|Home | About | Journals | Submit | Contact Us | Français|
A dog referred for lameness was diagnosed with culture-negative endocarditis. Antibodies to Bartonella spp. were detected. Antibiotic treatment resulted in transient clinical improvement, but the dog developed cardiac failure and was euthanized. Bartonella vinsonii subsp. berkhoffii genotype IV was identified within the aortic heart valve lesions by PCR amplification and DNA sequencing.
Endocardite à Bartonella vinsonii subsp. berkhoffii chez un chien de Saskatchewan. Un diagnostic d’endocardite, négative à la culture, a été posé chez un chien référé pour faiblesse. Des anticorps contre Bartonella spp. ont été détectés. Une antibiothérapie a conduit à une amélioration passagère mais l’animal a développé une insuffisance cardiaque et a été euthanasié. Bartonella vinsonii subsp. berkhoffii génotype IV a été identifié dans les lésions de la valvule aortique par ACP et séquençage d’ADN.
(Traduit par Docteur André Blouin)
A 9 year-old, intact male, German shepherd dog was referred to the Western College of Veterinary Medicine (WCVM) with a 2-month history of intermittent mild lameness involving the right hind and the left fore limb. There were no historical abnormalities other than the lameness and reluctance to exercise. The dog had never traveled outside Saskatchewan. Lameness had become more pronounced during the week prior to referral. Radiographs of the hips and elbows performed by the referring veterinarian had revealed right-sided hip dysplasia with mild arthritic change and degenerative changes in the left elbow, secondary to a fragmented coronoid process. A complete blood (cell) count (CBC) had revealed a leukocytosis characterized by a neutrophilia (segmented neutrophils 37.5 × 109/L; reference range, 4.80 to 13.9 × 109/L), left shift (bands 1.13 × 109/L; reference range, 0 to 0.1 × 109/L), and monocytosis (3.0 × 109/L; reference range, 0.08 to 1.0 × 109/L). A serum biochemical profile was normal. No treatment was initiated and the dog was referred to the WCVM for further evaluation.
On physical examination, the dog was in good body condition with mild generalized muscle atrophy. The dog was extremely agitated, and constant vocalization and panting made auscultation of the heart and lungs difficult. Body temperature was 39.7°C, femoral pulses were strong and regular (96 beats/min), mucous membranes were pink, and capillary refill time was normal (< 2s). Results from abdominal palpation were unremarkable, but the prostate gland was slightly enlarged, symmetrical, and nonpainful on transrectal examination. The dog walked with a stiff, stilted gait and resisted manipulation of both coxofemoral joints and both elbows. The left elbow joint was thickened with a decreased range of motion and palpable effusion.
Blood and urine were collected for analysis. The CBC results were similar to those obtained by the referring veterinarian 1 wk earlier (Table 1). Hyperproteinemia was present (total solids 83 g/L; reference range, 56 to 74 g/L). An analysis of urine collected by cystocentesis was normal and the urine culture was negative. A serum biochemical profile was not repeated.
The dog was sedated with acepromazine (Atravet; Wyeth-Ayest, Guelph, Ontario), 0.05 mg/kg bodyweight (BW), IM, and hydromorphone (Hydromorphone HP 10; Sabex, Quebec), 0.1 mg/kg BW, IM, for diagnostic imaging. Thoracic radiographs revealed mild generalized cardiomegaly and left atrial enlargement with no extracardiac signs of heart failure. Abdominal radiographs and abdominal ultrasonographs did not identify any abnormalities.
Auscultation was repeated while the dog was sedated; an arrhythmia and a grade III/VI diastolic cardiac murmur, loudest over the left side of the heart base were noted. Atrial premature contractions (APCs) were detected on an electrocardiogram (ECG). Indirect blood pressure measurements were within normal limits (mean 110 mmHg).
Echocardiography showed moderate irregular hyperechoic thickening of all aortic valve leaflets, most consistent with endocarditis. The left ventricle was slightly dilated and the left atrium was moderately dilated. There was also diffuse thickening of the cranial mitral valve leaflet, consistent with either endocardiosis or early endocarditis. There was aortic valve regurgitation during diastole and mitral valve regurgitation during systole.
