Fax +41 61 306 12 34E-Mail karger@karger.chwww.karger.com

Novel Insights from Clinical Experience

Cardiology 2009;114:208–211 DOI: 10.1159/000228645

Infective Endocarditis by Bartonella quintana Masquerading as Antineutrophil Cytoplasmic Antibody-Associated Small Vessel Vasculitis

Hiroaki Sugiyama a Makoto Sahara a Yasushi Imai a Minoru Ono b

Koh Okamoto c Ken Kikuchi d Ryozo Nagai a

Departments of a Cardiovascular Medicine, b Cardiothoracic Surgery, and c Infectious Diseases, University of Tokyo Hospital, and d Center of Excellence for Infection Control Science, Graduate School of Medicine,Juntendo University, Tokyo , Japan

vessel vasculitis. However, a variety of infections can result in a false-positive ANCA test, and especially subacute bacte-rial endocarditis (SBE) with the presence of ANCAs occasion-ally mimics the clinical manifestations of an ANCA-associat-ed vasculitis such as skin purpura and glomerulonephritis. In contrast, noninfectious endocardial involvement is known to be part of the spectrum of the manifestations of the ANCA-associated vasculitis. Therefore, it is crucial to distin-guish an ANCA-positive SBE from an ANCA-associated vas-culitis with endocardial compromise, because the misdiag-

Key Words

Bartonella � Blood culture-negative endocarditis � Antineutrophil cytoplasmic antibodies

Abstract

The Bartonella species have been recently recognized as im-portant causative agents of culture-negative bacterial en-docarditis. Antineutrophil cytoplasmic antibodies (ANCAs) have been associated with the spectrum of idiopathic small

Received: April 24, 2009 Accepted after revision: May 2, 2009 Published online: July 15, 2009

Hiroaki Sugiyama, MD Department of Cardiovascular Medicine, University of Tokyo Hospital 7-3-1, Hongo, Bunkyo-ku Tokyo 113-8655 (Japan) Tel. +81 3 3815 5411, Fax +81 3 5800 9780, E-Mail hsugiyama-tky@umin.ac.jp

© 2009 S. Karger AG, Basel0008–6312/09/1143–0208$26.00/0

Accessible online at:www.karger.com/crd

Established Facts

• The Bartonella species are important causative agents of culture-negative infective endocarditis. • Subacute bacterial endocarditis (SBE) occasionally exhibits a positive antineutrophil cytoplasmic

antibody (ANCA) test and mimics ANCA-associated small vessel vasculitis. Conversely, noninfec-tious endocardial involvement is part of the spectrum of the manifestations of ANCA-associated vasculitis.

Novel Insights

• A blood culture-negative SBE caused by B. quintana can exhibit a positive ANCA test and simulate an ANCA-associated vasculitis such as skin purpura and glomerulonephritis.

• The combination of a serological test with a polymerase chain reaction restriction fragment length polymorphism analysis is helpful for the correct diagnosis and appropriate treatment of a blood cul-ture-negative SBE caused by B. quintana.

Dow

nloa

ded

by:

Uni

vers

ity o

f Hon

g K

ong

14

7.8.

204.

164

- 9/

9/20

13 1

2:34

:01

AM

Bartonella Endocarditis Simulating ANCA-Associated Vasculitis

Cardiology 2009;114:208–211 209

nosis of an SBE as an ANCA-associated vasculitis can lead to an inappropriate immunosuppressive therapy with cata-strophic consequences. The differential diagnosis is some-times difficult, especially in the case of culture-negative in-fective endocarditis with a positive ANCA test. We describe here a case of a culture-negative SBE caused by Bartonella quintana, accompanied with a positive cytoplasmic ANCA test and clinical findings masquerading as ANCA-associated vasculitis. Both a serological test for Bartonella and poly-merase chain reaction restriction fragment length polymor-phism analysis were helpful for a correct diagnosis and ap-propriate treatment.

