Parts of this paper were presented as a poster at the 2007 ACVIM Congress in Seattle, USA (title: “Clinical features of 15 dogs naturally infected with Anaplasma phagocytophilum”) and at the 2007 DVG Congress in Berlin, Germany (title: “Canine granulozytäre Anaplasmose bei 16 natürlich infizierten Hunden”).
Corresponding author: Prof. Dr Barbara Kohn, Clinic for Small Animals, Faculty of Veterinary Medicine, Free University of Berlin, Oertzenweg 19b, D-14163 Berlin, Germany; e-mail: firstname.lastname@example.org.
Background: Anaplasma phagocytophilum, the causative agent of canine granulocytic anaplasmosis (CGA), is a Gram-negative intracellular organism transmitted by ixodid ticks. Thus far, only a few clinical studies evaluating dogs with CGA have been published.
Objectives: Evaluation of dogs naturally infected with A. phagocytophilum in which known co-infections were excluded.
Animals: Eighteen dogs with CGA.
Methods: Prospective study. The diagnosis of CGA was based on a positive PCR test result; dogs with co-infections were excluded. History, clinical findings, CBC, clinical biochemistry, infectious disease screening, diagnostic imaging, and the course of disease were evaluated.
Results: CGA was diagnosed based on a positive PCR test for A. phagocytophilum; 10 dogs also had morulae in neutrophils. Six of 18 dogs were seronegative to A. phagocytophilum, the others were seropositive. All dogs were acutely ill. The most common clinical findings were lethargy, inappetence, fever, and splenomegaly. Abnormal laboratory results included thrombocytopenia, anemia, lymphopenia, hypoalbuminemia, and abnormally high plasma alkaline phosphatase activity. In 6 of 10 dogs tested, the platelet-bound antibody test was positive; Coombs' test was negative in 9 dogs. All dogs were treated with doxycycline and recovered. PCR testing as well as blood smear analysis for morulae were negative in 14 tested dogs 2–8 weeks after beginning treatment.
Conclusions and Clinical Importance: Clinical findings in dogs with CGA were nonspecific. Positive platelet-bound antibody test results suggest immune-mediated platelet destruction as an important pathogenic mechanism. With correct diagnosis and treatment, prognosis is good.
Anaplasma phagocytophilum (A. phagocytophilum) is the new name of the species previously known as Ehrlichia phagocytophila, Ehrlichia equi, and human granulocytic ehrlichiosis agent. A. phagocytophilum is a Gram-negative, obligate intracellular bacterium of the family Anaplasmataceae. It is found in cytoplasmic vacuoles, forming distinctive inclusion bodies, so-called morulae. A. phagocytophilum mostly afflicts neutrophilic and rarely eosinophilic granulocytes.1 It is transmitted by ticks of the genus Ixodes: Ixodes ricinus in Europe,2 and Ixodes scapularis and Ixodes pacificus in the United States.3 Further pathways for transmission are perinatal transmission described in humans,4 transplacentary infection in cattle,5 and transmission via infected blood.6A. phagocytophilum is the causative agent of diseases such as canine, feline, equine, and human granulocytic anaplasmosis, and of tick-borne fever in ruminants.7
The pathogenesis of granulocytic anaplasmosis is not yet fully understood. Transmission of A. phagocytophilum from the tick to mammals is accomplished approximately 24–48 hours after the tick bite.8,9 It spreads within the mammal via either blood or lymph.10 The incubation period varies from 1 to 2 weeks.6 After adhesion via P-selectin, which is abundant on the surface of the bacteria, they enter the neutrophils by endocytosis.11 After the bacteria are incorporated into phagosomes, they multiply by binary fission forming the morulae.12 Experimental studies have shown that seroconversion in dogs may occur as soon as 2–5 days after the appearance of morulae in peripheral blood.6 Rupturing of the phagosome and cell membrane of the host cells results in a release of A. phagocytophilum, thus facilitating the infection of other cells and consequently, infection of multiple organs.10,11 The bacterium can activate pathogenic mechanisms preventing its destruction by cells of the immune system.13 Malfunction of infected phagocytes interferes with immune defense mechanisms, which may facilitate secondary infections.10
Infection with A. phagocytophilum can be detected either directly by the presence of morulae in granulocytes in the peripheral blood or by PCR, or indirectly by measuring antibody titer.11
Thus far, only a few experimental studies,6,14,15,a several case reports,16–23 and a few clinical studies24–28 describing dogs suffering from canine granulocytic anaplasmosis (CGA) have been published. There are differences among these clinical studies mainly with regard to exclusion of other infectious agents and the extent of hematologic and biochemical examinations.
