Introduction

Infective endocarditis (IE) can be challenging to definitively diagnose antemortem, and because of the frequency, it is commonly missed as a differential. IE of the pulmonic valve is extremely rare, and when it has been reported, the pulmonic valve is historically affected in combination with other valves.¹ In a retrospective case series, a single case of IE involving only the pulmonic valve was reported, but no details were given, and the bacterial species was not identified.² The current case is unique in that the pulmonic valve is the only valve affected, and it is the first report of bacterial endocarditis in a dog in which Empedobacter brevis was identified by blood culture.

Case Report

An 11 yr old female spayed golden retriever weighing 30.3 kg was referred to BluePearl Veterinary Partners for evaluation for lethargy, anorexia, tachypnea, stiff gait, and nonlocalized pain that had been progressive over the previous 3 wk. Initial diagnostics with the primary care veterinarian revealed thrombocytopenia, mild neutrophilia, and mild nonregenerative anemia. The dog was treated supportively with pain medication (carprofen, tramadol) and antibiotics (ciprofloxacin) and sent home for continued monitoring. A week later, the patient was referred to the specialty hospital for persistent stiff gait, worsening lethargy, and continued anorexia in addition to the laboratory abnormalities.

At the time of presentation, the owners reported that the patient had been walking with a stiff, stilted gait at home and seemed uncomfortable when ambulating. The appetite was currently decreased, and a mild cough was noted during physical exam, although this was not a prominent part of the history. Vaccination for distemper, leptospirosis, parvovirus, and rabies were current according to records from the primary veterinarian.

On physical examination, the patient was febrile (40.2°C, 104.4°F), tachycardic (168 beats per min), and depressed. The body condition score was estimated at five out of nine, mucous membranes were tacky but pink, and the capillary refill time was <2 s. No cardiac murmurs were appreciated, but an arrhythmia was noted with pulse deficits. An electrocardiogram showed a sinus rhythm with uniform ventricular premature complexes and periods of ventricular bigeminy. No joint pain or effusion was appreciated on exam. However, based on the history of progressive stiffness with an abnormally wary gait, fever, and the clinical progression, polyarthritis was suspected in this case, with infectious or autoimmune disease as possible underlying etiologies. The cause for the arrhythmia was undetermined at this time.

Hematologic abnormalities at presentation included normocytic, normochromic, nonregenerative anemia (hematocrit [Hct] 29.2%), leukocytosis (36.41 × 10³/μL; reference range, 5.05–16.76 × 10³/μL) characterized by neutrophilia (30.29 × 10³/μL; reference range, 2.95–11.64 × 10³/μL) and monocytosis (3.28 × 10³/μL; reference range, 0.16–1.12 × 10³/μL), and thrombocytopenia (23 × 10³/μL; reference range, 148–484 × 10³/μL). The blood chemistry panel, including electrolytes, was normal with the exception of mild hypoalbuminemia (2.1 g/dL; reference range, 2.2–3.9 g/dL). A urinalysis showed urine specific gravity of >1.050, hematuria (4+ on the urine strip), and proteinuria (3+ on the urine strip). A urine culture was not submitted at this time.

