Ehrlichia canis is a rickettsial organism transmitted by

Transcription

1 J Vet Intern Med 1999;13: A Retrospective Study of Ehrlichiosis in 62 Dogs from North Carolina and Virginia Johanna R. Frank and Edward B. Breitschwerdt The purpose of this retrospective study is to report the clinical signs, clinicopathological findings, and response to therapy of 62 dogs from North Carolina and Virginia. Ehrlichiosis was diagnosed in all of these dogs, and previous retrospective studies of ehrlichiosis in dogs were used as a basis for comparison. In this study, the clinical signs commonly associated with ehrlichiosis were observed less frequently than in earlier studies, although previously reported laboratory abnormalities were similar. Flow cytometry revealed an inverted CD4 : CD8 ratio in 3 of 4 dogs tested. This finding is suggestive of potential immune dysregulation that could predispose infected dogs to additional disease processes. Concurrent diseases were frequently reported and often contributed to death. Response to therapy was variable, with timely, complete recovery reported in only 27% of dogs; a slow, gradual, but complete recovery in 18% of dogs; an incomplete treatment response in 25% of dogs; and a presumed treatment failure in 16% of dogs. Key words: Ehrlichia canis; Response to therapy; Infectious disease; Flow cytometry. Ehrlichia canis is a rickettsial organism transmitted by the tick vector Rhipicephalus sanguineus. E. canis is prevalent worldwide and occurs at rates proportional to the geographic distribution of the vector. Disease prevalence is especially high in the southern United States. The purpose of this study is to document clinical and laboratory findings as well as the responses to treatment of dogs from North Carolina and Virginia with ehrlichiosis. Retrospective analysis of 62 cases is presented, and the results are compared with the 5 previously reported retrospective studies, including Troy et al, 1 Kuehn and Gaunt, 2 Waddle and Littman, 3 Woody, 4 and Harrus et al. 5 Clinical findings differed among the dogs in this study. The frequency of clinical signs commonly associated with dog ehrlichiosis is less than previously reported; this discrepancy is likely due to the wide variation in cases that were evaluated serologically. Most laboratory abnormalities were comparable to those found in other reports. Response to therapy was variable, and recovery was not as complete as described in previous retrospective studies. Concurrent diseases were common. Material and Methods Between January, 1986, and December, 1994, 62 dogs were diagnosed with canine ehrlichiosis at the North Carolina State University (NCSU) College of Veterinary Medicine. The sole inclusion criteria for this study was a positive indirect immunofluorescent antibody test (IFA) to E. canis antigen ( 1 : 20) in our diagnostic laboratory. Dogs with granulocytic ehrlichiosis were omitted from the study. Granulocytic ehrlichiosis was diagnosed based on the observation of Ehrlichia morulae within peripheral blood or synovial fluid granulocytes. Dogs in this study were admitted to the NCSU College of Veterinary Med- From the Department of Companion Animal and Special Species Medicine, College of Veterinary Medicine, North Carolina State University, Raleigh, NC. Dr. Frank is presently affiliated with Colorado Veterinary Specialists, Littleton, CO. Reprint requests: Johanna Frank, Colorado Veterinary Specialists, 223 West County Line Road, Littleton, CO 80126; jofrank@rmi.net. Received February 10, 1997; Revised December 1998; Accepted December 22, Copyright 1999 by the American College of Veterinary Internal Medicine /99/ /$3.00/0 icine for a variety of presenting complaints, including suspected ehrlichiosis and numerous seemingly unrelated medical and surgical conditions. Geographic distribution of the patient population included the states of North Carolina and Virginia. Medical records were analyzed retrospectively for signalment, history, clinical signs, clinicopathologic data, and response to therapy. The complete blood count, serum biochemical profiles, and urinalysis were analyzed using routine methods in the NCSU Clinical Pathology Laboratory. Antibody reactivity to E. canis antigen was detected using an indirect IFA test. Serum protein electrophoresis (SPE) and flow cytometric analysis of lymphocyte subset populations were performed on a small number of patients. SPE was performed in the Immunology Laboratory at the NCSU College of Veterinary Medicine using standard