Clinicopathological Abnormalities and Treatment Response in 24 Dogs Seroreactive to Bartonella vinsonii (berkhoffii) Antigens
Bartonella vinsonii (B. vinsonii) subspecies berkhoffii is a recently recognized cause of endocarditis, myocarditis, and granulomatous disease in dogs. In an effort to elucidate other potential disease manifestations, the case records of 24 dogs that were seroreactive to B. vinsonii (berkhoffii) antigens were studied retrospectively. Diagnoses included immune-mediated hemolytic anemia, neutrophilic or granulomatous meningoencephalitis, neutrophilic polyarthritis, cutaneous vasculitis, and uveitis. Repeated B. vinsonii (berkhoffii) antibody titers became negative after treatment. This study indicates that a diverse spectrum of disease manifestations and clinicopathological abnormalities can be detected in dogs that are seroreactive to B. vinsonii (berkhoffii) antigens.
Introduction
Most Bartonella (B.) spp. have been discovered within the past 10 years. This group of organisms is composed of highly adapted, vector-transmitted, intraerythrocytic bacteria that frequently induce persistent infections. An increasingly diverse spectrum of Bartonella-associated manifestations have been recognized in humans as a result of the acquired immunodeficiency syndrome (AIDS) epidemic, improved culture techniques, and the advent of serological and molecular diagnostic testing.1–4 Although descriptions of cat scratch disease, Trench Fever, and Oroya Fever have existed for more than half a century, B. henselae, B. quintana, and B. bacilliformis, as well as other previously unknown Bartonella spp. have received increasing attention in the contemporary literature.
Based upon human data, the clinical manifestations and histopathological lesions associated with Bartonella infection can vary depending on the immunological status of the host, the site of tissue injury, and the temporal evaluation of the lesion.1 To date, B. henselae infection in humans has been associated with lymphadenopathy,2 meningitis,3 bacillary angiomatosis,4 neuroretinitis,35 encephalopathy,6 osteomyelitis,7 peliosis hepatis,8 granulomatous hepatitis and splenitis,9 and polyarthritis.10 In addition, six Bartonella species (i.e., B. henselae,11 B. quintana,12 B. elizibetheae,13 B. washoensis,14 B. vinsonii subspecies berkhoffii,15 and B. vinsonii subspecies arupensis16) have been isolated from humans with endocarditis since 1993. During the past decade, there has been rapid expansion of clinical and pathological data related to the spectrum of disease in humans, as well as increased recognition of the diversity of Bartonella spp. that induce specific conditions, such as endocarditis.
In dogs, a novel Bartonella subspecies, B. vinsonii (berkhoffii), is emerging as a potentially important cause of a diverse group of pathologic disorders. To date, B. vinsonii (berkhoffii) has been implicated (on the basis of culture, serological, or molecular evidence) as a cause of endocarditis,1718 myocarditis,18 and granulomatous disease19 in dogs. Recently, Chomel et al. identified B. clarridgeiae as a cause of endocarditis in a dog, suggesting that additional Bartonella spp. may be implicated as a cause of endocarditis in dogs in the future.14
The duration of Bartonella infection following vector transmission in nature has not been established. Dogs presumably develop chronic intravascular infections that result in an insidious debilitating illness, which at times can be accompanied by acute, life-threatening decompensation secondary to endocarditis or myocarditis.1718 Seroreactivity to B. vinsonii (berkhoffii) antigens is frequently found in dogs seroreactive to other tick-borne pathogens, including Ehrlichia canis (E. canis), Babesia canis (B. canis), and Rickettsia rickettsii (R. rickettsii).20–23 In these animals, coinfection complicates the task of determining to what extent certain clinical signs are attributable to infection with a Bartonella spp. For this reason, dogs with serological evidence of exposure to other tick-transmitted organisms were excluded from this retrospective study.