Thoracic radiographs showed a mild diffuse bronchointerstitial pattern with patchy alveolar infiltrate, scant pleural effusion or pleural thickening, and possible right caudal lung lobe mass (Figure 1). Cardiovascular structures were normal, and there was no evidence of thoracic lymphadenopathy. Abdominal ultrasound was unremarkable, except for mild bilateral adrenal gland enlargement (largest diameter 0.91 cm left adrenal, 0.82 cm right adrenal) that was considered unlikely to be related to the current clinical signs. A tick-borne disease polymerase chain reaction panel with Lyme titers was submitted, which ultimately came back negative. The initial most likely differentials for the fever included tick-borne infection, autoimmune disease such as immune-mediated polyarthritis, and neoplasia (lymphoma, other). The diffuse changes in the lungs were suspected to be vasculitis in combination with degenerative aging changes; however, atypical pneumonia, hemorrhage, or neoplastic process were considered as well. Differentials for the ventricular arrhythmia included primary heart disease, systemic disease causing secondary arrhythmia, myocarditis, or endocarditis. An echocardiogram was recommended but declined by the owner. The patient was scheduled to undergo general anesthesia for thoracic computed tomography, joint taps, and bone marrow biopsy the next day. However, a productive cough developed overnight with frothy, blood-tinged sputum noted the following morning. With the progressive cough overnight, ventricular arrhythmia, and undetermined cause for the lung pattern, there was additional concern at this time for heart failure despite the seemingly normal cardiac silhouette on radiographs the day before. At this point, an echocardiogram was strongly recommended again prior to anesthesia, and the owner consented. The echocardiogram revealed irregular proliferative lesions on the pulmonic valve leaflets, moderate pulmonic regurgitation, and trivial tricuspid regurgitation (Figure 2). Systolic function was within normal limits. Based on the appearance of the pulmonic valve, primary differentials included IE, thrombus, severe valvular endocardiosis, or neoplasia. However, given the context of the case, IE was considered the most likely differential, and the originally planned diagnostics were not performed.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Three blood samples were collected aseptically, each 30 min apart, from different veins (left jugular, left saphenous, and left cephalic) and were submitted for bacterial aerobic and anaerobic blood cultures. The cause for the pulmonary infiltrates and cough was never definitively determined but was likely due to systemic inflammatory response and vasculitis, although hemorrhage from thrombocytopenia or pneumonia remained possible.

IV antibiotic therapy was initiated after all blood samples were collected, consisting of ampicillin-sulbactam^a (30 mg/kg q 8 hr) and enrofloxacin^b (10 mg/kg q 24 hr). The patient was placed in the intensive care unit on maintenance fluid therapy, constant electrocardiogram, and monitoring of the respiratory rate and effort. Within the next 12 hr, the fever had resolved (38.06°C, 100.5°F) and the frequency of the ventricular arrhythmia had decreased from 28 ventricular premature complexes per min to 8 per min within a 5 hr period. By the following morning, the ventricular complexes were only noted once in a 12 hr period. Recheck labwork showed worsening anemia (Hct 21.2%), improved leukocytosis (30.98 × 10³/μL; reference range, 5.05–16.76 × 10³/μL), and persistent thrombocytopenia (9 × 10³/μL; reference range, 148–484 × 10³/μL). Partial thromboplastin time and prothrombin time were normal. The patient remained in the intensive care unit for the next 48 hr and was then discharged for continued therapy at home due to financial constraints of the owner. The patient was sent with an IV catheter in place with detailed instructions for catheter care and administering intravenous antibiotic therapy at home. Intravenous therapy with enrofloxacin and ampicillin-sulbactam was continued at home for 2 wk under the supervision of the primary care veterinarian. Continued intravenous therapy was recommended as opposed to switching to subcutaneous administration to increase the concentration of the antibiotics in the bloodstream.

The blood cultures grew alpha-hemolytic streptococci and E brevis, previously known as Flavobacterium brevis, individually in separate blood culture samples. The susceptibility profiles were slightly different between these organisms, with the streptococci spp. showing in vitro resistance to enrofloxacin, marbofloxacin, and amikacin, whereas E brevis showed susceptibility to all tested antibiotics, including enrofloxacin and amoxicillin. The dog was rechecked by the referring veterinarian after 2 wk of intravenous therapy as recommended for bloodwork and physical exam. The hematologic abnormalities had improved and thrombocytopenia resolved. At this time, the patient was transitioned to oral amoxicillin, discontinuing the enrofloxacin due to the susceptibility profile. During this visit, the owner reported a good appetite and energy level. The tachypnea, stiffness, and coughing had also resolved at this time.