The objective of this study was to describe the clinical signs, laboratory abnormalities, diagnosis, treatment, and course of disease in naturally infected dogs living in areas with high I. ricinus populations and to exclude, to the extent possible, potential co-infections.
Material and Methods
Dogs presented at the Clinic for Small Animals, Free University of Berlin, were tested for anaplasmosis if they had clinical signs described in the literature in association with CGA (eg, lethargy, fever, lameness, joint pain, reluctance to move, weakness, splenomegaly, and meningitis) or if unexplained thrombocytopenia, anemia, or leukopenia were present.10,11,24,26,28 The cases were collected over a time period of 26 months (June 2005 to July 2007).
During the study period, we tested 258 sick dogs; 22 dogs were PCR positive. In 4 of these 22 dogs, co-infections were detected: Ehrlichia canis (2), Leishmania infantum (1), and septic arthritis with methicillin-resistant Staphylococcus aureus (1).
Signalment, history, findings of clinical and laboratory examinations and tests for canine vector-borne infectious agents, treatment, and course of disease were evaluated in 18 dogs with positive A. phagocytophilum PCR test results and no known co-infections identified. In all 18 dogs, the following diagnostic information was available: CBC,b clinical biochemistry using lithium heparin plasma,c manual differential blood cell count, and radiography of the thorax and abdomen. In addition, ultrasonography of the abdomen was performed in 15 dogs. Moreover, coagulation parameters (n = 10) such as activated partial thromboplastin time (aPTT)d and prothrombin time (PT)e were determined.f Urinalysis (n = 8), direct differentiated Coombs' test (n = 9),29 and a platelet-bound antibody test (n = 10)30 (Immunology Unit, University of Veterinary Medicine, Hanover, Germany) also were performed. In 1 dog with signs of polyarthritis, several joints were tapped under general anesthesia. Synovial analysis included evaluation of macroscopic appearance, viscosity, nucleated cell count, cytology, protein concentration, and aerobic and anaerobic microbiologic culture.31
All 18 dogs were tested for E. canis, Borrelia burgdorferi, Babesia canis, Dirofilaria immitis, Bartonella spp., and Mycoplasma haemocanis; 7 dogs in addition were tested for L. infantum.
Examinations for detection of A. phagocytophilum using direct (real-time-PCR targeting the msp2 gene of A. phagocytophilum and morulae in neutrophils, Fig 1) and indirect immunofluorescent antibody titer (IFAT) methods were performed.32,33
In addition, antibody titers against E. canis, B. canis, and L. infantum were determined by IFAT; PCR tests were used to test for Mycoplasma haemocanis and Bartonella spp.32–34 The SNAP-4Dxg was used to test for B. burgdorferi. This procedure also tests for A. phagocytophilum, E. canis, and D. immitis.
For statistical analysis, a nonsymmetrical distribution of the parameters was assumed. Thus, evaluation was performed by nonparametric statistical analysis giving minimal, maximal, and median values. In addition, mean values and standard deviation were calculated. SPSS 14.0 for Windowsh was used as statistical software.
All 18 dogs showed signs of an acute infectious disease. The disease was diagnosed mainly between April and September (17 dogs), and only once in November.
The dogs were presented because of lethargy (17), inappetence (15), decreased activity (12), fever (4), polydipsia and polyuria (3), diarrhea (3), vomiting (2), difficulty getting up (1), or collapse (1). Tick infestation had been noticed by the owner in 14 dogs. Five of the 18 dogs had either been regularly treated with fipronil or permethrin once a month, 11 were not (7/11) or only irregularly (4/11) treated with different ectoparasiticides. For 2 dogs, insufficient information was available. Eleven dogs had never left their living area whereas 6 dogs had traveled to different European countries (eg, Denmark, Croatia, Poland, and Slovenia). For 1 dog, no travel information was obtained.