Three 10-mL blood samples were collected aseptically at 30-min intervals from alternate jugular veins and submitted for standard aerobic and anaerobic culture and for Bartonella spp. culture. The Bartonella spp. culture was incubated in 7% CO2 for 30 d. Culture results were negative.
Synovial fluid from the right and left carpi, elbows, stifles, and hocks was obtained by arthrocentesis. Increased cellularity was detected in the synovial fluid from the right elbow, the right stifle, and the left stifle. Mononuclear cells predominated in the right elbow and the right stifle, consistent with degenerative joint disease. Segmented, nondegenerate neutrophils predominated in the left stifle. Synovial fluid from this joint was cultured for aerobic and anaerobic bacteria and was negative.
The dog was diagnosed with culture-negative infective endocarditis and secondary polyarthritis. Treatment was initiated with cephalexin (Novo-Cephalexin; Nu-Pharm, Richmond Hill, Ontario), 22 mg/kg BW, PO, q8h.
When reevaluated 2 wk later, the dog’s attitude, physical status, and lameness had improved. Results from a CBC were essentially unchanged (Table 1). Femoral arterial pulses were strong and regular prior to the dog being sedated, but after being sedated with hydromorphone (Hydromorphone HP 10; Sabex), 0.1 mg/kg BW, IM, and diazepam (Valium; Sabex), 0.25 mg/kg BW, IM, a diastolic murmur and arrhythmia were apparent, and an electrocardiogram (ECG) revealed APCs and ventricular premature contractions (VPCs). Thoracic radiographic images and echocardiogram readings were unchanged. Blood cultures were repeated and were negative. Enrofloxacin (Baytril; Bayer, Toronto, Ontario), 3.6 mg/kg BW, PO, q24h, and metronidazole (Apo-Metronidazole; Apotex, Toronto, Ontario) 12.0 mg/kg BW, PO, q12h, were added to the treatment protocol. Meloxicam (Metacam; Boehringer-Ingelheim, Burlington, Ontario), 0.1 mg/kg BW, PO, q24h, was also administered to manage the lameness.
When reevaluated 6 wk later (8 wk after initial presentation), the dog had gained 2 kg and the lameness had resolved. The heart murmur was unchanged and an ECG showed APCs and VPCs. A CBC revealed a mild monocytosis, but the neutrophilia had resolved (Table 1). Thoracic radiographs and echocardiography revealed modestly increased dilation of the left ventricle and atrium, but the valvular lesions and the magnitude of the aortic and mitral regurgitation were unchanged. The multianti-biotic protocol was continued and the dog was reevaluated every 6–8 wk. The dog remained clinically normal for approximately 6 mo, with no reported lameness, exercise intolerance, or systemic signs of illness.
Approximately 8 mo after initial presentation, the dog developed early congestive heart failure, characterized by weight loss (3 kg), cough, and exercise intolerance. Radiographs of the thorax revealed increased left atrial prominence, pulmonary venous congestion, and mild perihilar edema. Echocardiography confirmed that the left atrial dilatation was dramatically worsened and the left ventricular fractional shortening reduced. The aortic valvular lesions were enlarged, although the magnitude of aortic regurgitation was unchanged. Results from a CBC revealed normal neutrophil numbers, a monocytosis and severe thrombocytopenia (Table 1). A coagulation profile consisting of a prothrombin time (PT), partial thromboplastin time (PTT), and fibrin degradation products (FDPs) was normal. Treatment was initiated with furosemide (Apofurosemide; Apotex), 1 mg/kg BW, PO, q12h, and benazepril (Fortekor; Novartis Animal Health, Mississauga, Ontario), 0.25 mg/kg BW, PO, q24h. The multiantibiotic regime was continued. The meloxicam was discontinued and 5 d later, daily aspirin therapy was initiated (Acetylsalicylic acid; Vita Health Products, Winnipeg, Manitoba), 5 mg/kg BW, PO, q24h, to treat presumed embolic disease.