protocol. Flow cytometry was also performed by the Immunology Laboratory by using a fluorescence-activated cell sorter (FACS) analyzer. E. canis was cultured from the blood in 5 dogs by using a DH82 tissue-culture cell line. Response to therapy was assessed by analyzing follow-up laboratory data and through a telephone interview with the owners of the dogs or the referring veterinarians. The information obtained included initial response to therapy, whether the dog was currently alive or expired, and, if applicable, the date and presumed cause of death. Results Records from 62 dogs that fulfilled the inclusion criteria were reviewed. Twenty-seven purebred breeds comprised 72% of the dogs; the remaining dogs were mixed breeds. Six large-breed dogs, Labrador Retrievers, German Shepherd Dogs, Rottweilers, Boxers, Golden Retrievers, and Collies, constituted 40% of the purebred dogs. The mean age of dogs was 6.5 years, with a range of 8 months to 13 years. There was no significant difference in the gender of the dogs affected (52% female dogs and 48% male dogs). Historically, known tick exposure was reported by the owners in only 6 out of the 62 dogs (10%). There was no seasonal distribution for disease presentation, with 58% of dogs presenting between the months of March to November. Clinical signs, summarized in Figure 1, were nonspecific and variable among dogs. The most common clinical signs included anorexia (21 dogs; 34%), lethargy (19 dogs; 31%), inactivity (17 dogs; 27%), and weight loss (15 dogs; 24%). Clinically evident hemorrhagic disorders were observed in 37 dogs (60%). Bleeding was manifested by petechial hemorrhages (10 dogs), epistaxis (6 dogs), ecchymoses (5

2 Retrospective Study of Canine Ehrlichiosis 195 Fig 1. Frequency of clinical findings in 62 dogs with ehrlichiosis. dogs), retinal hemorrhages (5 dogs), melena (5 dogs), hyphema (3 dogs), or hematuria (3 dogs). Of the 6 dogs with epistaxis, only 2 had platelet counts less than 40,000, the degree of thrombocytopenia compatible with spontaneous bleeding. Neurological abnormalities present in 10 of the dogs included ataxia (3 dogs), paraparesis (2 dogs), conscious proprioceptive deficits (4 dogs), head tilt (2 dogs), nystagmus (2 dogs), and head tremors and seizures (2 dogs). Review of each dog s medical records indicated an association with ehrlichiosis in 5 dogs, of which only 3 were thought to have primary central nervous system (CNS) infection. Cerebrospinal fluid analysis, performed in 2 dogs, identified elevated protein concentrations and increased nucleated cell counts. Two other dogs had neurological signs that may have developed secondary to the ehrlichiosis but were not considered to be caused by primary CNS infection. One dog that presented with paraplegia and absence of deep pain had spinal hemorrhages on postmortem examination that were speculated to be caused by severe thrombocytopenia. Head tremors and ataxia in another dog were suspected to be associated with serum hyperviscosity secondary to hypergammaglobulinemia; however, this relationship remained unproven. Five dogs had neurological signs that were thought to be unassociated with the E. canis infection. These clinical abnormalities accompanied a variety of disease processes, including fly-biting episodes, hepatic encephalopathy, lumbosacral disease, T-cell lymphoma, and idiopathic trigeminal neuropathy. Ocular lesions manifested in both eyes (oculi unitas) as hyphema (2 dogs), retinal detachment (3 dogs), uveitis (4 dogs), retinal degeneration (1 dog), choroidal hemorrhage (1 dog), optic neuritis (1 dog), and secondary glaucoma (1 dog) and were observed in 10 dogs. The clinicopathological data are summarized in Table 1. Twenty-six dogs were anemic (43%). A regenerative response, as assessed by a reticulocyte count, was present in 5 of 22 dogs in which the test was performed. There was nonregenerative anemia in 21 dogs (81%), leukocytosis in 15 dogs, and leukopenia in 6 dogs. There was neutrophilia without a left shift in 18 dogs and neutropenia in 8 dogs. Lymphopenia occurred in 28 dogs. Thrombocytopenia was identified in 47 dogs (77%). Pancytopenia occurred in 4 dogs. Hypoalbuminemia occurred in 42 dogs, and hyperglobulinemia occurred in 40 dogs. Serum alkaline phosphatase and alanine transaminase were elevated in 29 and 23 dogs, respectively. Azotemia was observed in a small percentage of cases, with an elevation of urea (blood urea nitrogen) in 13 dogs and an elevation of serum : creatinine concentration in 7 dogs.