Six dogs were mixed breed, the other 12 dogs were of 9 different breeds. Thirteen dogs were male (6 neutered); 5 were female (1 neutered). The age of the dogs ranged from 2 to 13 years (median, 8 years).
Abnormal findings were fever (11) with rectal temperatures ranging from 39.2 to 41.0 °C (median, 40.0 °C), pale mucous membranes (5), tense abdomen (5), diarrhea (3), vomiting (2), lameness (2), petechiae (2), epistaxis (1), and melena (1). Lateral recumbency, ventricular tachycardia, and tachypnea were noticed in 1 dog each; another dog collapsed during examination. Radiologic examination revealed splenomegaly in 17 dogs and hepatosplenomegaly in 1 dog. The spleen was homogeneous in the 15 dogs examined by ultrasonography.
Table 1. CBC results in 18 dogs with canine granulocytic anaplasmosis.
Plant (× 103/μL)
RBC (× 106/μL)
WBC (× 103/μL)
Table 2. Plasma biochemistry results in 18 dogs with canine granulocytic anaplasmosis.
Platelet counts in 16 thrombocytopenic dogs ranged from 5.2 to 164 × 103/μL (median, 32.4). Only 3 dogs with platelet counts <45 × 103/μL displayed hemorrhage. The hematocrit (Hct) in 11 anemic dogs ranged from 19 to 39% (median, 32%), 2 of these dogs were severely anemic (Hct < 20%). For 8 of 9 anemic dogs, a reticulocyte count was performed. Anemia was regenerative at initial presentation (reticulocytes >60,000/μL) in 3 dogs only. Of 4 anemic dogs, erythrocytic agglutination was present in 2, and did not persist after saline washing. Coagulation profiles were performed in 10 dogs. The aPTT was mildly prolonged in 6 dogs and the PT was prolonged in 3 dogs, respectively. Disseminated intravascular coagulation was suspected in 2 dogs with thrombocytopenia and mildly prolonged aPTT and PT.
Urinalysis was performed in 8 dogs. Specific gravity ranged from 1.008 to 1.049 (median, 1.023). The urine-protein/creatinine ratio (U-P/C) was measured in 3 dogs (0.15, 0.33, and 0.81; reference range, <0.5). Examination of urine by standard urine dipstick methodology (n = 8) revealed mild to moderate proteinuria in 7 dogs, glucosuria in 1 dog, bilirubinuria in 4 dogs, blood in 7 dogs, and erythrocytes in 5 dogs. Urine sediment evaluation revealed casts in 4 and epithelial cells in 6 dogs.
A platelet-bound antibody test was performed for 10 dogs: 9 dogs had thrombocytopenia and 1 dog had a platelet count within the reference range. For 6 of the thrombocytopenic dogs, the test results were positive. A direct Coombs' test was performed in 9 dogs with Hct ranging from 19 to 40 % (median, 33%). All dogs tested negative.
The synovial analysis of several joints of 1 dog indicated an increased volume; the synovial fluid was turbid, had severely decreased viscosity, and nucleated cell count ranging from 15,800 to 21,000/μL. Over 90% of these cells were neutrophilic granulocytes. Aerobic and anaerobic cultures were negative.
Examinations for Canine Vector-Borne Infectious Agents
According to the inclusion criteria, PCR tests were positive for all dogs. Morulae were found in neutrophilic granulocytes in 10 dogs. At initial presentation, serum antibody titers tested by IFAT were negative in 6 and positive in 12 dogs (≥1:64). Of the 6 dogs that tested negative with the IFAT, 2 were positive using the 4Dx test. On the other hand, of 12 IFAT-positive dogs, 5 were seronegative with the 4Dx test.
All dogs were tested for E. canis, B. burgdorferi, B. canis, D. immitis, Bartonella spp., and M. haemocanis, and 7 dogs in addition were tested for L. infantum. All test results were negative, except in the case of 1 dog that had a positive antibody titer against B. burgdorferi on 4Dx testing. This dog was seronegative using IFAT, and there were no signs of joint or any other disease indicative of borreliosis. Urinalysis was unremarkable.
