Etiologic diagnosis of infective endocarditis is very important for appropriate treatment, however culture-negative endocarditis has accounted for a substantial proportion of all endocarditis cases (1). Recent efforts to demonstrate the etiologic microorganisms of culture-negative endocarditis by use of the serologic tests and 16S rRNA gene sequencing of the valves identified Bartonella species as one of important causes in culture-negative endocarditis (2). Cases of endocarditis caused by Bartonella species have been reported from several countries; however there has been no report from Korea to date. Herein, we report the first case of Bartonella endocarditis confirmed by 16S rRNA gene sequencing in Korea.

A previously healthy 75-yr-old Korean woman was hospitalized with a 7-day history of exertional dyspnea and fever on August 16, 2011. She had not raised any pet dogs or cats. On admission, her vital signs showed body temperature 38.3℃, blood pressure 147/63 mmHg, heart rate 86/min, and respiratory rate 18/min. On physical examination, a grade 3 diastolic heart murmur was detected at left upper sternal border. Her complete blood count showed WBC 7,910/µL (neutrophils 90.8%, lymphocytes 7.5%), hemoglobin 12.7 g/dL and platelet 71,000/µL. Chemistry profiles showed total bilirubin 0.8 mg/dL, AST 16 IU/L, ALT 4 IU/L, alkaline phosphatase 79 IU/L, total protein 6.1 g/dL and albumin 2.8 g/dL. BUN 17.1 mg/dL, creatinine 0.87 mg/dL and CRP 1.4 mg/dL. Arterial blood gas analysis showed pH 7.537, pO₂ 65.9 mmHg, pCO₂ 31.3 mmHg, O₂ saturation 92.2%, and HCO₃ 26 mM. Urinalysis revealed microscopic hematuria and microalbuminuria. A chest radiograph and electrocardiogram showed no abnormality. Echocardiography demonstrated the presence of oscillating vegetation (0.46 × 1.77 cm) attached to the aortic valve, as well as moderate aortic regurgitation. Ceftriaxone was empirically given in a dose of 2 g/day after repeated blood cultures. On hospital day 2, she underwent urgent valve replacement surgery due to the risk of embolization. Intraoperative findings revealed 0.5 × 1.5 cm sized oscillating mass on left coronary cusp. Blood cultures obtained prior to antibiotic administration were all negative. Microbiological examination of the excised aortic valve and vegetation revealed no organism on the Gram stain and the culture grew no organism. Histology revealed fibromyxoid valvulopathy and chronic active inflammation with vegetation, which were consistent with infective endocarditis. Empirical ceftriaxone treatment was continued under the diagnosis of culture-negative endocarditis.

Bacterial DNA was extracted from surgically excised valve tissue and vegetation, directly. A 16S rRNA fragment was amplified with the universal primers 4F (5'-TTGGAGAGTTTGATCCTGGCT-3'), 534R (5'-TACCGCGGCTGCTGGCAC-3'), 27F (5'-AGAGTTTGATCMTGGCTCAG-3'), and 801R (5'-GGCGTGGACTTCCAGGGTATCT-3').

The 16S rRNA gene sequence showed 100% (707/707) similarity to that of the Bartonella quintana strain (accession no. BX897700 and AJ250247) in the GenBank database (http://blast.ncbi.nlm.nih.gov/). The next closest sequences were those of Bartonella henselae, and it showed 98.7% (699/708) similarity (accession no. DQ645426 and BX897699). Identification to the Bartonella species can be accepted if there is a ≥ 99% match with the 16S rRNA gene sequence and a difference of greater than 0.8% between those of other different species, according to the Clinical and Laboratory Standards Institute (CLSI) guideline MM18-A (3). B. quintana was also only the best-matched strain in the BIBI database (http://pbil.univ-lyon1.fr/bibi/). Based on the molecular identification using GenBank and BIBI databases, we concluded that B. quintana was the best matched species.

In addition, serologic testing with the indirect fluorescent antibody (IFA) method, which was performed at the Korea Center for Disease Control and Prevention showed an IgG titer of 1:512 to Bartonella spp. After she was finally diagnosed with B. quintana endocarditis, gentamicin was added to ceftriaxone in a dose of 1.7 mg/kg q8h. She was discharged after ceftriaxone treatment for 4 weeks in combination with gentamicin for the last 2 weeks.