3 196 Frank and Breitschwerdt Table 1. Clinicopathologic findings from 62 dogs with ehrlichiosis. a Value Hematocrit (%) White blood cell Platelets Plasma protein Neutrophils Band neutrophils Lymphocytes Reticulocyte count (%) Albumin (g/dl) Globulin (g/dl) Serum alkaline phosphatase (U/L) Serum alanine transaminase (U/L) Total bilirubin (mg/dl) Ureal (mg/dl) Creatinine (mg/dl) Total Protein (mg/dl) Number of Dogs Tested Reference Range Range Median % Low % High , , Reciprocal E. canis titer 20 20,840 5,120 NA a % Low indicates the proportion of dogs with values below the reference range. % High indicates the proportion of dogs with values above the reference range. NA indicates not applicable Proteinuria was identified through a qualitative protein test in 49 dogs. A urine protein : creatinine ratio was determined for 19 dogs in whom concurrent urine sediments were noninflammatory. The urine protein : creatinine ratio ranged from Seven dogs had values greater than 2.0; of these, 2 dogs were concurrently infected with Dirofillaria immitis. In the remaining dogs with normal urinary protein : creatinine ratios, urinary tract infections were detected by microbial culture in 5 dogs. Although gross hematuria was reported in only 3 dogs, microscopic hematuria was evident on urinalysis in 35 dogs. Serum electrophoresis, performed in 21 dogs, was used to identify polyclonal gammopathy in 9 dogs and monoclonal gammopathy in 10 dogs. Serum from 2 of these dogs had a normal distribution of albumin and globulin. Bone marrow cytology was examined in 23 dogs. Erythroid hypoplasia was noted in 7 dogs, erythroid hyperplasia in 3 dogs, myeloid hyperplasia in 6 dogs, myeloid hypoplasia in 1 dog, megakaryocytic hyperplasia in 8 dogs, and megakaryocytic hypoplasia in 5 dogs. Three dogs had erythroid hypoplasia with myeloid, megakaryotic, or myeloid megakaryotic hyperplasia. Two pancytopenic dogs had hypoplasia of all 3 cell lines. Fourteen dogs had bone marrow plasmacytosis. A Coomb s test, performed in 12 dogs, was positive in 2 dogs with moderate anemia. A 3rd dog with marked anemia had autoagglutination with hemoglobinemia, so a Coomb s test was not performed. Antinuclear antibodies were detected in 2 of 15 dogs (reciprocal titers 80 and 2,560). All 62 dogs were positive for E. canis antibodies. The reciprocal serologic titers ranged from 20 to 20,480 (Figure 2). E. canis was isolated from 3 of the 5 dogs in whom isolation was attempted. Babesia canis specific antibodies were detected in 1 of 4 dogs tested. Serologic testing for Rocky Mountain spotted fever was performed in 26 dogs: 18 dogs tested negative and 8 dogs tested positive. These results indicated exposure to spotted fever group rickettsiae with no rise in convalescent titers. Thoracic radiographs revealed a generalized pulmonary interstitial pattern of mild to moderate severity in 9 of 33 dogs for whom radiographs were obtained. Abdominal ultrasound, performed on 18 dogs, identified splenomegaly in 9 dogs, hepatomegaly in 3 dogs, and hyperechogenicity of the renal cortex in 6 dogs. A kidney biopsy obtained from 1 dog revealed glomerulonephropathy. Further characterization of the glomerular disease was not available. We used flow cytometry to identify an abnormal distribution of T-cell lymphocytes in 3 of 4 dogs tested (Table 2). Posttreatment FACS analysis revealed normalization of the CD4 : CD8 ratio in the 1 dog for whom serial analysis was available. There was an inversion of the CD4 : CD8 ratio because of increases in CD8 cells in dogs A and B. The results of the flow cytometry for dog B were initially suggestive of a T-cell tumor. However, a diagnosis of ehrlichiosis was made, and treatment was initiated. One year later, this dog developed B-cell lymphoma; flow cytometry revealed a normal CD4 : CD8 ratio and a high percentage of B cells (96.22% IgG). Dog C exhibited a normal distribution of subsets and a normal CD4 : CD8 ratio. The results of the flow cytometry for dog D showed no circulating B cells, and 18% of the T cells for this dog were pan T-cells positive for both CD4 and CD8. Repeated samples obtained over several months revealed a similar distribution. Concurrent disease diagnoses included systemic lupus erythematosus (SLE; 1 dog); heartworm disease (4 dogs); hepatic insufficiency (6 dogs); T-cell lymphoma (1 dog); interstitial cell tumor, seminoma, cutaneous histiocytoma, and amelanotic melanoma (1 dog); iris melanoma (1 dog); retropharyngeal fibrosarcoma (1 dog); multiple myeloma (1 dog); disseminated intravascular coagulation (DIC; 2 dogs); sepsis (1 dog); pneumonia (3 dogs); renal failure (5 dogs);