Similar to other members of the genus Bartonella, B. vinsonii (berkhoffii) is extremely difficult to culture from the blood of healthy or sick dogs. Because of the limitations associated with culture, serology is currently the most useful tool for screening dogs for exposure to or infection with B. vinsonii (berkhoffii).22–25 Bartonella vinsonii (berkhoffii) seroprevalence in canine populations is generally low but appears to vary with the degree of tick exposure.22–25 In an epidemiological study involving sick dogs from North Carolina and Virginia, B. vinsonii (berkhoffii) seroprevalence was 3.6% (69/1920 samples).22 In a study from Israel, involving dogs suspected of having a tick-borne disease, B. vinsonii (berkhoffii) seroprevalence was 10% (4/40 samples).23 In a cross-sectional study of healthy dogs working for the United States government (a dog population with high exposure to insect vectors), the B. vinsonii (berkhoffii) seroprevalence was 8.7% (162/1872 samples).24 In a study from Rhode Island, assessing the association between the geographic distribution of Ixodes scapularis and seropositivity to various tick-transmitted organisms, B. vinsonii (berkhoffii) seroprevalence was 2.2% (6/277 samples).25 In contrast, B. vinsonii (berkhoffii) seroprevalence was 93% (25/27 samples) in a flea- and tick-infested walker hound kennel in North Carolina.20 Coincidentally, all of the dogs in the walker hound kennel that were B. vinsonii (berkhoffii) seroreactive also had antibodies to one or more other tick-borne pathogens, including E. canis, Ehrlichia equi, R. rickettsii, or B. canis.20 In an epidemiological study of coyotes from California that experienced extensive tick exposure, 28% (31/109 samples) were bacteremic, and 76% (83/109 samples) were seroreactive to B. vinsonii (berkhoffii) antigens.26 It is not known why B. vinsonii (berkhoffii) can be cultured from a relatively high percentage of seroreactive coyotes, in contrast to the low isolation rate from seroreactive domestic dogs.18–20
The purpose of this retrospective study was to determine the historical and clinicopathological abnormalities and the treatment outcomes for dogs that were seroreactive to B. vinsonii (berkhoffii) antigens but not seroreactive to other tick-transmitted pathogens, including E. canis, R. rickettsii, B. canis, Babesia gibsoni (B. gibsoni), or Borrelia burgdorferi antigens, at the time of clinical evaluation.
Materials and Methods
For inclusion in this study, potential cases were identified from the database of the North Carolina State University, College of Veterinary Medicine, Vector-Borne Disease Diagnostic Laboratory (NCSU-CVM-VBDDL). The immunofluorescent antibody (IFA) techniques, used to detect antibodies to E. canis, R. rickettsii, B. canis, and B. gibsoni have been described previously.20–22 A commercially available enzyme-linked immunosorbent assay (ELISA) testa was used to detect serological evidence of natural exposure to B. burgdorferi. In a previous study testing serum from experimentally infected dogs, cross reactivity was not detected between B. vinsonii (berkhoffii) and E. canis, B. canis, or R. rickettsii antigens.22 For diagnostic purposes, the NCSU-CVM-VBDDL currently defines a B. vinsonii (berkhoffii) seroreactive sample as a reciprocal titer of ≥64,22–25 and although incompletely characterized, a single titer is used to implicate prior exposure or chronic infection with B. vinsonii (berkhoffii).17–21 For this study, serological inclusion criteria consisted of a B. vinsonii (berkhoffii) reciprocal IFA titer of ≥64, with a concurrent lack of seroreactivity to E. canis (IFA, ≤32), R. rickettsii (IFA, ≤64), B. burgdorferi (ELISA, negative), B. canis (IFA, ≤64), and B. gibsoni (IFA, ≤64). In addition, only those cases with a complete medical record were included in the study. To avoid factors that could influence the interpretation of the clinical and laboratory data, cases with concurrent diseases of known causation (e.g., neoplasia) were also excluded in an effort to limit case selection to dogs in which the clinical manifestations might reasonably be associated with B. vinsonii (berkhoffii) seroreactivity.
Data abstracted from each medical record included historical and physical examination abnormalities, diagnostic test results, and treatments administered during the period of evaluation (i.e., from the time of initial testing until death or resolution of clinical signs). In cases for which multiple complete blood counts (CBC) and serum biochemical profiles were reported, the values selected for descriptive analysis were based on close temporal association with the initial B. vinsonii (berkhoffii) antibody titer. When available, the results of other diagnostic tests (e.g., bone-marrow cytopathology, cerebrospinal fluid analysis) were reviewed. Reference ranges, as cited in this study, were established by the Clinical Pathology Laboratory, North Carolina State University. Repeated serological testing was performed in 11 of 24 dogs, by the same technician, using the same IFA methodology. Telephone follow-up was attempted with owners of dogs for whom treatment outcome was not documented in the medical record.
