19 (90%) cases; and monocytopenia (<0.15 × 103/μL) in 10/19 (53%) dogs. In 13 dogs, platelet counts ranged from 5 to 26 × 103/μL (median, 10 × 103/μL; reference interval, 175 to 500 × 103/μL), while in the remaining six dogs, the platelet count ranged from 40 to 96 × 103/μL (median, 50.5 × 103/μL) [Table 2]. In 18/18 dogs, the anemia was nonregenerative, based upon a reticulocyte production index (RPI) of <2 (reference range, >2). An agglutination test (in-saline slide; i.e., Coombs’ test) was positive in 2/14 (14%) dogs. Serum biochemical abnormalities included elevated ala-nine aminotransferase (ALT, 13/18 [72%]; range, 40 to 1800 U/L; median, 60; reference value, ≤35 U/L) and alkaline phosphatase (ALP, 6/17 [35%]; range, 300 to 2400 U/L; median, 716 U/L; reference value, ≤210 U/L), hypoalbuminemia (10/19 [53%]; range, 1.4 to 2.2 g/dL; median, 2.0 g/dL; reference value, ≥2.5 g/dL), increased urea nitrogen concentration (7/18 [38%]; range, 38 to 96 mg/dL; median, 43 mg/dL; reference value, ≤36 mg/dL), increased creatinine (2/18 [11%]; range, 2.2 to 3.1 mg/dL; mean, 2.65 mg/dL; reference value, ≤1.4 mg/dL), hyperglobulinemia (6/19 [32%]; range, 4.8 to 7.2 g/dL; median, 5.75 g/dL; reference value, ≤4.5 g/dL), hyperproteinemia (4/19 [21%]; ≤8 g/dL), hypoproteinemia (3/19 [16%]; range, 5.0 to 5.6 g/dL; median, 5.3 g/dL; reference value, ≥6 g/dL), and increased total bilirubin (2/6 [33%]; range, 3.0 to 4.0 mg/dL; mean, 3.5 mg/dL; reference value, ≤0.8 mg/dL) concentrations. Urinalyses revealed glomerular proteinuria (6/19; 31.5%) as indicated by a positive Heller (nitric acid) reaction in the context of inactive sediment; microscopic hematuria (1/18; 5.5%); and low specific gravity (<1020; 1/18 [5.5%]).
Cytopathology
Monocytic Ehrlichia spp. morulae were found in 2/19 (10.5%) dogs (case nos. 1, 17); the buffy-coat smears from both of these dogs contained one morula per 1000 OIF. In addition, the bone marrow aspirate from case no. 1 contained two morulae per 1000 OIF and a lymph node aspirate from case no. 17 contained one morula per 1000 OIF. In case no. 8, morulae were found in the lung imprints, but not in buffy-coat, lymph node, or bone marrow samples. No tick-transmitted pathogens were visualized in samples obtained from most of the dogs (n=16). Leishmania infantum amastigotes were detected in lymph node and bone marrow smears from only one dog (case no. 12).
The percentage of megathrombocytes in buffy-coat smears ranged from 0% to 1%, and there were up to five megakaryocytes per bone marrow smear. Hypoplasia involving the erythroid, myeloid, and megakaryocytic cell lines was detected in all 19 dogs. Plasma cells ranged from 0% to 20% (median, 4%), with 13/19 (68%) dogs having bone marrow plasmacytosis (plasma cells, >2%).
Serology
Immunofluorescent antibody or ELISA serology indicated exposure to at least one vector-borne pathogen in all dogs [Table 3]. Reciprocal titers to E. canis antigens ranged from 128 to >8192 (median, >8192). Seroreactivity to R. conorii antigens was found in 10/19 (52%) dogs, with reciprocal titers ranging from 64 to 512 (median, 128). One dog was seroreactive to B. vinsonii (berkhoffii) antigens (case no. 14), and another (case no. 12) was ELISA-positive for L. infantum antibodies. In 12/13 (93%) dogs tested, there was agreement between the E. canis IFA and E. canis Western immunoblot results.9 Although negative by Western blot criteria, case no. 8 (which was presented for evaluation of depression and epistaxis) had a PCV of 4.8%, a neutrophil count of 0.4 × 103/μL (reference interval, 3 to 11 × 103/μL), a platelet count of 50 × 103/μL (reference interval, 175 to 500 × 103/μL), a reciprocal E. canis titer of 128, and was PCR-positive in both laboratories.