4 Retrospective Study of Canine Ehrlichiosis 197 Fig 2. Serological titers of Ehrlichia canis (determined by immunofluorescent antibody test) for dogs with ehrlichiosis. The magnitude of the titer was unknown in 1 dog because it was diagnosed by the referring veterinarian. heartworm disease (4 dogs); hemolytic anemia (2 dogs); uveitis (3 dogs); prostatitis (1 dog); cutaneous pythiosis (1 dog); mycotic pleuritis (Aspergillus species) (1 dog); disseminated phaeohyphomycosis (1 dog); and hepatopathy (5 dogs). Treatment with tetracycline hydrochloride (Global Pharmaceutical Corp, Philadelphia, PA) was used in 31 dogs; doxycycline (Vibramycin, Pfizer Inc, Lee s Summit, MO) was used in 11 dogs; and chloramphenicol (Phoenix Pharmaceutical Inc, St Joseph, MO) was used in 8 dogs. Concurrent antibiotics, including amoxicillin clavulanic acid (Clavamox, SmithKline Beecham Animal Health, Exton, PA), cephalexin (Keflex, Dista Products Co, Atlanta, GA), and enrofloxacin (Baytril, Bayer Corporation, Shawnee Mission, KS) were administered to 14 dogs. Prednisone (Prednistab, Vedco Inc, St Joseph, MO) was used in the treatment of 8 dogs. A blood transfusion was required in only 2 dogs. Adjunctive chemotherapy with either vincristine (Oncovin, Eli Lilly and Co, Indianapolis, IN) or cyclophosphamide (Cytoxan, Bristol Myers Squibb, Princeton, Table 2. Flow cytometry in 4 dogs with ehrlichiosis. Dog T Cells CD4 CD8 CD4 : CD8 B Cells A B C D 67.32% 589.7/ L 96.65% 9,665/ L 77.94% 1,378/ L 95% 8,204/ L 11.98% 6.65% 665/ L 47.67% 842/ L 38% 3,282/ L 60.17% 91% 9,100/ L 23.64% 418/ L 25% 2,159/ L % 17.52/ L 2.98% 298/ L 28.96% 512/ L 1% 86.36/ L Total Lymphocyte Count 876/ L 10,000/ L 1,768/ L 8,636/ L

5 198 Frank and Breitschwerdt NJ) was employed in the management of 4 dogs with severe thrombocytopenia or anemia for which an immune-mediated cellular destruction was suspected. Response to therapy was variable. Nine dogs were alive at the time of the survey, 34 were dead, 18 were lost to follow-up, and the status of 1 dog was unknown. Eighteen dogs died or were euthanized for other medical causes at least 1 year (mean 2.6 years) after diagnosis. Nine dogs died less than 1 week after diagnosis, and 3 dogs died in less than 1 year but more than 1 week after diagnosis (mean 4.5 months). The time of death after diagnosis was unknown for 4 dogs. The cause of death was unknown for 5 dogs. Twentyone dogs died as a result of unrelated disease processes, including renal failure, neoplasia, trauma, SLE, congestive heart failure, and respiratory disease. Two dogs died as a direct result of the E. canis infection: 1 died from intractable seizures that were likely caused by serum hyperviscosity, and the other was euthanized because of posterior paralysis secondary to spinal hemorrhage and thrombocytopenia. In 6 dogs, the cause of death was unclear and may have been related to ehrlichiosis. For example, 1 dog died from disseminated phaeohyphomycosis and DIC; 1 from pneumonia, DIC, and cardiac arrest; and another from renal failure 1 year after diagnosis with ehrlichiosis and glomerulonephritis. When owners were questioned about their dog s initial response to therapy, 12 of them said their dogs responded rapidly and completely (defined as no recurrence of disease requiring further treatment); these owners considered their dogs healthy. Eight dogs experienced gradual resolution slowly over time (defined as slow but complete recovery), and 11 dogs responded to the treatment incompletely, as evidenced by recurrence of disease manifestations and requirement for multiple treatments. Seven dogs had a poor response and died either because of the ehrlichiosis or because of concurrent disease processes; 2 dogs responses were difficult to interpret because of concurrent SLE and chronic renal failure, respectively. Four dogs were not treated for the Ehrlichia infection. One dog died 6 days after presentation secondary to renal failure and cardiac arrest; 1 dog was euthanized shortly after diagnosis because of T- cell lymphoma; 1 dog with severe chronic renal failure and another dog that presented with a scapular fracture were, for no apparent reason, not treated for the ehrlichiosis. Posttreatment blood testing was performed in 12 dogs at various intervals. In 3 dogs, hematologic recovery was evident by 1 month after therapy; in 2 dogs, improvement was noted by 2 months after therapy; and in 5 dogs, there was improvement at 3 months after therapy. Thrombocytopenia persisted in 1 dog for at 1 year despite multiple treatments. Follow-up serologic testing was performed in 8 dogs. At 1 month after diagnosis, all 8 dogs had persistently elevated titers. Of these 5 dogs, 1 had a persistently elevated titer at 6 months, 1 at 7 months, 1 at 1 year, and 1 at 3 years. Discussion Several features of canine ehrlichiosis reported in this study are similar to those previously reported by other investigators. Signalment and lack of seasonal prevalence were comparable with other reports. 1 5 History and clinical signs, however, were often vague and were inconsistent from dog to dog. Clinical signs were nonspecific, and the most common symptoms included lethargy, inactivity, weight loss, and anorexia. However, these nonspecific signs of ehrlichiosis occurred at much lower frequencies than reported in previous retrospective studies. In this study, the percentage of dogs experiencing anorexia, depression, and weight loss ranged from 25 35% in comparison with 40 78% in other studies. 