Fifteen dogs were treated in the hospital for a time period between 1 and 7 days (median, 3); 3 dogs were treated as outpatients. All dogs received doxycycline (5 mg/kg PO q12h). Doxycycline was given over a time period of 2, 3, and 4 weeks in the case of 3, 9, and 5 dogs, respectively. Six days after beginning treatment, an increase in liver enzyme activities was noticed in 1 dog. Therefore, treatment with doxycycline was discontinued and replaced by imidocarb (6 mg/kg SC once). In addition, all dogs were treated according to their clinical signs (crystalloid infusions, n = 15; ranitidine, n = 12; metamizole, n = 8; and metoclopramide, n = 2). One dog with petechiae and melena, in which immune-mediated thrombocytopenia was suspected, was treated with prednisolone for 7 days (1 mg/kg PO q12h). The prednisolone was gradually tapered and completely discontinued after 2 months. One dog with petechiae, severe thrombocytopenia (12 × 103/μL), and anemia (Hct, 19%) that collapsed during examination was transfused on the day of presentation (7 mL/kg packed red blood cells and 11 mL/kg fresh whole blood). Immune-mediated thrombocytopenia was assumed, and therefore this dog received prednisolone 1 mg/kg PO q24h for 2 weeks; after that, prednisolone was gradually tapered off. Another dog with reactive polyarthritis received prednisolone at a dosage of 1 mg/kg PO q12h, because there was no substantial improvement after 6 days of treatment with carprofen and doxycycline. The dog's lameness improved quickly, and the prednisolone was tapered over 2 weeks.
Course of Disease
In 14 of 18 dogs, evaluations for canine vector-borne infectious agents were repeated after 14 to 56 days (median, 27). At that time, all dogs were without clinical signs of disease. The PCR results were negative for all dogs, and no morulae were found in neutrophils.
Three of 6 dogs with negative antibody titers at initial examination had positive antibody titers after 5–6 weeks; 3 dogs remained seronegative. In 2 seropositive dogs, an increased antibody titer was detected after 6 and 8 weeks, respectively. Four dogs with a high antibody titer at 1st presentation had the same high antibody titer at the second examination after 3–5 weeks. In 2 of the dogs that were seropositive initially, the antibody titers were lower or negative, respectively, 3 weeks after initial presentation.
Repetition of altered laboratory parameters depended on clinical signs and the decision of the attending clinicians. In all dogs with low platelet counts at initial presentation, platelet counts were within or slightly below the reference range after 1–24 days (median, 6).
In 9 of 11 dogs with mild to severe anemia, the Hct was within or slightly below the reference range after 1–28 days (median, 16). Leukocyte counts were within the reference range in 16 of 18 dogs at the end of treatment after 2–4 weeks. In 4 of 10 dogs with hypoalbuminemia, serum albumin concentrations were within the reference range after 8–21 days (median, 17); 6 dogs remained slightly hypoalbuminemic after 4–121 days (median, 19.5).
Eighteen dogs with CGA were diagnosed between June 2005 and July 2007. The data suggested a trend toward exposure in certain months. Seventeen dogs (94%) were diagnosed with CGA between April and September. In other studies, seasonality was described as well. However, the months varied, which may be because of the different time periods during which the vectors are active or climatic differences depending on the various geographical locations.24–26,28
As described in previous studies, all dogs were diagnosed during the acute stage of disease. Various studies have presented different results regarding tick exposure observed by the owners. For example, tick infestation was not described for any of the dogs examined by Poitout et al.28 In a Swedish study, tick exposure was described for 13 of 14 dogs.26 In our study, 80% of the owners had observed infestation with ticks.