Results
Case Material
From July 1, 2000 through June 30, 2001, 790 canine serum samples were tested for antibodies to B. vinsonii (berkhoffii) antigens at the NCSU-CVM-VBDDL, of which 90 (11.4%) samples were reactive to B. vinsonii (berkhoffii) antigens (reciprocal IFA titers, ≥64). Of these 90 samples, 13 (14%) were concurrently reactive to E. canis antigens, and 17 (19%) were reactive to R. rickettsii antigens, including one sample that was reactive to B. vinsonii, E. canis, and R. rickettsii antigens. Because of the potential of concurrent infection, these cases were excluded from the present study. None of the 90 B. vinsonii (berkhoffii) seroreactive samples were concurrently reactive to B. canis, B. burgdorferi, or B. gibsonii antigens. In addition to the dogs that were seroreactive to other test antigens, 37 dogs were also excluded because of a lack of complete medical records or the presence of confounding diseases, such as neoplasia. Of the 90 B. vinsonii (berkhoffii) seroreactive dogs detected during a 1-year period, 24 cases met the previously defined inclusion criteria. The reciprocal B. vinsonii (berkhoffii) antibody titers for these dogs were 64 (n=5), 128 (n=8), 256 (n=7), 512 (n=2), 1024 (n=1), and 2048 (n=1).
The medical records for dogs reported here originated from the following sources: the NCSU Veterinary Teaching Hospital (n=12), Texas A&M Veterinary Medical Center (n=4), the University of Florida Veterinary Teaching Hospital (n=1), a private specialty clinic in Colorado (n=1), and general practitioners residing in the southeastern United States (n=6). The 24 dogs included in this study were from North Carolina (n=11), Virginia (n=4), Texas (n=4), Florida (n=2), South Carolina (n=1), Alabama (n=1), and Colorado (n=1). Seven dogs were male, and 17 were female. There were 19 purebred dogs of various breeds and five mixed-breed dogs. Ages ranged from 1 to 14 years (mean, 7 years; median, 8 years; standard deviation [SD], ±4 years). There was no seasonal pattern noted as to the time of presentation. Duration of illness prior to serological testing for B. vinsonii (berkhoffii) antibodies ranged from 2 days to 2 years (mean, 84 days; median, 30 days; SD, ±155 days) [Tables 1, 2, 3].
Clinical Signs
Clinical signs varied among the 24 dogs, with most dogs having multiple clinical signs attributed to more than one organ system [Tables 2, 3]. Overall, 16 (67%) dogs were lethargic; 13 (54%) had weight loss, anorexia, and/or inappetence; and seven (29%) had muscle and/or joint pain. Vomiting accompanied other signs of illness in two dogs. At initial presentation, three of the dogs had cutaneous lesions consistent with a vasculitis. Two dogs (case nos. 2, 5; Table 3) had erythremic macules on the trunk, which did not blanche with diascopy. The third dog (case no. 4; Table 2) had cicatricial alopecia of the tips of the pinnae, similar alopecic patches on the hind limbs, and lymphocytic plasmacytic vasculitis confirmed by biopsy. Five (21%) dogs had hind-limb paresis, and five dogs were pyretic.
Four (17%) dogs presented with bilateral ocular abnormalities, including one dog each with anterior uveitis and multifocal chorioretinitis, extensive hyphema (in conjunction with oral mucosal petechial hemorrhages), and complete retinal detachment (in association with systemic hypertension). Neutrophilic polyarthritis was diagnosed concurrently in the dog with uveitis. Four (17%) dogs were ataxic, four (17%) were recumbent, three (12.5%) had collapsing episodes, and three (12.5%) had seizures. Mitral-valve endocarditis was diagnosed via echocardiography in one dog. At necropsy two dogs were found to have myocardial disease. A 9-year-old, spayed female dog (case no. 11; Table 3) treated for acute, rapidly progressive immune-mediated hemolytic anemia/immune thrombocytopenia (IMHA/ITP) had multifocal, neutrophilic myocarditis at necropsy, indicative of a septicemic episode. A 3-year-old, spayed female rottweiler (case no. 9; Table 3) with acute, rapidly progressive IMHA/ITP accompanied by collapsing episodes, ventricular tachyarrhythmias, and progressive pulmonary edema, had myocardial fibrosis and mineralization involving the left atrium and left ventricle and mineralization of the proximal aorta, based on necropsy results. Despite the presence of myocardial fibrosis and mineralization in the rottweiler, the duration of illness in these latter two dogs was only 2 weeks.