1 5 Pyrexia occurred in 17% of dogs, a lower frequency when compared with the 50% figure reported by Troy et al, 1 the 30% reported by Waddle and Littman, 3 the 49% reported by Woody, 4 and the 60% reported by Harrus et al. 5 Tick exposure was reported in only 10% of dogs in this report. This rate is comparable to the rate in Waddle s study, in which tick exposure was identified in 11% of dogs. 3 However, the frequency of exposure was much lower than that reported by Woody 4 and Troy 1, who reported exposure in 39% and 40% of dogs, respectively. The discrepancy in these results may be explained by the diversity of cases in which serologic testing was performed. The dogs in this study were referred to the NCSU College of Veterinary Medicine for a variety of clinical signs, some of which would not be considered typical of ehrlichiosis. Since ehrlichiosis is endemic in this region, serological testing may have been pursued because of an increased index of suspicion. Ehrlichiosis was subsequently diagnosed as a result of serologic testing. In addition, these discrepancies may be related to the intrinsic limitations of the retrospective study design in that there may be inaccuracy in data collected from medical records as well as differences in interpretation of clinical signs by different clinicians. This may confound the true prevalence of certain clinical signs. The presence of a positive antibody titer to E. canis antigens was considered significant for exposure to monocytic Ehrlichiae. Only 8 dogs (13%) had reciprocal titers less than 80. The clinical and laboratory abnormalities in dogs with a low antibody titer did not differ from the abnormalities in dogs with high titers. Recent work using western blot analysis of sera from dogs with low antibody titers has demonstrated patterns consistent with exposure to monocytic Ehrlichiae. 6 Based on the low correlation between antibody titer and disease manifestations and the high prevalence of ehrlichiosis in this geographic region, we consider the possibility of false-positive results to be unlikely. Future studies that are currently in progress should attempt to correlate the magnitude of the antibody titer, the presence of a positive Ehrlichia culture result, and the detection of Ehrlichia DNA by polymerase chain reaction. The ability to discriminate between the acute and chronic phase of an E. canis infection in clinical cases is impossible. Therefore, the phase of infection could not be determined in the dogs in this study. Ehrlichia morulae are difficult to detect, and were only seen in 1 dog, thus indicating acute disease. The number of dogs with known tick exposure differed tremendously from previous reports; this may be caused by the variation in disease phase. 1,4 The presence of ticks would be more common during the acute phase, or it may be a coincidental finding during the chronic phase.

6 Retrospective Study of Canine Ehrlichiosis 199 Overall, hemorrhagic tendencies (60%) were more commonly observed in this study; however, the types of bleeding disorders were similar to those previously reported. 4,5 Epistaxis occurred in 10% of the dogs; this figure is lower than that described by other authors, which has ranged from 20 57%. 1 5,7 Many dogs in this study that experienced bleeding disorders did not have thrombocytopenia of sufficient severity to cause spontaneous bleeding. Coagulation profiles performed in 3 of 6 dogs with epistaxis identified normal figures for platelet counts, prothrombin time, partial thromboplastin time, fibrinogen, and fibrin degradation products. Thrombocytopathia has been postulated to be a cause of bleeding in dogs with sufficient platelet numbers and normal coagulation profiles. Data from naturally infected dogs are lacking, although recently a naturally infected dog was reported with thrombocytopathia (defined by reduced platelet aggregation in response to collagen and no response to epinephrine or ristocetin). 8 Inhibition of platelet aggregation has been described in response to collagen, ADP, epinephrine, and thrombin in experimentally induced acute ehrlichiosis. 9 Mechanisms for the thrombocytopathia have been theorized. Decreases in platelet adhesiveness have been demonstrated by platelet retention tests. 10 Interference with platelet factor 3 release has also been postulated. 10 A platelet migration inhibitory factor produced by lymphocytes of E. canis infected dogs contributes to the thrombocytopathia by inhibiting platelet pseudopod formation. 11 Hyperglobulinemia may also interfere with platelet migration and adhesiveness. 11 Antiplatelet antibodies have been demonstrated in dogs with experimentally induced ehrlichiosis. 9 Inhibition of platelet aggregation and antiplatelet antibodies were detected concurrently, suggesting a causal relationship. 9 However, in the same study, platelet aggregation was inhibited in some dogs at a time in which antiplatelet antibodies were absent, indicating that other factors are also important. 9 Clinicopathologic abnormalities found in this study were similar to abnormalities found in previous reports. 