Serum was submitted to the Intracellular Pathogens Research Laboratory, College of Veterinary Medicine, North Carolina State University for determination of antibody titers to Bartonella spp. Serologic testing for antibodies to other tick-borne agents of diseases associated with thrombocytopenia was not done, as Saskatchewan is not considered an endemic region for these diseases. Reciprocal antibody titers to B. henselae and B. vinsonii subspecies berkhoffi were 2048 and 512, respectively (titers greater than 1:64 indicate prior exposure or current infection). Blood submitted to Healthgene (Toronto) for polymerase chain reaction (PCR) for Bartonella DNA was negative. A presumptive diagnosis of Bartonella endocarditis was made. Azithromycin (Zithromax; Pfizer Canada, Toronto, Ontario), 5.25 mg/kg BW, PO, q24h for 7 d and then q48h for 6 wk was added to the antibiotic treatment. As well, because of continued deterioration in cardiac function, pimobendan (Vetmedin; Boehringer Ingelheim), 0.25 mg/kg BW, PO, q12h, was administered.
The dog’s cardiac status continued to decline. The dog also developed progressive rear limb weakness and incoordination, with slightly decreased conscious proprioception and normal spinal reflexes. The dog returned to the WCVM for euthanasia approximately 10 mo after the initial presentation. Immediately following the euthanasia, the heart was removed aseptically, opened, and fixed in formalin. The fixed heart and the remainder of the carcass were submitted for postmortem examination. The heart was large with dilatation of ventricles and auricles, and hypertrophy of the right and left ventricular walls. Severe endocarditis was present, with mineralization affecting the aortic and mitral valves. The aortic opening was occluded and the septal and right semilunar valvulae were covered by firm, mottled white, yellow, and red plaques. Attached to the left valvula was a white gritty mass measuring 1 × 0.5 × 0.5 cm. On the surface of the mitral cusps were multiple small (up to 1 mm diameter) yellow and white firm nodules. There was a 2-cm diameter focal area of hemorrhage in the left myocardium. Severe pulmonary edema and ascites consistent with left- and right-sided heart failure, respectively, were present. There were multiple depressed wedge-shaped infarcts in the renal cortices. The right hip and both elbow joints showed moderate to severe degenerative changes. Microscopically, the aortic valvulae were infiltrated at the base by moderate numbers of neutrophils, fewer lymphocytes, and numerous hemosiderinladen macrophages. The core of the valvulae was expanded by collagen, amorphous mineralized material, and mucin, and the free edges of the valves showed hypertrophy of endothelial cells, and prominent neovascularization. The surface of the mitral valve was eroded and multifocally covered by amorphous mineralized material. There was multifocal acute and chronic infarction of the myocardium associated with embolization of mineralized material in the coronary arteries. Neutrophilic inflammation was associated with necrotic myofibers in acute infarcts. Light microscopy confirmed moderate chronic passive congestion of the lungs, acute passive congestion of the liver, multiple chronic renal infarcts, and chronic mild arthritis. Bone marrow was normal and there was multifocal mild to moderate thoracic and lumbar degenerative myelopathy. Samples of the formalin-fixed paraffin embedded aortic and mitral heart valves were submitted for Bartonella PCR to the Intracellular Pathogens Research Laboratory, College of Veterinary Medicine, North Carolina State University. Bartonella vinsonii subsp. berkhoffii DNA was amplified and sequenced from the aortic valve. Bartonella henselae DNA was not amplified.
Bartonella spp. comprise a group of small gram-negative bacilli that cause chronic intraerythrocytic and endothelial cell infections (1). With the exception of B. bacciliformis and B. quintana, most Bartonella spp. have been isolated and identified as human pathogens only since the early 1990s. In 1993, B. quintana, B. elizabethae, and B. henselae (the usual agent of cat scratch disease) were identified for the 1st time as causal agents of endocarditis in humans; since then, Bartonella infection has become known as an important cause of culture-negative endocarditis in humans (2). Most human cases of endocarditis are related to infection with B. quintana and B. henselae.