Approximately 4 wk later, the patient presented again to the specialty hospital for 1 wk duration of lethargy and decreased appetite. The patient was still receiving oral amoxicillin (13 mg/kg) twice daily. Physical exam was normal aside from a fever (39.9°C, 103.9°F). Hematologic abnormalities included normocytic, normochromic, nonregenerative anemia (Hct 31.5%) and leukocytosis (29.19 × 10³/μL; reference range, 5.05–16.76 × 10³/μL) characterized by neutrophilia (24.54 × 10³/μL; reference range, 2.95–11.64 × 10³/μL) with regenerative left shift and monocytosis (2.49 × 10³/μL; reference range, 0.16–1.12 × 10³/μL). The platelet count was normal. The blood chemistry panel, including electrolytes, was normal with the exception of mild hyperchloremia. Electrocardiogram showed rare uniform ventricular premature complexes, and recheck thoracic radiographs showed resolution of the pulmonary infiltrates with scant pleural effusion. There was no evidence of the previously suspected pulmonary mass at this time. An echocardiogram showed similar findings as the previous exam. Blood cultures were submitted again, and the patient was discharged on oral enrofloxacin (4.5 mg/kg) once daily in addition to the amoxicillin. The cultures ultimately were negative. However, the patient had a significantly improved appetite and energy level after starting the additional antibiotic, and the fever had resolved. Therefore, therapy with both enrofloxacin and amoxicillin was recommended long-term (minimum 6 mo). The patient recently presented (about 10 mo after initial diagnosis) to the specialty hospital for an unrelated problem, and the owner remarked that the dog had continued to be clinically normal at home with no systemic signs of illness. The owner was still giving the oral enrofloxacin but had discontinued the amoxicillin on his own.

Discussion

IE is a rare disease in small animal veterinary medicine characterized by vegetative lesions of the heart valves. The most recently reported incidence rates <0.1%, and it is most common in middle-aged to older male large-breed dogs.^3–6 Pathophysiology is complex, but bacteremia is necessary at some point in the disease process for the development of bacterial IE.^4,5 In this case, the inciting cause for bacteremia was never determined. The dog had not had any recent hospitalization or catheter placement. There was no history of immunosuppressive medication, IV drugs, or evidence of systemic disease prior to the recent clinical signs. Additional diagnostics to investigate possible sources of bacteremia were limited due to finances and the need to pursue long-term antibiotic therapy. The most commonly affected valves in dogs are the aortic and mitral valves.^3,4 Variable presenting clinical complaints include fever, lethargy, inappetence, weakness, lameness, stiffness, mentation changes or neurologic signs, and occasionally cough.^3,4 Potential complications of IE include thromboembolic disease, congestive heart failure, cardiac arrhythmias, and consequences of bacteremia, such as septic arthritis, splenic abscesses, and septic emboli. Various autoimmune diseases (immune-mediated polyarthritis, glomerulonephritis, vasculitis) have also been associated with IE.^3–5 This patient presented originally for fever, stiff gait with nonspecific pain, weakness, and lethargy. Although these signs are fairly nonspecific, immune-mediated polyarthritis was initially suspected based on the history and clinical findings, as this is a common diagnosis for fever of unknown origin in dogs.⁷ Although joint taps would have provided additional information as to the cause of the stiff gait, they were not performed in this case due to the severity of progressive respiratory signs and concern for sedation. Although we cannot confirm polyarthritis in this patient, polyarthritis and IE can coexist, and both can present with nonspecific signs and fever.

Antemortem definitive diagnosis of IE is difficult. Because this patient was still alive at the time of this report, obtaining a definitive diagnosis of IE at necropsy was not an option. In 1994, a series of major and minor criteria were proposed for the antemortem diagnosis of IE in human medicine.⁸ This system has become known as the Duke criteria and has been validated by previous studies.^8,9 The Duke criteria has since been adapted for veterinary medicine for use in obtaining a definitive diagnosis of IE.^5,10,11 A case qualifies for an antemortem definitive diagnosis of IE if it meets at least two major criteria or at least one major and two minor criteria. Major criteria include two or more positive blood cultures with identification of a typical organism and positive echocardiography findings supportive of IE, such as vegetative lesions or new valvular insufficiency. Minor criteria include predisposing heart disease, the presence of a new cardiac murmur, fever, presence of vascular complications including septic pulmonary infarcts and arterial emboli, immunologic disease including polyarthritis or vasculitis, a positive blood culture, large-breed dogs (>15 kg), and known predisposing causes to bacteremia such as indwelling catheter, repeated IV drug administration, and immune-compromised patients. Using the Duke criteria, this case fulfills two major criteria (vegetative valvular lesion, new valvular insufficiency) and at least three minor criteria (fever, >15 kg, positive blood culture).¹¹

IE of the pulmonic valve is very rare in veterinary medicine.^1–9 Most reported cases of pulmonic valvular bacterial endocarditis were noted in combination with additional affected valves and/or congenital defects.^1,2,6 To the authors' knowledge, only one case of pulmonic valvular bacterial endocarditis in the absence of congenital anomalies and concurrent valve involvement in a dog is mentioned in the literature. This case was reported as part of a retrospective analysis through the University of Pennsylvania, and other details of the case were not given.² The current case report is an older dog who had no history of a heart murmur or other signs of congenital disease. The echocardiogram did not show evidence of pulmonic stenosis or other definitive congenital heart disease, and the remaining valves were normal with the exception of trivial regurgitation through the tricuspid valve. Although we cannot completely rule out very early involvement of the tricuspid valve given the regurgitation without a necropsy, there was no evidence of vegetative lesions on the tricuspid valve on echocardiogram. Additionally, unlike the mitral and aortic valves, which are commonly affected together in patients with bacterial endocarditis, the pulmonic and tricuspid valves are not in fibrous continuity. The tricuspid valve is also upstream from the pulmonic lesion, so hematologic spread from a primary lesion on the pulmonic valve is not considered likely.

IE is most commonly caused by bacterial infection in veterinary medicine, although rickettsial and fungal endocarditis have been reported.^4,10 Greater than 50% of cases of bacterial endocarditis in dogs are due to staphylococci spp. and streptococci spp. infections, but several other bacterial species have been reported.^1–4,10 The current case is the first report of E brevis bacteremia in a dog. There is one report of postsurgical meningitis in the dog involving this organism isolated in significant numbers in the cerebrospinal fluid; however, blood cultures were not performed.¹² This gram-negative, short, nonmotile rod is widely found in the environment, typically in the soil and water sources. It is most commonly involved in suspect nosocomial infections in humans.^12,13 When performing cultures, colony-forming units can be helpful to determine the likelihood of a contaminant versus a pathogenic organism. The blood cultures in this case were performed using broth media; therefore, quantitative analysis was not possible. However, given that this is the second report of E brevis isolation from a dog with systemic infection, it is the authors' opinion that this organism should be considered a relevant veterinary pathogen.¹²

Treatment for bacterial endocarditis consists of appropriate IV antibiotic therapy for 1–2 wk, ideally guided by culture and sensitivity results, followed by transition to subcutaneous and/or oral administration.^3–5 Total antibiotic duration is variable, and 3–6 mo up to a year of therapy is recommended in some cases.⁶ This patient initially improved with antibiotic therapy with amoxicillin and enrofloxacin, only to have recurrent clinical signs about 4 wk after discontinuing the enrofloxacin. The decision to discontinue the enrofloxacin was made according to the in vitro susceptibility pattern available for both bacteria, as both were reportedly sensitive to therapy with amoxicillin. It is unclear whether one or both of the bacteria developed resistance to the amoxicillin, or if the streptococci sp. was nonpathogenic. There are reports of decreased in vivo susceptibility of E brevis to beta-lactam antibiotics due to the presence of a beta-lactamase gene, blaEBR-1. Therefore, these antibiotics are typically avoided when treating this pathogen in humans.¹³ In this patient, it was clear that the addition of the enrofloxacin led to rapid clinical improvement. The persistent clinical response to enrofloxacin is suggestive that of the two pathogens identified on culture, E brevis was likely the most pathogenic organism, as the streptococci sp. was resistant to enrofloxacin based on in vitro sensitivity testing. Clinical improvement after the addition of this antibiotic would not be expected if the streptococci sp. was the primary pathogenic organism causing clinical signs in this patient. In the authors' opinion, this information provides additional evidence to the pathogenicity of E brevis in dogs and is supportive that this organism was not a contaminant in this case.

Conclusion

IE can be difficult to diagnose antemortem because of the nonspecific and variable presentations as well as limited diagnostic capabilities. This case illustrates a rare and unique presentation of IE in the dog. Involvement of the pulmonic valve has been rarely reported as the sole location of infection, and this patient represents the first reported case of IE due to E brevis in a dog.