1 5 Anemia and thrombocytopenia were the most common hematologic abnormalities. However, the frequency of anemia (43% of dogs) in this study was less than that previously reported (67 90%). 1 5 The majority of dogs (81%) had a nonregenerative anemia. A Coomb s test was positive in 2 of 12 dogs. A positive Coomb s test implies an immunologic reaction that is presumably caused by the adherence of immunoglobulins to red cell membranes or to cross-reactivity of Ehrlichia with red cell antigens. However, concurrent hepatic carcinoma and disseminated phaeohyphomycosis were diagnosed in the Coomb s-positive dogs; these may have influenced the positive test result. Seventy-one percent of dogs in this study were thrombocytopenic; these results are similar to those reported by Khuen, 2 Waddle, 3 and Harrus 5 (64%, 84%, and 73%, respectively). However, these values are lower than those documented by Woody (96%) 4 and Troy (100%). 1 The discrepancy in the number of thrombocytopenic dogs in the latter studies may be attributable to the analysis of E. canis serology because of thrombocytopenia. The number of E. canis seroreactive dogs with normal platelet counts (29% in this study) may be greater in an endemic region in which serologic testing may be used more frequently in dogs with nonspecific clinical findings. It is important to note that the absence of thrombocytopenia should not exclude E. canis infection from the differential diagnosis. Hypoalbuminemia and hyperglobulinemia were the most frequent biochemical abnormalities. In our study, 70% of dogs were hypoalbuminemic; this number is higher than the number given in previous studies (29 45%). 2 5 Causes for hypoalbuminemia can include hepatic insufficiency, nephropathy with protein loss, gastrointestinal losses, vasculitis, and hyperglobulinemia. 12 In this study, there is clinicopathologic support suggesting that both hyperglobulinemia and proteinuria contribute to hypoalbuminemia. Hyperglobulinemia was documented in 68% of dogs. Both monoclonal and polyclonal gammopathies have been previously reported in dogs with ehrlichiosis, 7 and both distributions were observed within the population of dogs in this study. Differential diagnoses for monoclonal gammopathy include multiple myeloma, plasma cell leukemia, lymphoid leukemia, macroglobulinemia, cutaneous amyloidosis, and idiopathic monoclonal gammopathy. 7 However, ehrlichiosis is an important differential diagnosis for monoclonal gammopathies, especially with regard to a more favorable prognosis for long-term survival following treatment. Protein-losing nephropathy has been documented in the experimentally induced acute phase of ehrlichiosis. 13 An immunologic mechanism for glomerular injury was proposed because of the deposition of anti-canine IgM, IgG, and C 3 in the glomerular tufts and mesangium. 13 Glomerulonephritis and nephropathy with protein loss have been associated with chronic ehrlichiosis. An immunologic mechanism is likely to come, but one has not yet been investigated experimentally. Although proteinuria was detected in 49 dogs in this study, a renal biopsy was obtained from only 1 dog. There was evidence of gross hematuria in only 3 dogs, although microscopic hematuria was detected through urinalysis in 35 dogs. Severe thrombocytopenia (platelet counts less than 40,000) may have contributed to the hematuria in 7 dogs. Microscopic hematuria was not explained by either marked thrombocytopenia or urinary tract infection in 31 dogs. The significance of this observation is unknown. Cystocentesis is the most frequently used technique for obtaining a urine specimen in our hospital; this technique may result in iatrogenic hematuria. Microscopic hematuria has not been previously reported as a frequent finding in dogs with ehrlichiosis, but it may be an early marker of glomerular injury prior to the development of proteinuria. 14 Alternatively, thrombocytopathia associated with E. canis infections may also result in spontaneous bleeding. Bone marrow hyperplasia and bone marrow hypoplasia were detected in a similar number of dogs. Conventionally, megakaryocytic and myeloid hyperplasias with erythroid hypoplasia are common findings in acute ehrlichiosis, and bone marrow hypoplasia and pancytopenia can occur in the chronic phase. 15 However, pancytopenia of the bone marrow was uncommon in this study, occurring in only 2 of 23 dogs (9%). Plasmacytosis of the bone marrow was documented in % of dogs in which a bone marrow aspirate was performed. This finding has been previously docu-