Infestation of I. ricinus with A. phagocytophilum as a prerequisite for endemic occurrence of anaplasmosis in Germany was confirmed in Bavaria, Baden-Württemberg,35,36 and Thuringia.37 In different regions of Germany, antibodies against A. phagocytophilum were detected by IFAT in 50.1% (563/1,124), 42.9% (48/112), and 43.4% (108/249) of the tested dogs. None of these dogs had been outside Germany.38,39,i
Adapted from Centers for Disease Control and Prevention, CGA is diagnosed if there was a tick infestation, a blood transfusion, or both in addition to clinical signs or hematologic findings, a positive PCR test result, detection of morulae in neutrophils, or a 4-fold increase of the antibody titer within 4 weeks.10 Season, history, clinical signs, hematologic (especially seeing morulae in neutrophils) and clinical biochemistry data, and serology all can assist in making a strong presumptive diagnosis of A. phagocytophilum infection, but definitive diagnosis relies on PCR analysis.11 In earlier studies, direct detection of the pathogen was accomplished by verification of morulae in neutrophilic granulocytes.24–28 Moreover, in several dogs, a PCR test was performed, which was positive in the majority of patients.25,27,28 In most studies, antibody titers against A. phagocytophilum were determined in addition to direct detection methods; 27 of 38 dogs with CGA tested in different studies were seropositive.25–28 However, a diagnosis of CGA cannot be based on serology, because dogs can be seronegative and clinically ill or vice versa. Moreover, there is no standardization between the serologic tests. In our study, discordance of the IFAT and ELISA 4Dx test results was found in 39% of the 18 dogs. Even PCR-positive dogs can lack clinical and laboratory abnormalities.i Therefore, PCR testing for A. phagocytophilum is strongly recommended in dogs used as blood donors. However, with direct detection methods, only positive results are relevant. The diagnosis of CGA for the dogs of this case series was based on clinical signs, laboratory findings, positive PCR test results, detection of morulae in neutrophils, exclusion of known co-infections, and response to doxycycline treatment.
In earlier studies, 38 dogs with CGA without known co-infections were described.24,26,28 The most common clinical signs were lethargy (68%), fever (66%), anorexia (50%), reluctance to move (39%), and lameness (24%). Less commonly reported were lymphadenopathy (8%), coughing (8%), diarrhea (5%), pale mucous membranes (5%), joint pain (3%), and hemorrhage (3%). The clinical findings were similar to those observed in our study.
Splenomegaly was present in all dogs of this study. Seven dogs infected experimentally were examined pathologically; the spleens of all these dogs were slightly to moderately enlarged, congested, and had a somewhat fleshy consistency. Microscopically, the spleens showed reactive hyperplasia with enlarged activated lymph nodules and increased numbers of macrophages and plasma cells in the red pulp.6
For 22 dogs of earlier studies in which co-infections were excluded, hematologic data were available for comparison with our findings.26,28 The most common abnormalities were thrombocytopenia (87%), lymphopenia (36%), and leukopenia (27%). Less common abnormalities were anemia (13%), neutropenia (13%), and leukocytosis (5%). Our data were similar. Mild to severe anemia was present in 61%, neutrophilia in 44%, and monocytosis in 39% of the dogs.
Mild to moderate thrombocytopenia was detected in 89% of the dogs in our study. Thrombocytopenia is common in humans and animals infected by a wide range of Ehrlichia species.26,40,41 It may be attributed to increased platelet consumption because of disseminated intravascular coagulation, sequestration in an enlarged spleen, immunologically mediated platelet destruction, or production of inhibitory factors.42–44 Complex immunologic processes might contribute to the blood cell disorders. In vitro, A. phagocytophilum-infected neutrophils stimulated increased production of IL-8 and other cytokines (macrophage inflammatory protein [MIP]-1α, MIP-1β, monocyte chemoattractant protein [MCP]-1, and regulated and normal T-cell expressed and secreted [RANTES]). MIP-1α, MIP-1β, and IL-8 inhibit hematopoiesis in vitro.45 In humans infected with A. phagocytophilum, up to 80% of patients had positive antiplatelet antibody test results.43 The identification of platelet-bound antibodies in CGA has been described in 1 dog thus far.19 Because 6 of 10 dogs in our study had positive test results for platelet-bound antibodies, secondary immune-mediated thrombocytopenia appears to be an important pathomechanism for thrombocytopenia in dogs with CGA.