Polymerase Chain Reaction Assays
Polymerase chain reaction assays (performed at NCSU) resulted in the amplification of Ehrlichia genus DNA from 13/19 bone marrow samples (case nos. 1, 2, 4–9, 12, 13, 17–19). Following individual amplification with each set of species-specific primers, all 13 bone marrow samples contained E. canis. In case nos. 7 and 17, A. platys and A. phagocytophilia were also amplified, respectively. Polymerase chain reaction failed to amplify DNA from other Ehrlichia spp. Neither Rickettsia or Babesia genus DNA was amplified from the bone marrow of any dog. Amplicons of L. infantum were obtained in one sample (case no. 12), which was from the only dog with detectable L. infantum antibodies on ELISA. Based on PCR amplification at Aristotle’s University of Thessaloniki, only 5/19 (26.3%) bone marrow samples contained E. canis DNA (case nos. 1, 3, 4, 8, 16).
Discussion
Because the timing of E. canis transmission in natural infections is unknown, in contrast to experimental infections, distinctions between the acute, subclinical, and chronic phases of ehrlichiosis are more difficult to define.11 However, it is generally accepted that E. canis-seropositive dogs, with bi-or pancytopenia associated with bone marrow hypoplasia, satisfy the diagnostic criteria for the chronic phase of the disease.2 In this regard, the combination of megathrombocyte percentages in buffy-coat smears and megakaryocyte counts in bone marrow smears may be quite useful for staging the phase of E. canis infections. In a group of 50 dogs from Greece with presumably acute natural E. canis infections, megathrombocytes in buffy-coat smears ranged from 2% to 20% (median, 9%), and megakaryocytes in bone marrow ranged from 20 to 400 (median, 101) per smear, further supporting the acute phase.12 Acute infection in those dogs was defined by compatible clinical and/or clinicopathological evidence of disease, seroreactivity to E. canis antigens, detection of E. canis DNA by bone marrow PCR amplification, and complete clinical recovery after a 4-week treatment with oral doxycycline.12 In the present study, both megathrombocytes and megakaryocytes were markedly reduced, strongly indicating the presence of chronic disease. Limited information from clinically healthy dogs has shown that normal megathrombocyte counts range from 0.8% to 1.2% and normal megakaryocyte counts range from 10 to 60.1314
In the study reported here, other causes of severe bone marrow suppression, including drugs, radiation exposure, and endogenous hyperestrogenism, were excluded. Myelophthisis and myelodysplasia were also ruled out by bone marrow cytopathology. Severe bone marrow suppression further substantiated a chronic disease course, particularly when considered in conjunction with the high E. canis antibody titers, which were detected in most (18/19) of the dogs. Bone marrow plasmacytosis, which was found in 13/19 (68%) dogs, presumably reflected a chronic, exaggerated, and ineffective humoral immune response to E. canis.4
Interestingly, case no. 8, which was PCR-positive for E. canis DNA in both laboratories, had a very low reciprocal IFA titer for E. canis. A Western immunoblot did not confirm the IFA results, suggesting that the severe bone marrow changes in this dog may have developed during the acute phase of infection. Alternatively, IFA titers have been reported to drop sharply in severely pancytopenic or terminally ill dogs, and case no. 8 was both pancytopenic and severely ill.15
Unlike acute E. canis infections, where the majority of the dogs develop disease manifestations during the warm period of the year, no difference in seasonality was documented in this study.16 This may be explained by the extremely variable duration of the subclinical phase of the disease, as well as the current lack of understanding as to what factors cause a subclinically infected dog to develop overt clinical signs.14
The typical clinical manifestations that have been associated with chronic E. canis infection were seen more frequently in the dogs of this report than in similar studies.1116–18 This difference may reflect the strict entry criteria used in this study. It is of interest that the cases reported here were identified within a relatively short time span, potentially suggesting a unique feature of E. canis infections in dogs in Greece. In contrast with previous reports, all 19 dogs developed moderate to severe bleeding tendencies, with petechiation and ecchymoses being the most common. However, epistaxis was seen less often when compared to reports from other investigators.1618 In six cases, the hemorrhagic lesions could not be attributed to thrombocytopenia alone, as the platelet counts (range, 40 to 96 × 103/μL) were not considered low enough to result in spontaneous hemorrhage. One explanation for the clinicopathological diversity observed among these dogs would be the concurrent presence of E. canis-induced thrombocytopathia, secondary immunemediated vasculitis, or both.4 Although the ulcerative stomatitis seen in three dogs could have been a nonspecific, clinical manifestation of many pathological conditions, it was also consistent with a vasculitis or E. canis-induced immunosuppression.4