7 200 Frank and Breitschwerdt mented in dogs with ehrlichiosis, as has plasmacytosis of various other organs, such as the spleen, meninges, lungs, and kidneys. 2,4,15 Clinical signs of neurological disease were thought to be related to ehrlichiosis in 5 dogs. Commonly described neurological signs associated with ehrlichiosis include ataxia, seizures, paraparesis or tetraparesis with upper or lower motor-neuron deficits, stupor, vestibular disease, and back or neck pain. 4,5,15,16 Neurological signs are usually attributed to plasma cell infiltration of the meninges or to hemorrhage into the cerebral or spinal cord parenchyma. 15 Paraplegia resulted from hemorrhage into the spinal cord secondary to marked thrombocytopenia in 1 dog in this study. An interstitial pulmonary pattern was observed in 9 dogs. This nonspecific radiographic finding has been documented previously in dogs with ehrlichiosis and is thought to represent interstitial hemorrhage, septal hypertrophy, or pulmonary plasmacytosis. 17 Differential diagnoses include infectious diseases such as heartworm disease, pulmonary mycosis, pneumonia, pulmonary hemorrhage, and allergic pneumonitis. 17 In this study, concurrent heartworm disease was diagnosed in 1 dog with interstitial pneumonia, and congestive heart failure was diagnosed in another. Concurrent diseases occurred frequently in this study. In most dogs, the disease did not appear to have a direct causal relationship to the ehrlichiosis and was considered an incidental finding. Examples of concurrent diseases include systemic lupus erythematosus, heartworm disease, hepatic insufficiency, T-cell lymphoma, and neoplasia. Renal failure was diagnosed in 5 dogs. Renal failure is typically not directly associated with ehrlichiosis. In theory, however, deposition of antigen antibody immune complexes may result in glomerulonephritis that could predispose dogs to chronic renal failure and proteinuria. In this study, there were also a number of concurrent diseases that may be associated with ehrlichiosis as either a direct result of the E. canis infection or secondary to immunosuppression. These include DIC, sepsis, pneumonia, hemolytic anemia, uveitis, cutaneous pythiosis, mycotic pleuritis, and disseminated phaecomycosis. Immunodysregulation likely occurs as a result of ehrlichial infection as evidenced by the disruption of the CD4 : CD8 ratio documented using flow cytometry in several dogs. The mechanism of immunodysregulation is unknown. Cell-mediated immunity (CMI) was depressed in German Shepherd Dogs experimentally infected with E. canis, as demonstrated with a leukocyte migration inhibition test and by a decrease in delayed-type hypersensitivity response to specific antigens. 18 It is possible that immunodysregulation may predispose animals to concurrent infections, sepsis, or dissemination of a localized infection. For example, 1 dog in this study developed systemic dissemination of a localized infection with phaecomycosis. Alternately, diseases that compromise the immune system may make an animal more susceptible to develop the chronic clinical manifestations of ehrlichiosis. Response to therapy was variable, and overall response was less favorable than in other studies. 1,3,5 Of the 44 dogs for whom follow-up data was obtained, a swift complete response occurred in only 27% of dogs, and a slow complete response occurred in 18% of dogs. These percentages compare with a 72% complete response rate reported by Waddle and Littman 3, a 70% response rate reported by Troy et al, 1 and a 65% survivability rate in Harrus study. 5 In our study, 25% of dogs had an incomplete response, and 16% had a poor response. However, the cause of death in the majority of dogs in our study was thought to be associated with concurrent diseases such as renal failure, hepatic insufficiency, cardiac disease, and neoplasia. Although 2 dogs (4.5%) died as a direct consequence of ehrlichiosis, another 6 dogs (13.6%) died acutely of disease processes that were probably related to the E. canis infection. In this study, mortality due to ehrlichiosis was high (18%), and the overall success rate of therapy was lower than previously reported. The reason for this discrepancy remains unclear. Antibody titers generally remain positive for 3 to 9 months after therapy and may remain elevated for up to 31 months. 19 Bartsch and Greene reported persistence of antibody titers in 40% of dogs 1 year after diagnosis with E. canis infection. 20 In our study, 1 dog remained seropositive for 3 years and another dog remained seropositive at 1 year. The persistence of serum antibody titers does not reflect the presence of a CMI response or clearance of the organism. 11 Clinical recovery and amelioration of laboratory parameters often precede declines in serologic titers. However, the presence of persistently elevated antibody titers could imply the presence of the organism with concurrent stimulation of immune system, suggesting failure to completely eliminate the infection. For this reason, whether the length of medical therapy should be based upon clinicopathologic signs or on serologic titers is still not clear. The efficacy of current medical therapy (including tetracycline and doxycycline) is controversial, considering the persistence of antibody titers and variable response to treatment. In our study, treatment strategy varied widely, and the success of 1 therapy over another was not apparent. E. canis was isolated from experimentally inoculated dogs after therapy with tetracycline at 10 mg/kg PO once daily for 1 week. 21 Reductions in IFA titers of 16-fold magnitude occurred by 49 days after treatment in 2 dogs for whom the organism was cleared. Titers did not decline appreciably in the 3 dogs for whom the organism was not cleared. This particular study implied that current drug dosages may be insufficient and that treatment periods should be extended to longer than 1 week. 21 Current therapeutic protocols already suggest treatment periods of 3 to 4 weeks, since shorter treatment intervals may result in ineffective therapy and persistent infection. 