This is the first reported case of B. quintana endocarditis in a healthy adult, which has been confirmed for the first time in Korea by using serologic test and 16S rRNA gene sequencing. Successful treatment of infective endocarditis requires timely and effective antibiotic therapy based on microbiological identification of causative microorganism. However, sometimes microbiological diagnosis remains difficult, and blood culture-negative endocarditis (BCNE) accounts for 2.5% to 31% of all cases of endocarditis (4, 5). Slow-growing or fastidious organisms such as the HACEK group, anaerobes, Abiotrophia spp., Brucella spp., Bartonella spp., Legionella spp., and Mycoplasma spp., or obligate intracellular organisms such as Coxiella burnetii had been reported as causes of BCNE (4). Microbiological identification of those pathogens in patients with endocarditis was very difficult until PCR-based sequence analyses have been adopted in clinical practices. Two prospective studies in French reference center from 1983 to 2001 and from 2001 to 2009, in which the etiology of BCNE was determined using a multimodal strategy including molecular methods, revealed that C. burnetii (37.0%-48.0%) and Bartonella spp. (12.4%-28.4%) were the most frequent microorganisms causing BCNE (2, 5).

Cases of endocarditis caused by Bartonella spp. have also been reported from other countries. An overview of published detailed cases of Bartonella endocarditis in Europe from 2000 to 2007 was presented, and the cases were reported from France, UK, Germany, Spain and the Netherlands (6). In addition, there have been reports of Bartonella endocarditis from Asian countries including India, Thailand and Japan as well as Australia (7). In contrast, in Korea, there has been no report of Bartonella endocarditis to date. However, recent reports have shown that Bartonella spp. were detected from the ticks and wild rodents in Korea (8), and B. henselae and Bartonella clarridgeieae were highly prevalent in Korean cats and dogs (9). In addition, there has been a paper reporting possible cases of cat scratch disease based on the seropositivity against B. henselae or B. quintana (10) and a report of cat scratch disease confirmed by molecular methods in Korea (11). These suggest that there is a possibility that cases of endocarditis by Bartonella spp. in Korea have not been diagnosed due to lack of suspicion. In fact, in many hospitals in Korea serologic tests for Bartonella spp. or gene sequencing from excised valves have not been generally performed for etiologic diagnosis for cases of culture-negative endocarditis.

Bartonella species are small fastidious Gram-negative rods, facultatively intracellular bacteria associated with a broad spectrum of diseases in humans, including cat scratch disease, trench fever, bacillary angiomatosis, chronic bacteremia, and endocarditis (12). Among various Bartonella species, B. quintana and B. henselae are the predominant human pathogens causing infective endocarditis (6, 12). Patients with B. henselae endocarditis frequently have a previous valvulopathy, and disease is associated with cat bites or scratches and cat flea exposure (13). In contrast, B. quintana endocarditis mostly develops in persons without any previous valvular injuries, and usually occurs in immunocompromised patients (12). Its other known risk factors include chronic alcoholism, homelessness, and body lice infestation (13). However, our patient did not have any alleged risk factors for B. quintana endocarditis.

Because of the fastidious nature of Bartonella spp., culture of these bacteria from clinical specimens remains very difficult. Therefore, serologic testing and/or molecular methods are necessary for Bartonella spp. identification. Previous study on a review of published cases of Bartonella endocarditis in Europe reported that pathogen identification was carried out by molecular methods and/or serologic testing in 16 out of 18 cases (6). In only two cases, Bartonella pathogens could be identified by culture methods. Currently, indirect immunofluorescent assay and enzyme-linked immunosorbent assay (ELISA) can be used for the serodiagnosis of Bartonella infection. Indirect immunofluorescent assay has a positive predictive value for Bartonella infection of 95% for an immunoglobulin G (IgG) titer to B. henselae and/or B. quintana of ≥ 1:800 in patients with endocarditis (13). However, serologic tests may not distinguish reliably between antibody responses to B. quintana and B. henselae (14). Furthermore, cross-reactions may occur at a low level with C. burnetii and significantly with Chlamydia spp. (14-16). In contrast, the molecular method using 16S rRNA gene sequencing or Bartonella species-specific 16S-23S rRNA intergenic region sequencing could resolve these problems and determine the species level of Bartonella.

Bartonella spp. are susceptible to most antimicrobial agents, including penicillins, cephalosporins, aminoglycosides, tetracyclines, macrolides, rifampicin, fluoroquinolones, and cotrimoxazole (17). However, only aminoglycosides have a bactericidal effect and it has been recommended to use aminoglycosides for minimum 2 weeks in combination with other class of antibiotics (12, 17, 18).

This report indicates that Bartonella endocarditis should be included in differential diagnosis of culture-negative endocarditis in Korea. Molecular genetic screening and serological testing methods for detection of Bartonella are necessary in cases of culture-negative endocarditis.