Laboratory Abnormalities
Twelve (50%) dogs were thrombocytopenic (platelet counts, <180 × 103/μL), with platelet counts ranging from 18 to 172 × 103/μL (mean, 87 × 103/μL; SD, ±50 × 103/μL). Eight (33%) dogs were anemic, with packed cell volumes (PCV) ranging from 11% to 31% (mean, 18%; SD, ±6.8%; reference range, 33% to 58%). In three instances, the anemia was nonregenerative, based on low reticulocyte counts (reticulocytes, <1.5%); however, two of these dogs had IMHA, and the time of sampling was inadequate for a demonstrable bone marrow regenerative response. Coombs’ test results were positive in three of four dogs tested. Cytopathological interpretation of bone-marrow aspirates was available in five of eight anemic dogs. Three of these dogs had megakaryocytic and/or myeloid hyperplasia with normal erythroid maturation; two of which had peracute IMHA and one had a 2-month history of lethargy, cough and hypoadrenocorticism. One of the five dogs had erythroid hyperplasia, and one dog with a 2-year history of poly-arthritis and hyperglobulinemia had erythroid hypoplasia.
Neutrophilic leukocytosis was the most common leukogram abnormality (12 dogs; 50%), with six dogs also having a regenerative left shift (bands >0.3 × 103/μL). Values for increased neutrophils ranged from 12.0 to 51.5 × 103/μL (reference range, 3.0 to 11.5 × 103/μL), with a mean of 28.7 × 103/μL and a SD±13.3 × 103/μL. Monocytosis (monocytes >1.35 × 103/μL) was detected in eight dogs. Eosinophilia was observed in seven (29%) dogs, ranging from 0.77 to 5.536 × 103/μL (reference range, 0.1 to 0.75 × 103/μL), with a mean of 1.9 × 103/μL and a median of 1.4 × 103/μL. The dog with the eosinophil count of 5.536 × 103/μL (case no. 7; Tables 1, 2) had an eosinophilic leukocytosis, accompanied by normal numbers of neutrophils, lymphocytes, and monocytes, and was initially diagnosed with idiopathic epilepsy.
Nine (37.5%) dogs had increased serum alkaline phosphatase (ALP) activity, and three had concurrent increases in serum alanine transaminase (ALT) activity. The abnormal levels of ALP ranged from 168 to 1971 IU/L (reference range, 12 to 150 IU/L), with a mean of 603 IU/L and a SD±588 IU/L. Corticosteroids had been administered to four of nine dogs with high ALP activity within a month of testing and may have influenced their results.
Cerebrospinal fluid was analyzed in four dogs that were evaluated for neurological abnormalities. Considerable variation was found in nucleated cell count, protein content, and differential cell count among the four samples [Table 1]. The fourth dog (case no. 7; Table 2), which began having seizures around 6 months of age and that increased in frequency over 1.5 years to every 2 weeks, had an unremarkable cerebrospinal fluid analysis. A high reciprocal serum antibody titer (1024) to B. vinsonii (berkhoffii) antigens was found in this dog, however.
Joint fluid analysis revealed neutrophilic polyarthritis in one dog, lymphocytic polyarthritis in one dog, and mononuclear phagocytic inflammation of the hock joint (consistent with chronic synovitis) in one dog. The reciprocal IFA B. vinsonii (berkhoffii) titers for these three dogs were 256, 256, and 2048, respectively. The dog with the neutrophilic polyarthritis had an acute history of lameness, whereas the other two dogs were chronically lame.
Treatment and Outcome
In addition to other drugs, a variety of antibiotics were used in the medical management of these cases. In over half (62.5%) of the cases, more than one systemic antibiotic was administered. Antibiotics used included doxycycline (58%), enrofloxacin (33%), azithromycin (29%), clavamox (12.5%), cephalexin (12.5%), clindamycin (12.5%), amoxicillin (12.5%), tetracycline (4%), erythromycin (4%), and ciprofloxacin (4%). Only one dog (case no. 18; Table 3) did not receive systemic antibiotics.
For 21 of 24 dogs, clinical outcome could be qualitatively assessed from the medical records or by follow-up telephone inquiries. Following antibiotic treatment, clinical signs resolved completely in 12 (57%) dogs and substantial clinical improvement was reported in four (19%) dogs. In the three dogs with cutaneous vasculitis, all dermatological lesions resolved with treatment for Bartonella, except for persistent alopecia in one dog that resulted from extensive scarring of the original lesions. Following treatment with azithromycin, the dog diagnosed with idiopathic epilepsy (case no. 7) has not seizured for 18 months; however, the animal has been maintained on phenobarbital (3 mg/kg per os [PO] q 12 hours) during this 18-month period. Neurological deficits improved in a 13-year-old dog with meningomyelitis (case no. 13; Tables 1, 2), but the dog continued to have mild motor deficits in the hind limbs following treatment. No change in condition occurred in the dog with chronic otitis following topical treatment. Because of the severity of their condition, four (19%) dogs died or were euthanized. Of these, one dog diagnosed with granulomatous meningoencephalitis and two dogs with IMHA/ITP were dead within 1 week of onset of their illness. The fourth dog, which also had IMHA/ITP, died 4 months later. Initially, this dog was treated for <2 weeks with enrofloxacin, while receiving an immunosuppressive dosage of prednisone. One month later, azathioprine was added because of persistent anemia and thrombocytopenia. Although the cause of death was not established, melena, accompanied by anemia (PCV, 22%) and hypoproteinemia (total solids, 3.8 g/dL; reference range, 6.0 to 8.0 g/dL), was documented within 24 hours prior to death. Because of the retrospective nature of the study, the absence of treatment control groups, and the concurrent use of more than one antibiotic, no statistical correlations could be established between the treatment administered and the clinical outcome.
Of the 11 dogs that were retested at various time intervals following the initial detection of seroreactivity to B. vinsonii (berkhoffii) antigens, all were found to be seronegative (i.e., <16) [Table 2].
Discussion
Based upon the results of this study and previous studies,17–19 infection with B. vinsonii (berkhoffii) should be considered in the differential diagnosis of dogs with endocarditis, myocarditis, lymphadenitis, cutaneous vasculitis, meningoencephalitis, uveitis, IMHA/ITP, and polyarthritis. Although the dogs reported here satisfy serological criteria for exposure to B. vinsonii (berkhoffii), epidemiological studies are needed to establish causation of the diverse clinicopathological abnormalities observed. Perhaps the strongest support for a pathogenic role for B. vinsonii (berkhoffii) as a cause of the disease manifestations reported in these dogs was derived from the 11 dogs for which a convalescent titer was available. In all 11 dogs, clinical improvement was accompanied by a decrease in posttreatment B. vinsonii (berkhoffii) antibodies to nondetectable levels [Table 2].
In an effort to exclude the confounding effects associated with coinfection, dogs with serological evidence of exposure to other tick-borne pathogens were excluded from this study. Although selection criteria excluded dogs exposed to several tick-transmitted organisms and dogs with known noninfectious diseases, it is probable that other cofactors including duration of infection, concurrent infection with other unknown pathogens, or variation in the immunological response of individual dogs, contributed to the diversity of disease manifestations observed in these dogs. Experimental infection with B. vinsonii (berkhoffii) induces CD (cluster differentiation) 8+ lymphocytopenia, modulation of CD8+ adhesion molecule expression, defective bacterial phagocytosis by monocytes, and impaired B-cell antigen presentation.27 Therefore, the extent to which concurrent infection with B. vinsonii (berkhoffii) induces clinically relevant immunosuppression or complicates the diagnostic evaluation and therapeutic management of dogs infected with other tick-borne pathogens requires future study. However, Bartonella-induced immunosuppression in coinfected dogs may prove to be relevant, as 32% of the B. vinsonii (berkhoffii) seroreactive dogs tested by the NCSU-CVM-VBDDL between July 2000 and June 2001 had antibodies to at least one other tick-borne pathogen.
Seroreactivity to B. vinsonii (berkhoffii) antigens was found in 11.4% of samples submitted to the NCSU-CVM-VBDDL during the 1-year study period. Compared to Pappalardo, et al’s study22 that identified a 3.6% seroprevalence in sick dogs presented to the NCSU-VTH during an earlier 1-year period, the higher percentage of seropositive samples found in this study may reflect an increased awareness of the pathological potential of Bartonella spp. in dogs. Because many veterinarians test for a panel of tick-transmitted organisms rather than test for an individual organism, the sample population is also biased by the selection of dogs that were tested for concurrent exposure to other tick-transmitted pathogens. Previous epidemiological evidence strongly supports transmission of B. vinsonii (berkhoffii) by Rhipicephalus sanguineus.22 More recent studies using polymerase chain reaction (PCR) to detect organism-specific deoxyribonucleic acid (DNA), indicate that Ixodes spp. in California and Holland contain and may potentially transmit a variety of Bartonella spp. to animals and humans.2829
For the clinician, there is considerable overlap among the clinical and pathological conditions described in these dogs and the disorders attributed to other tick-borne pathogens. Half of the dogs in this study were thrombocytopenic, and one-third were anemic. Both of these abnormalities are commonly associated with infection of Babesia,30 Ehrlichia,31 or Rickettsia species.32 One-third of the dogs reported here had bone, joint, or muscle pain and hind-limb weakness, which can also be found in association with both Ehrlichia3334 and Rickettsia spp. infections.32 Similar to some (29%) of the dogs in this report, poorly localized pain involving bone, muscle, or joints is very consistent with contemporary reports81235 of human bartonellosis and with historical descriptions of Trench Fever in soldiers during World War I.36
Neurological dysfunction, which can also accompany ehrlichiosis,33 babesiosis,30 or Rocky Mountain Spotted Fever (RMSF),32 was the primary presenting complaint for four (16%) dogs. The mechanisms by which Bartonella spp. induce seizures, status epilepticus, or coma in humans are poorly understood.3637 Bartonella henselae can remain viable within in vitro cultures of feline microglial cells for at least 28 days without inducing ultrastructural abnormalities.37 Bartonella spp. may be able to cause chronic infection of the central nervous system (CNS), as suspected in the dog (case no. 7; Table 1) diagnosed with idiopathic epilepsy. The implication that Bartonella spp. can cause neurological dysfunction in dogs is consistent with recent evidence in people that Bartonella spp. may contribute to a broad spectrum of neurological manifestations36 and may complicate the clinical course of B. burgdorferi infection (i.e., Lyme disease) in humans.38
Three (13%) dogs had cutaneous lesions that were considered to be consistent with vasculitis. Leukoclastic vasculitis has been reported in a child with B. henselae infection.39 Vasculitis is a well-established component of RMSF.3132 However, unlike RMSF, which is an acute, fatal, or self-limiting disease, the duration of illness in the three dogs with cutaneous vasculitis in this report ranged from 4 to 7 months. These durations were more consistent with chronic intravascular infection and all three dogs exhibited resolution of the vasculitis following antimicrobial therapy.
In five (21%) dogs, clinical signs were attributed to cardiovascular dysfunction, further supporting previous evidence that infection with Bartonella spp. can contribute to cardiac abnormalities.141718 Although the aortic valve appears to be preferentially infected by Bartonella spp.,141718 one dog in this study was diagnosed echocardiographically with mitral valve endocarditis. In previous studies involving dogs diagnosed with endocarditis or myocarditis accompanied by serological, cultural, or molecular evidence of Bartonella infection, there was substantial variation in the duration of illness.1718 Similarly, over half of the dogs in this study were ill for variable periods of time, ranging from 1 month to 2 years. In many instances, owners reported very subtle abnormalities such as intermittent anorexia, lethargy, weight loss or lameness that persisted for months without overt abnormalities on physical examination. Like a previous study,18 fever was infrequently documented in these dogs and it was most often associated with lymphadenitis, bone and joint pain, or endocarditis. Collectively, these data suggest that B. vinsonii (berkhoffii) can cause chronic infections in dogs that may induce acute decompensation from underlying myocardial or CNS pathology.
This study also indicates that Bartonella infection should be considered in differential diagnosis of uveitis, retinal hemorrhage, and potentially retinal detachment in dogs. Although uveitis, retinal hemorrhage, and retinal detachment occur in dogs infected with Ehrlichia spp.40 or R. rickettsii,41 these reports predate recognition that dogs might be coinfected with B. vinsonii (berkhoffii).20–23 In humans, recent reports have suggested that Bartonella infection should be considered in any individual with posterior-segment ocular disease.5 Although a somewhat far-reaching recommendation, this statement indicates the rapidly increasing importance attributed to Bartonella spp. as a cause of ocular disease in people.