21 The role of chronic E. canis infection in the development of concurrent disease processes remains unclear. Controlled studies are needed to evaluate the influence of E. canis infection on CMI. Flow cytometry may prove useful in detecting changes in the lymphocyte subset distribution suggestive of ehrlichiosis. The response to therapy in this study was variable, yet the majority of deaths were presumably caused by unrelated disease processes. However, it is important to note that 18% of the mortality rate was presumed to be secondary to ehrlichiosis. Based on these findings and on the fact that serologic titers remained persistently elevated in a large percentage of cases, we believe that further investigation utilizing more advanced molecular techniques such as poly-

8 Retrospective Study of Canine Ehrlichiosis 201 merase chain reaction (PCR) should be made to evaluate the persistence of the organism following current treatment protocols in order to better evaluate response to therapy. Future studies should attempt to correlate the antibody titer, Ehrlichia culture results, and the detection of Ehrlichia DNA by PCR. Acknowledgments We acknowledge the many clinicians who managed the cases summarized in this study. We thank the staff of the NCSU Clinical Pathology Laboratory, the Immunology Laboratory, the FACS Analysis Laboratory, and the Tickborne Diseases Diagnostic Laboratory. In particular, we acknowledge Ms. Robin Gager and Dr. Michael Levy for providing serologic testing for E. canis to the Veterinary Teaching Hospital. References 1. Troy GC, Vulgamott JC, Turnwald GH. Canine ehrlichiosis: A retrospective study of 30 naturally occurring cases. J Am Anim Hosp Assoc 1980;16: Kuehn NF, Gaunt SD. Clinical and hematological findings in canine ehrlichiosis. J Am Vet Med Assoc 1985;186: Waddle JR, Littman MP. A retrospective study of 27 cases of naturally occurring canine ehrlichiosis. J Am Anim Hosp Assoc 1988; 24: Woody BJ. Canine EhrlichiosisRetrospective review of 132 clinical cases from Texas A&M University1979 to American Veterinary Medical Association Meeting, New Orleans, LA, July Harrus S, Kass PH, Klement E, Waner T. Canine monocytic ehrlichiosis: A retrospective study of 100 cases, and an epidemiological investigation of prognostic indicators for the disease. Vet Rec 1997; 141: Hegarty BC, Levy MG, Gager RF, Breitschwerdt EB. Immunoblot analysis of the immunoglobulin G response to Ehrlichia canis in dogs: An international survey. J Vet Diagn Invest 1997;9: Breitschwerdt EB, Woody BJ, Zerbe CA, et al. Monoclonal gammopathy associated with naturally occurring canine ehrlichiosis. J Vet Intern Med 1987;1: Varela F, Font X, Valladares JE, Alberola J. Thrombocytopathia and light-chain proteinuria in a dog naturally infected with Ehrlichia canis. J Vet Intern Med 1997;11: Harrus S, Waner T, Eldor A, et al. Platelet dysfunction associated with experimental acute canine ehrlichiosis. Vet Rec 1996;139: Lovering SL, Pierce KR, Adams LG. Serum complement and blood platelet adhesiveness in acute canine ehrlichiosis. Am J Vet Res 1980;41: Greene RT. Canine ehrlichiosis: Clinical implications for humoral factors. In: Bonagura JD, Kirk RW, eds. Current Veterinary Therapy XII, Small Animal Practice. Philadelphia, PA: WB Saunders; 1995: Turnwald GH, Barta O. Immunologic and plasma protein disorders. In: Willard MD, Tvedten H, Turnwald GH, eds. Small Animal Clinical Diagnosis by Laboratory Methods. Philadelphia, PA: WB Saunders; 1989: Codner EC, Maslin WR. Investigation of renal protein loss in dogs with acute experimentally induced Ehrlichia canis infection. Am J Vet Res 1992;53: Rose BD, Renke HG. Urinalysis and approach to the patient with renal disease. In: Renal Pathology, the Essentials. Baltimore, MD: Williams and Wilkins; 1994: Woody BJ, Hoskins JD. Ehrlichial diseases of dogs. Vet Clin North Am Small Anim Pract 1991;21: Woody BJ, McDonald RK. Canine ehrlichiosis. Miss Vet J 1985;1: Codner EC, Roberts RE, Ainsworth AG. Atypical findings in 16 cases of canine ehrlichiosis. J Am Vet Med Assoc 1985;186: Nyindo M, Huxsoll DL, Ristic M, et al. Cell mediated and humoral immune responses of German Shepherd Dogs and Beagles to experimental infection with Ehrlichia canis. Am J Vet Res 1980;41: Perille AL, Matus RE. Canine ehrlichiosis in six dogs with persistently increased antibody titers. J Vet Intern Med 1991;5: Bartsch RC, Greene RT. Post-therapy antibody titers in dogs with ehrlichiosis: Follow up study on 68 patients treated primarily with tetracycline and/or doxycycline. J Vet Intern Med 1996;10: Iqbal Z, Rikihisa Y. Reisolation of Ehrlichia canis from blood and tissues of dogs after doxycycline treatment. J Clin Microbiol 1994; 32: