Culture-Negative Infective Endocarditis Due to Bartonella henselae

Abstract

Bartonella henselae can cause culture-negative endocarditis, and almost all antibiotics included in empirical therapy regimens for endocarditis are ineffective against this pathogen. We report a case of culture-negative B henselae endocarditis in which identification of the pathogen avoided repeated valve destruction and surgery.

Graphical Abstract

History of Presentation

A 40-year-old man presented at a nearby secondary care hospital with dyspnea and weight loss lasting for 3 months. He was afebrile, with a respiratory rate of 18 breaths/min and an oxygen saturation of 95% without oxygen. His heart rate was 76 beats/min, and his blood pressure was 120/61 mm Hg. Physical examination revealed jugular vein distention and lower leg edema. He was hospitalized after being diagnosed with congestive heart failure. However, because his condition did not improve after initial treatment with loop diuretics for 6 days, he was transferred to our hospital on day 7.

Take-Home Message

• •This case highlights the clinical importance of understanding the appropriate workup and management of Bartonella henselae endocarditis because there is a discrepancy between the empirical therapy regimen for infectious endocarditis and B henselae–sensitive antibiotics.

On arrival to our hospital, he was afebrile at 35.9 °C and had a respiratory rate of 24 breaths/min with an oxygen saturation of 98% on 1 L/min supplemental oxygen. His heart rate was 91 beats/min, and his blood pressure was 105/58 mm Hg. Physical examination revealed a grade 3/6 systolic murmur best heard in the mitral valve area and an early diastolic murmur in the third to fourth intercostal space of the parasternal left margin of the sternum.

Past Medical History

The patient was under observation for primary mitral regurgitation. Ten months before this admission, the patient was admitted for an examination of splenomegaly accompanied by slight fever and lymphadenopathy, but the cause was unknown.

Differential Diagnosis

The initial differential diagnosis included refractory heart failure with valvular heart disease, pulmonary embolism, lymphoma, chronic infection, severe anemia, and pneumonia.

Investigations

His blood test results on admission were as follows: white blood cell count, 9,800/μL; hemoglobin, 11.6 mg/dL; total protein, 4.9 g/dL; total bilirubin, 2.1 mg/dL; creatinine, 0.98 mg/dL, C-reactive protein (CRP) 1.97 mg/dL; and brain natriuretic peptide, 1,812 pg/mL. Chest x-ray demonstrated pulmonary congestion, enlarged cardiac shadow, and pleural effusion (Figure 1A). Transesophageal echocardiography (TEE) on day 8 showed severe mitral regurgitation because of flail of mitral valve leaflets secondary to valve destruction with several highly mobile structures (Figure 1B). Significant aortic regurgitation was noted at the coaptation line of thickened right and left coronary cusps (Figure 1C). Secondary severe tricuspid regurgitation was observed, but there were no findings of infective endocarditis (IE) in the tricuspid and pulmonary valves. Based on these findings, the patient was diagnosed with congestive heart failure due to mitral and aortic valve destruction resulting from IE.

Figure 1.

Test Results From the Patient

(A) Severe pulmonary congestion and enlarged cardiac shadow were seen on the chest x-ray. (B and C) Transesophageal echocardiography showed mitral valve destruction with several highly mobile structures (red arrows) and severe mitral regurgitation (B) and significant aortic regurgitation at the coaptation line of thickened right and left coronary cusps (C). (D) Intraoperative images. In the aortic valve, vegetations are attached to the RCC and LCC, with significant valve destruction (left image). In the mitral valve, significant valve destruction is observed in all leaflets except P1, and vegetations are also seen on the chordae tendineae attached to both leaflets from the posterior papillary muscle, with several chordae tendineae ruptured (right image). LA = left atrium; LCC = left coronary cusp; LV = left ventricle; NCC = noncoronary cusp; PA = pulmonary artery; RCC = right artery cusp; RV = right ventricle; S-G = Swan-Ganz catheter.

Management

The patient’s heart failure symptoms progressively worsened because of low output syndrome despite the use of dobutamine and diuretics. Emergency open heart surgery (aortic and mitral valve replacement, tricuspid valve repair, and left atrial appendage closure) was performed on day 8, and intraoperative macroscopic findings were similar to those of the echocardiography (Figure 1D). After surgery, his hemodynamics stabilized. The contrast-enhanced computed tomography performed postoperatively showed no signs of embolism in any organs, including the brain, and no infectious cerebral aneurysms were observed.

Two sets of blood cultures before antibiotic administration collected at the previous hospital were negative. The culture of the infective aortic valve tissue showing bacterial masses on pathology (Figure 2) was also negative. The patient’s overall health and blood data were improving with the continued use of antibiotics (beta-lactam antibiotics and vancomycin), which are recommended in the current guidelines as empirical therapy for IE before pathogen identification. However, CRP levels failed to return to normal and began to rise again on day 24.

Figure 2.

Blood Cultures

Pathologic examination of the aortic left coronary cusp showed neutrophil infiltration and bacterial masses (red dotted line), revealing abscess formation. Scale bar, 50 μm.

A medical interview conducted again revealed intense contact with cats at work regularly. Thus, culture-negative IE secondary to cat scratch disease was suspected. Indirect immunofluorescence assays detected an IgG antibody titer of >1:800 for Bartonella henselae, and polymerase chain reaction analysis of the infectious valve tissue detected DNA of B henselae. Accordingly, the patient was diagnosed with endocarditis caused by B henselae. The patient was switched to antibiotic therapy with doxycycline and rifampicin, and on day 43, CRP levels were found to be normal. Follow-up TEE performed on day 53 showed new vegetation not seen preoperatively on the tricuspid valve (Figure 3), but oral doxycycline and rifampicin treatment was continued because of a lack of findings suggestive of valve destruction. The second TEE follow-up on day 70 showed the vegetation on the tricuspid valve was still present but had not increased in size.

Figure 3.

Transesophageal Echocardiography From 3 Different Time Frames

(A) Transesophageal echocardiography (TEE) conducted preoperatively showed no findings of endocarditis on the tricuspid valve. (B and C) Two follow-up TEEs revealed new mobile string structures that were suggested endocarditis on the valve. Red arrows show new vegetation not seen preoperatively on the tricuspid valve. RA = right atrium; RV = right ventricle.

Discussion

IE is an uncommon but highly fatal infectious disease.^1,2 Mortality rates can be substantially high among patients with IE, especially when there is a delay in identifying the causative pathogen and appropriate antibiotic therapy is not initiated promptly. Bacterial culture tests, in particular blood cultures, are the gold standard for diagnosing IE and identifying the causative pathogen. However, blood culture–negative endocarditis can occur in 5% to 20% of cases.^1,3 The most common reason for culture-negative endocarditis is the previous administration of antibiotics before cultures are obtained. Another reason is infection with pathogens that are not isolated by routine culture techniques.^1,2,4

Bartonella species was first described as a cause of IE in 1993 and is now considered a significant cause of culture-negative endocarditis.⁵ B henselae, which is known as a causative pathogen of cat scratch disease, can cause culture-negative endocarditis, especially in individuals with valvular disorders.⁴ A previous report showed more than 40% of patients with B henselae endocarditis had previous valve disease.⁶ In the present case, the splenomegaly accompanied by slight fever and lymphadenopathy for which the patient underwent examination 10 months before this admission was most likely due to cat scratch disease. The persistent bacteremia could have caused IE because of the existing valvular disease.

Some case reports have mentioned the relation between extensive cerebral ischemia caused by septic embolism and B henselae IE.^7,8 De Cristofaro et al⁷ reported a case of B henselae IE with a small vegetation on the mitral valve, which was related to multiple cerebral septic embolic events. Fortunately, our case did not show any obvious findings of a cerebrovascular disorder despite remarkable valve destruction. The reason for this difference is unknown, but it suggests that B henselae IE can cause both significant heart and cerebrovascular disorders.

The recently proposed 2023 Duke-International Society for Cardiovascular Infectious Diseases Criteria for Infective Endocarditis include enzyme immunoassay IgG titer and polymerase chain reaction for B henselae as new major clinical criteria.⁴ In the present case, these tests conducted according to the new IE criteria revealed B henselae to be the causative pathogen, highlighting the indispensability of the new IE criteria to identify the causative pathogen in culture-negative endocarditis cases. Additionally, a detailed medical interview, which revealed intense contact with cats, triggered suspicion of IE secondary to cat scratch disease. This observation strongly suggests the importance of medical interviews in culture-negative endocarditis cases.

Before identifying the causative pathogen or if culture test results are negative, empirical therapy is generally required. Empirical antibiotic therapy regimens in the current guidelines mainly include broad-spectrum beta-lactam antibiotics and vancomycin or daptomycin for methicillin-resistant Staphylococcus aureus.^1,2,9 However, almost all these antibiotics are ineffective against B henselae, because this pathogen is an intracellular gram-negative bacterium. In the present case, IE recurred on the tricuspid valve during the search for a causative pathogen after surgery. Fortunately, identification of B henselae as the causative pathogen and the subsequent switch to appropriate antibiotics avoided repeated valve destruction and surgery. There is a discrepancy between the empirical therapy regimen for IE and B henselae–sensitive antibiotics. Therefore, we strongly suggest the importance of identifying the causative pathogen by conducting detailed medical interviews and examinations according to the new IE criteria in culture-negative endocarditis.

Follow-Up

The patient was discharged from our hospital on day 81. He will continue the oral antibiotic therapy for at least 6 months after initiation.

Conclusions

The present case suggests that identifying the causative pathogen is clinically essential even in culture-negative endocarditis to avoid reactivation and exacerbation of endocarditis.

Funding Support and Author Disclosures

The authors have reported that they have no relationships relevant to the contents of this paper to disclose.