In 1993, the subspecies B. vinsonii subsp. berkhoffii was first isolated and then classified from the blood of a dog with intermittent epistaxis and endocarditis (3,4). Although sero-conversion and, rarely, disease can occur following infection with other Bartonella spp. (5–8), B. vinsonii subsp. berkhoffii has been implicated on a molecular basis in most cases of clinical disease in dogs, with endocarditis being the most common manifestation reported to date (9–12). Bartonella vinsonii subsp. berkhoffii has been associated with a wide variety of other manifestations, including, myocarditis, fever, granulomatous lymphadenitis, rhinitis, epistaxis, cutaneous vasculitis, anterior uveitis, and immune mediated hemolytic anemia (6,9,13–16). Lameness attributed to polyarthritis has been reported in dogs with and without endocarditis (6,12,13,15). Thrombocytopenia, anemia, neutrophilic leukocytosis, monocytosis, and eosinophilia are hematologic findings reported in dogs seropositive to B. vinsonii subsp. berkhoffii antigens (1,10,15). In 2000, B. vinsonii subsp. berkhoffii was isolated from an endocarditis lesion on a heart valve from a Portuguese man, raising new concerns about the zoonotic potential of this organism (17). Recently, B. vinsonii subsp. berkhoffii DNA was detected in the blood of non-immunocompromised veterinary professionals, further increasing this concern (18).
Current data suggest that B. vinsonii subsp. berkhoffii is transmitted throughout much of the USA and in most tropical and subtropical regions of the world, but seroprevalence data are limited (1). In a 1998 study, approximately 3.6% of sick dogs from North Carolina and Virginia were found to have antibodies against this organism (19); heavy flea or tick infestation and living in a rural environment were risk factors associated with positive antibody status (19,20). In similar studies, 2% of sick dogs in Northern California (13) and 2% of dogs from Rhode Island (21) had antibodies against B. vinsonii subsp. berkhoffii. Seroprevalence was higher (8.7%) in healthy U.S. Government-owned dogs, with the highest prevalence in dogs from the southern and plains regions of the USA (22). In Morocco, a much higher seroprevalence was reported for stray dogs (38%) than for pet dogs (4%) (23). The relatively low seroprevalence and chronic bacteremia in dogs makes the role of pet dogs as an important reservoir of infection for B. vinsonii subsp. berkhoffii uncertain. Coyotes (Canis latrans) have been proposed as the primary natural wildlife reservoir, with 35% to 76% of serum samples from coyotes in California containing antibodies to B. vinsonii subsp. berkhoffii antigens (24,25) and as many as 28% being bacteremic with this organism at the time of sampling (6,24,25). Gray foxes (Urocyon cinereoargenteus) may also serve as a reservoir host for B. vinsonii subsp. berkhoffii genotype III (26).
No seroprevalence data are available for B. vinsonii subsp. berkhoffii in western Canada, though the prevalence of all tick-borne diseases in this region is very low. A serologic survey performed in 1995 in southern regions of Saskatchewan and Alberta found low seroprevalence of B. henselae and B. clarridgeiae infection in cats (17.8%, 18.6%), but no evidence of Bartonella spp. infections in dogs (27). The prevalence of dedicated rodent strains of Bartonella in ground squirrels in Saskatchewan, thought to be transmitted by fleas, is high (> 50%) (28).