Copyright © 2009 S. Karger AG, Basel

Introduction

Blood culture-negative infective endocarditis remains a diagnostic and therapeutic challenge for clinicians. Cas-es of blood culture-negative infective endocarditis account for 5–14% of all endocarditis cases [1, 2] . The main reason for negative blood cultures is fastidious or slowly growing bacteria, for example, the Bartonella species (spp.) or Cox-iella burnetii, as well as an antibiotic treatment preceding the collection of the blood cultures. The Bartonella spp., especially B. quintana, have been recognized as important causes of culture-negative endocarditis. However, the de-tection and identification of these bacterial pathogens can be met with difficulties, and an effective antibiotic thera-py often tends to be delayed [3] .

Antineutrophil cytoplasmic antibodies (ANCAs) that are directed against either proteinase-3 or myeloperoxi-dase correlate well with a certain group of small vessel vasculitis that includes microscopic polyangiitis and Weg-ener’s granulomatosis [4] . Infective endocarditis occa-sionally accompanies a variety of immunologic phenom-ena including small vessel vasculitis such as cutaneous purpura and glomerulonephritis. Several infectious dis-eases, particularly subacute bacterial endocarditis (SBE), have been reported to exhibit positive ANCA tests and to mimic ANCA-associated idiopathic vasculitis, which may lead to an incorrect diagnosis and treatment [5] . Conversely, noninfectious endocardial involvement is known to be part of the spectrum of the manifestations of the ANCA-associated idiopathic vasculitis [5] . There-fore, it is crucial to distinguish an ANCA-positive SBE with small vessel vasculitis from an ANCA-associated idiopathic vasculitis with endocardial compromise, be-cause the treatment strategies for both disease conditions are entirely different. Here, we report a case of a culture-

negative SBE caused by B. quintana with a high cytoplas-mic ANCA (C-ANCA) titer, who presented with features simulating ANCA-associated small vessel vasculitis.

Case Report

A 64-year-old construction worker was admitted to our hos-pital with a 3-month history of general malaise and dyspnea on exertion. He had a history of hypertension along with insidious mild renal dysfunction. No organic heart disease had ever been documented during the routine health checkups. There was no history of any tooth extractions, drug addiction or skin injury. He did not have any recent contact with cats.

The significant physical findings on admission included a temperature of 37.0 ° C, loud diastolic decrescendo murmur over the right parasternal border, and purpuric rash and mild pitting edema on the anterior aspect of his legs. The basic laboratory data revealed the following values: leukocyte count 7.5 ! 10 3 /mm 3 , with 43.5% neutrophils, 49.5% lymphocytes, 5.5% monocytes and 0.5% eosinophils; hemoglobin 7.3 g/dl; total protein 7.7 g/dl; albu-min 2.4 g/dl; serum creatinine 1.33 mg/dl, and C-reactive protein 2.65 mg/dl. The plasma brain natriuretic peptide level was re-markably increased up to 491 pg/ml. The rheumatoid factor was 453 IU/ml (reference range ! 20) and the antinuclear antibody was strongly positive with a titer of 1: 320. The serum complement

3 level was reduced (59.0 mg/dl, reference range 65–220), while the total complement and complement 4 levels were not. The

C-ANCA titer was as high as 60 ELISA units (reference range ! 10); however, the perinuclear ANCA titer was not elevated. The serological tests for viruses including HIV were negative. A uri-nalysis revealed (+++) occult blood and (++) protein (24-hour proteinuria of 846 mg/day); a microscopic examination of the urine sediment revealed the presence of a lot of oddly shaped red blood cells and several types of casts suggesting glomerular in-jury. The urine culture was negative. The estimated glomerular filtration rate on admission was 46 ml/min/1.73 m 2 .