The prevalence of pancytopenia was higher in this study than in similar reports, but comparison is hindered by differences in case selection criteria.1116 Transient pancytopenia can accompany acute E. canis infection, but it is usually associated with bone marrow hypercellularity.212 In contrast, the chronic phase of the disease is associated with severe bone marrow hypocellularity, as observed in the pancytopenic dogs of this report. Clinical experience among veterinarians in Greece indicates that many dogs with chronic E. canis infections have severe pancytopenia, as was also the case with the United States (USA) military dogs during the Vietnam war.19 This is in contrast to the experience reported recently from the USA, and may possibly be attributed to variability in the pathogenicity of E. canis strains in different geographical areas or to perhaps other, as yet undefined factors.11
Hypoalbuminemia, which was documented in 10/19 (53%) dogs, is consistent with E. canis infection.1120 In this study, hypoalbuminemia was attributed to blood loss (10/10), protein-losing nephropathy (5/10), potential liver dysfunction, and hyperglobulinemia-related down-regulation of albumin synthesis.4 Despite the presumed chronic nature of the infection in these dogs, hyperglobulinemia was less frequently encountered (32%) than in previous reports.1617 This may have resulted from decreased antibody production from profound pancytopenia.20 Mildly elevated serum ALT and ALP activities (72% and 35% of the dogs, respectively) may have arisen from anemic hypoxia, intralobular necrosis secondary to hemorrhage, or secondary bacterial sepsis.411 Unfortunately, a lack of blood cultures makes the latter option quite speculative. The pathogenesis of glomerular proteinuria in dogs chronically infected with E. canis is poorly understood and was found in six (31.5%) dogs in this report. Glomerular proteinuria has been documented following acute E. canis infection and is known to arise from an immune-mediated glomerulopathy.911 Interestingly, microscopic hematuria was observed in only one dog, despite the overt bleeding diatheses observed in all 19 dogs.
At least 12 (64%) dogs died as a direct consequence of pancytopenia from either hemorrhage or presumed sepsis. Specific treatment with doxycycline and imidocarb dipropionate, as well as vigorous supportive therapy, were of no benefit in 10 dogs (all died within 3 weeks). In general, it is believed that leukopenia, pancytopenia, and a bleeding tendency herald a poor prognosis.16 Apart from financial issues, the poor prognosis associated with E. canis infections is the main reason why the owners of six dogs refused any kind of treatment. In Greece, it has long been known that large-breed dogs with E. canis-associated pancytopenia usually succumb to the disease despite all treatment efforts. In the present study, the mortality rate may have been higher partly because German shepherd dogs were over-represented. This breed is reported to be the most likely to develop severe disease manifestations, such as pancytopenia.16 In contrast, E. canis-infected small-breed dogs in Greece (which can present with identical clinical and hematological abnormalities, including pancytopenia) have, in the authors’ experience, a good to excellent long-term prognosis. Most infected small-breed dogs eventually recover despite a prolonged clinical course and the need for repeated treatments. During the recovery period, neutrophils are the first cell line to return to normalcy, followed by platelets and red blood cells.
The role of other vector-borne infections on the clinical course, disease severity, and treatment outcome in dogs infected with E. canis has yet to be systematically investigated. Although it is likely that coinfection influences the severity of clinical disease, descriptive studies are not adequate to determine if there is an association between coinfection and mortality rate, particularly when studying small numbers of dogs. As coinfections with L. infantum, A. phagocytophilia, or A. platys were each documented in only three dogs, it is possible that the Greek E. canis strains are more myelosuppressive than North American strains.