Bartonella vinsonii subsp. berkhoffii is presumed to be transmitted to dogs by the bite of an infected tick (9). In some regions of the USA, dogs naturally infected with Ehrlichia canis or Babesia canis are likely also to be seropositive for B. vinsoni (36% and 52%, respectively), suggesting that their tick vector (Rhipcephalus sanguineous) may also be involved in the transmission of B. vinsonii subsp. berkhoffii (1,20). Ixodes scapularis ticks in the USA are known to be infected frequently with rodent Bartonella spp. (1,29). Epidemiologic evidence has also implicated Dermacentor and Amblyomma americanum ticks in the transmission of B. vinsonii subsp. berkhoffii (1). More than 95% of the ticks identified on dogs in Saskatchewan are D. variabilis, although small numbers of D. andersoni can be found in the western part of the province. There is a small population of mature I. scapularis, but most of these ticks are thought to be adventitious and carried into the province by migratory birds, as the temperatures are too cold in winter to support a self-sustaining population (30). Rhipcephalus sanguineous ticks have not been identified in Saskatchewan. The dog in this report lived in a rural area and had been known to periodically have both attached ticks and fleas.
Infective endocarditis is defined as infection of the endothelial surface of the heart, almost always affecting the aortic or mitral valve in affected dogs (31–33). Bacteremia is required for the development of infective endocarditis, but an underlying reason for transient or chronic bacteremia is rarely identified in dogs or humans (12,32). Structural abnormalities of heart valves and immunosuppression may make an individual more susceptible (32). The most commonly reported etiologic agents of infective endocarditis in dogs are staphylococci, streptococci, gram-negative bacilli, and Bartonella spp. (11,12,33,34).
Infective endocarditis is most common in middle-aged (4–8 y), medium to large breed, male dogs (11,33). Clinical manifestations typically include fever, lethargy, weight loss, and lameness (3,12). Some dogs will be presented in acute fulminating cardiac failure (35). Physical examination of dogs with endocarditis usually reveals a cardiac murmur, which may be of recent onset (12,31–34). Arrhythmias occur in 39% to 50% of affected dogs and can indicate cardiac decompensation, concurrent infective myocarditis, or regional myocardial ischemia caused by embolization of vegetative lesions (9,10,12,34). Fever has been reported to occur in 38% to 72% of dogs with infective endocarditis, but it may be less common in dogs with infective endocarditis caused by Bartonella than with other organisms (10,12,34). Protein losing nephropathy or azotemia occurs in 28% to 60% of affected dogs, regardless of the causative organism (12,21,33). Lameness, a stilted gait, or joint swelling is common in dogs with infective endocarditis due to the presence of immune-complex mediated “reactive” polyarthritis or, less commonly, septic arthritis (11,12,15,33,34). The dog reported here was presented for veterinary evaluation because of polyarticular lameness that, initially, was presumed to be related to chronic orthopedic abnormalities and degenerative joint disease, but an inflammatory leukogram prompted referral for further evaluation.
The most common CBC finding in dogs with infective endocarditis is a mature neutrophilia, although some dogs will also have an increase in band neutrophils and monocytes (10,12,33). The magnitude of the inflammatory response is not different between dogs with Bartonella-induced endocarditis and endocarditis due to other infectious agents (11,12,35). More than 50% of dogs will be thrombocytopenic, regardless of the causative organism, a finding usually attributed to a consumptive coagulopathy, septic embolization, or endotoxin-induced microvascular injury (10–12,33,34).
Diagnosis of endocarditis should be suspected in any dog with suggestive clinical signs and a new or progressive heart murmur (12,34). The diagnosis is confirmed by echocardiographic identification of a typical vegetative or destructive lesion on the mitral or aortic heart valve (12,32,34). The mitral valve is more commonly affected than the aortic valve in dogs with infective endocarditis caused by organisms other than Bartonella spp. (10,12,34), but in most (> 90%) Bartonella-infected dogs, the aortic valve is affected either alone or together with the mitral valve (9–12,34), Blood cultures will reveal the causative organism in 40% to 80% of non-Bartonella-infected dogs, but they will rarely be positive in Bartonella-infected dogs. Recently, approximately 20% to 50% of the culture-negative infective endocarditis cases in dogs in California have been reported to result from Bartonella spp. infection (11,12,34). Histologically, examination of heart valves from dogs with Bartonella endocarditis reveals a combination of fibrosis, mineralization, endothelial cell proliferation, and neovascularization that is distinct from endocarditis caused by culturable bacteria (11).