In an experimental study, 9 dogs inoculated with A. phagocytophilum developed mild, normocytic, normochromic anemia resembling anemia of inflammation.14 Anti-erythrocytic antibodies and agglutination of erythrocytes have been detected in the sera of 3 dogs infected with a granulocytic Ehrlichia strain in the United States.27 Hemolysis was a possible pathomechanism in some of our cases because 5 dogs with mild hyperbilirubinemia also had anemia. Three of these dogs with hyperbilirubinemia and anemia, another 5 dogs with anemia, and 1 dog with a normal Hct had negative direct Coombs' test results. Therefore, the importance of immune-mediated erythrocyte destruction in CGA warrants further studies. For 3 dogs with hemorrhage, blood loss contributed to the anemia.
The results of clinical biochemistry were difficult to compare with other studies, because the parameters measured differed substantially.24,26,28 Most reported abnormalities included increased alkaline phosphatase activity (7/15), hyperglobulinemia (6/8), hypophosphatemia (5/8), and hyperproteinemia (4/8). In our study, common abnormal laboratory findings were increased liver enzyme activity, hyperbilirubinemia, hypokalemia as well as hyperproteinemia, hyperglobulinemia, and hypoalbuminemia. During an acute phase reaction, hepatic production of albumin is decreased and that of α- and β-globulins is increased, which might explain the presence of hypoalbuminemia and hyperglobulinemia.45
Thirty-two dogs described in earlier studies with CGA were treated with doxycycline; 2 dogs died, the other dogs recovered as did the dogs in our study.24,26,28 In our study, 1 dog was treated with imidocarb, but efficacy in CGA has not been proven. Currently, only data on the efficacy of imidocarb are available in cases of canine monocytic ehrlichiosis,46 which was not confirmed in a recent publication.47 In the case of 3 dogs in which immune-mediated disease (eg, reactive polyarthritis, secondary immune-mediated thrombocytopenia) was suspected, prednisolone was administered in addition to doxycycline.
In an experimental study, persistent infections were established in 2 dogs using a human isolate of cultivated A. phagocytophilum. Both animals were positive on all PCR assays. As seen with E. canis and A. marginale infections, doxycycline treatment did not eliminate the organism in these infected dogs.a In our study, all dogs that were retested had negative PCR test results using EDTA-anticoagulated blood and no morulae were detected in neutrophils after 2–8 weeks. However, dogs were not evaluated after 8 weeks and thus it is possible that infection in these dogs may have persisted at a level below that required for detection or may have persisted in organs such as bone marrow, liver, or spleen.15
CGA is difficult to diagnose. It should be on the list of differential diagnoses if nonspecific clinical signs such as fever, lethargy, inappetence, splenomegaly, lameness, or laboratory abnormalities such as thrombocytopenia, anemia, lymphopenia, and hypoalbuminemia are present. The diagnosis is based on the exclusion of other infectious diseases, on a positive PCR result, observation of morulae in neutrophils, and clinical improvement after initiation of doxycycline treatment.
aAlleman AR, Chandrashaker R, Beall M, et al. Experimental inoculation of dogs with a human isolate (Ny18) of Anaplasma phagocytophilum and demonstration of persistent infection following doxycycline therapy. J Vet Intern Med 2006;20:763 (abstract)
bCell-Dyn, Abbott GmbH, Wiesbaden, Germany
cKone Lab 30i, Thermo Electron GmbH, Dreieich, Germany
fCoagulometer according to Schnitger & Gross, H. Amelung, Lemgo, Germany
gSNAP-4Dx, IDEXX, Westbrook, ME
hSPSS 14.0 for Windows, Microsoft, Redmond, WA
iPfister K, Galke D, Fumi C, et al. Prevalence of Anaplasma (A.) phagocytophilum infection in dogs in Germany. 21st International Conference of the World Association for the Advancement of Veterinary Parasitology, Ghent 2007 (abstract)
We would like to thank Bayer Vital GmbH, Germany, and the H. Wilhelm Schaumann Stiftung, Germany for financial support, as well as the laboratory staff of the Institute of Parasitology, LMU for performing the examinations for canine vector-borne infectious agents.