Anemia, thrombocytopenia, neutrophilia, monocytosis, and eosinophilia were prevalent hematological abnormalities in this study. Fifty percent of the dogs were thrombocytopenic, 50% had a neutrophilic leukocytosis, 33% were anemic, 33% had a monocytosis, and 29% were eosinophilic. To date, thrombocytopenia (generally mild in degree) is the most consistent abnormality reported in humans chronically infected with B. quintana.36 In contrast, infection with B. bacilliformis in people of South America results in a severe and, if untreated, frequently fatal hemolytic anemia.42 As the cause of IMHA in dogs is not often established,43 it is potentially relevant that at least five (21%) dogs had clinical or laboratory evidence of IMHA. Because two of the dogs died within the first week of management of IMHA, future studies should address the extent to which immunosuppressive therapy might interfere with antimicrobial elimination of the organism or potentiate the severity of illness in dogs with IMHA and occult Bartonella infection.
Although nonspecific findings, neutrophilic leukocytosis, monocytosis, or eosinophilia were identified in a substantial number of dogs in this and in a previous case series.18 Chronic eosinophilia, which fluctuated in severity, has also been reported in cats experimentally infected with B. henselae or B. clarridgeae by blood transfusion.44 From a hematological perspective, Bartonella infection should be considered in the differential diagnoses of unexplained anemia, thrombocytopenia, neutrophilia, monocytosis, or eosinophilia—particularly if accompanied by nonspecific clinical signs, a history of bone or joint pain, a chronic afebrile course, or a history of insect exposure.
Mild to moderate elevations of serum ALP activity and, to a lesser extent, serum ALT activity were observed in some of these dogs. Bartonella henselae has been implicated as a cause of peliosis hepatis in both dogs45 and humans,8 and as a cause of granulomatous hepatitis in people, particularly children.9 Recently, B. henselae and B. clarridgeiae DNA were detected in hepatic specimens from two dogs with liver disease.46 As the clinical relevance of these observations has yet to be established, the role of Bartonella spp. infection in causing elevations in liver enzyme activities or as a contributor to chronic hepatic disease in dogs deserves future research consideration.
The antibiotic of choice as well as the duration of treatment have not been clearly established for Bartonella spp. infection in cats,47 dogs, or humans.4849 Because azithromycin attains very high intracellular antimicrobial concentrations and the macrolides appear to be efficacious for the treatment of Bartonella infection, azithromycin was administered to several dogs in this study, with apparent success. Based upon clinical and hematological improvement or a decrease in posttreatment antibody titers [Table 2], however, several antibiotics may be efficacious. As B. vinsonii (berkhoffii) antibodies were not detected in any of the 11 posttreatment serum samples, posttreatment serological testing may be useful when attempting to document therapeutic elimination of Bartonella infection in dogs. The extent to which early diagnosis accompanied by aggressive antimicrobial therapy might have prevented mortality in the dogs with IMHA, ITP, or GME is unknown. Because Bartonella spp. are intracellular bacteria that can infect erythrocytes, macrophages, and endothelial cells, selection of an antibiotic (e.g., azithromycin, doxycycline, enrofloxacin) that achieves high intracellular concentrations is presumably beneficial.47–49 Despite an extended duration of illness in over half of the dogs in this report, complete therapeutic resolution was achieved following antibiotic treatment in nearly all survivors. Based upon a trial involving cats infected with B. henselae or B. clarridgeae,47 doxycycline may not be efficacious for treating Bartonella infections unless used at high doses for an extended duration.
Conclusion
Characterizing the spectrum of disease manifestations associated with Bartonella infection in humans has been a complex and difficult task, and it appears that defining the pathological consequences of acute or chronic Bartonella infection in dogs is equally challenging. This retrospective study indicates that further controlled prospective studies examining the role of Bartonella spp. as pathogens in dogs are warranted.
Snap 3DX; IDEXX Laboratories, Inc., Westbrooke, ME
Acknowledgments
The authors thank the numerous clinicians and diagnosticians that facilitated the medical evaluations of the dogs in this report. They also acknowledge the staff of the NCSU-CVM Vector Borne Disease Diagnostic Laboratory for facilitating this study as well as their enhanced understanding of tick-transmitted infectious diseases.