Attempts to culture B. vinsonii subsp. berkhoffii from the blood of healthy or sick dogs with serologic or molecular evidence of infection are often unsuccessful (1,6,15). Serology and PCR amplification of Bartonella DNA are the mainstays of diagnosis. Seroprevalence is relatively low, even in endemic regions, so detection of antibody in a sick dog provides strong clinical evidence for prior exposure and active infection (13,15,19,21,22). A titer of 1:64 or greater is considered positive (1). Antibody detection by means of indirect fluorescent antibody assays is reliable, but coinfection with multiple Bartonella spp. has been reported and there may be some cross-reactivity between B. henselae, B. clarridgeiae, and B. vinsonii subsp. berkhoffii (12,13). It also appears that antibodies to B. vinsonii subsp. berkhoffii will disappear rapidly in antibiotic-treated dogs, which may well complicate serodiagnosis, depending on the timing of sample collection. Persistent release of Bartonella into the systemic circulation from infected heart valves most likely resulted in chronic antigenic stimulation and a detectable antibody titer in this dog, despite long-term administration of antibiotics.
An advantage of using PCR for diagnosis is that the DNA products can be sequenced, allowing precise identification of the species and strain of Bartonella involved (6). Based on 16S-23S intergenic spacer sequence differences, a classification scheme was proposed in 2006 to divide the B. vinsonii subsp. berkhoffii subspecies isolates into distinct genotypes (26). This typing scheme hopefully will allow tracking of geographic distribution, and determination of natural reservoir hosts, and will assist in clarifying modes of transmission. To date, 4 B. vinsonii subsp. berkhoffii genotypes have been identified. Genotype II is most prevalent in dogs and humans in the USA and has also been identified in coyotes in that country. Genotype I, isolated from a dog with endocarditis, has rarely been identified in dogs or coyotes in the USA. Genotype III has been identified in grey foxes in California, in a human endocarditis patient from Europe, and, recently, from the blood of a military working dog with endocarditis (Breitschwerdt, unpublished data). A unique B. vinsonii subsp. berkhoffii strain, genotype IV, was isolated from the dog in this report (26).
To date, an optimal protocol has not been established for the treatment of Bartonella infections in cats, dogs, or humans (1,6). A variety of antibiotics are used in humans with Bartonella endocarditis, but the mortality rate is high and most patients require heart valve replacement. Bartonella is an intracellular bacterium that infects erythrocytes, macrophages, and endothelial cells, so selection of an antibiotic such as a macrolide that achieves high intracellular concentrations is recommended — azithromycin is considered the oral antibiotic of choice, although fluoroquinolones may also have some effect (1,6,14,15). Doxycycline, although often recommended, has been associated with treatment failures and relapses in dogs and cats with Bartonellosis (1,6,14). Prolonged treatment (4–6 wk) is recommended in all dogs with presumed Bartonella infections (6). Therapeutic elimination of Bartonella organisms from endocarditis lesions may be impossible.
Consequences of infective endocarditis include the development of heart failure and cardiac arrhythmias. Thromboemboli commonly cause regional infarction of the spleen, kidneys, myocardium, brain, and skeletal muscle (11,33). The median survival time for dogs with infective endocarditis of any etiology has been reported to be 54 d (33). In a recent study, 80% of dogs with endocarditis due to Bartonella spp. died within 2 wk of diagnosis (median survival 3 d, mean survival 45 d), while dogs with endocarditis caused by other bacteria lived for 11–12 mo (median 330 d, mean 375 d) (12).
This case illustrates the importance of being aware of emerging infectious diseases, even in regions presumed to be non-endemic, and identifies B. vinsonii subsp. berkhoffii genotype IV as the cause of culture-negative endocarditis for the first time in Canada. The reservoir, geographic distribution, and mode of transmission for this unique genotype requires additional study.
We thank Dr. Karen Wagener at the Balgonie Veterinary Clinic for her insightful initial referral of this case and for her diligence in providing ongoing medical care. CVJ