Tick-borne pathogens in Greece have been well documented, based on visualization of organisms in buffy-coat smears (B. canis, H. canis, A. platys, E. canis), by serological evidence (E. canis, A. platys, A. phagocytophilia, Rickettsia spp.), and more recently on the basis of PCR testing (E. canis).12 The A. phagocytophilia infection in case no. 17 represents the first molecularly documented case of granulocytic ehrlichiosis in Greece. The same holds true for the A. platys in case no. 7. Since recent epidemiological studies on the tick population in Greece failed to recover Amblyomma americanum, the vector of E. ewingii, it is quite possible that the neutrophilic morulae observed in buffy-coat smears from dogs in the past by some of the authors (Mylonakis, Koutinas, Kontos) were infections with A. phagocytophilia.21 Seroreactivity to R. conorii antigens, also documented in this study (10/19; 52%), clearly indicates canine exposure to rickettsial antigens of the spotted fever group. There has been scant information regarding the clinical implications of R. conorii infection in the dog, which can be transmitted to humans in the Mediterranean region by Rhipicephalus sanguineus.22 Failure to detect rickettsial DNA in the dogs of this report seems to decrease the possibility of chronic spotted fever group rickettsial infection, as occurs with canine Rocky Mountain spotted fever in North America. Only one dog was seroreactive to B. vinsonii (berkhoffii) antigens, in contrast to a much higher seroprevalence found in E. canis seropositive dog populations in the USA and Thailand.68
In buffy-coat, lymph node, and bone marrow cytopathology, Ehrlichia morulae were detected in only two cases, thus rendering its diagnostic value low in the chronic phase of the disease. This is in marked contrast to acute ehrlichiosis, where sensitivities as high as 74% have recently been documented for morulae detection in Greek dogs.12 Although superior to cytopathology, the sensitivity of PCR for documenting E. canis infection in pancytopenic dogs with chronic disease needs further optimization. In this study, for example, there were considerable differences in the number of dogs found positive on two different PCR assays in two different laboratories, despite the fact that laboratories both received bone marrow aliquots from the same aspiration and from dogs that had not been previously treated. In particular, case nos. 3 and 16 were found positive with the Aristotle’s University of Thessaloniki PCR but negative with the NCSU PCR, while the opposite was true for case nos. 2, 5–7, 9, 12, 13, 17–19. Nevertheless, in 4/19 dogs (case nos. 10, 11, 14, 15), Ehrlichia spp. DNA was not amplified by either of the methods, and in three other dogs (case nos. 1, 4, 8) the result was positive in both laboratories. The use of different primers, different DNA extraction techniques, as well as different bone marrow aliquots with each PCR, may account for the discordant results. Another explanation could be the low numbers of infected leukocytes owing to the severe peripheral blood leukopenia and bone marrow hypocellularity. Future studies are required to determine if tissues other than bone marrow (e.g., lymph nodes, spleen, lung) are more suitable for PCR testing in the chronic phase of E. canis infection, as has been suggested for the subclinical phase.3
Conclusion
Based on this study, canine chronic ehrlichiosis caused by Ehrlichia canis was associated with severe bone marrow hypoplasia, severe anemia, bleeding diatheses, and a high mortality rate. Coinfection with vector-borne agents did not appear to alter the course of disease, but further systematic investigation is required. Results of bone marrow PCR seemed to be less sensitive in these chronic cases, as compared to the acute phase of E. canis infections in dogs.
QBC Vet Autoread; IDEXX Laboratories, United Kingdom
Snap Leishmania Test Kit; IDEXX, United Kingdom
DNA blood mini-kit; Qiagen, Valencia, CA
Stat Xact for Windows version 4.1; Cytel Software Corporation, Cambridge, MA
Ronaxan; Merial, Lyon, France
Imizol; Schering-Plough, Animal Health Corporation
Acknowledgments
Research funding for the Vector Borne Diseases Diagnostic Laboratory was provided by the Hellenic Ministry of Health (Interreg-2) and in part by the state of North Carolina.
Citation: Journal of the American Animal Hospital Association 40, 3; 10.5326/0400174
Clinical signs and pathological conditions in 19 natural cases of chronic canine ehrlichiosis.