The echocardiographic examination demonstrated a tricuspid aortic valve with third-degree aortic valve regurgitation, vegeta-tions of 9 mm in diameter attached to the right coronary cusp and noncoronary cusp, and left ventricular dilation with preserved systolic function. A total of 6 sets of aerobic and anaerobic blood cultures obtained before the initiation of the antimicrobial ther-apy remained all negative. Based on the echocardiographic find-ings and his clinical course, a culture-negative SBE of the aortic valve was strongly suspected, and an intravenous administration of ceftriaxone and gentamicin was initiated 5 days after admis-sion as an empiric antimicrobial therapy. After that, serologic testing of his serum samples revealed significantly high IgG titers of 1 1: 1,024 (expected value ! 1: 128) against both B. henselae and B. quintana, whereas the immunoassays for the IgM antibodies for both the pathogens and IgM and IgG antibodies for C. burnetii remained negative. Thus, the antibiotic regimen was supplement-ed with a daily oral dose of 200 mg of doxycycline to cover the spectrum against Bartonella spp. Despite a few weeks of antibi-otic therapy, his general condition and inflammatory response due to the infection did not improve. Furthermore, he developed a relatively rapid decline in his renal function, and his serum cre-atinine level became elevated up to 3.81 mg/dl (estimated glomer-

Dow

nloa

ded

by:

Uni

vers

ity o

f Hon

g K

ong

14

7.8.

204.

164

- 9/

9/20

13 1

2:34

:01

AM

Sugiyama /Sahara /Imai /Ono /Okamoto /Kikuchi /Nagai

Cardiology 2009;114:208–211210

ular filtration rate 14 ml/min/1.73 m 2 ) by 23 days after admission. Because a complication of ANCA-associated rapidly progressive glomerulonephritis was suspected, a renal biopsy was performed; however, there was no evidence of rapidly progressive glomerulo-nephritis such as glomerular crescent formations in the renal bi-opsy specimens, which revealed focal glomerular sclerosis, mild interstitial inflammation along with complement depositions and mild tubular degeneration.

A prosthetic aortic valve replacement surgery was performed 40 days after admission due to drug-refractory endocarditis and congestive heart failure associated with severe aortic regurgita-tion. Intraoperatively, 2 pieces of vegetations were noted on the aortic leaflet: 1 on the noncoronary cusp and the other on the right coronary cusp (8 and 5 mm in diameter, respectively). The surgically isolated aortic valve was homogenized and cultured with the standard methods; however, not only routine Gram staining and periodic acid-Schiff staining but also Warthin-Star-ry silver staining of the removed tissue could not reveal any bac-teria. A histologic examination of the valve revealed calcifications and fibrotic changes with mild infiltration of inflammatory cells rather than granulomatous inflammation. Next, using the ge-nomic DNA extracted from the surgically isolated aortic valve, the polymerase chain reaction (PCR) restriction fragment length polymorphism (RFLP) analysis was performed. The PCR prod-ucts amplified with species-specific primers, including 3 Barton-ella genes, i.e., riboflavin synthase (ribC), citrate synthase (gltA) and heat shock protein (htrA), were digested with the restriction enzyme Taq I, and the restriction patterns were analyzed by poly-acrylamide gel electrophoresis. Consequently, the causative mi-croorganism of the endocarditis was identified as B. quintana.

After a surgical intervention, the same antibiotic therapy (cef-triaxone and doxycycline) was successively administered for a to-tal of 6 weeks. Then, his general condition as well as the results of the laboratory tests associated with the inflammation and uri-nary abnormalities gradually became ameliorated. A titer of the C-ANCA also exhibited a clearly declining trend and finally be-came negative with the disappearance of the cutaneous purpura of his lower extremities. His renal function was restored to the prehospital state (serum creatinine 1.0–1.5 mg/dl) at the time of discharge. Eight months later, the patient was generally doing fine and demonstrated no evidence of systemic infection, vasculitis or cardiac decompensation.

Discussion

Here, we present an elderly male patient with a cul-ture-negative SBE caused by B. quintana and the clinical manifestations masquerading as an ANCA-associated small vessel vasculitis such as cutaneous purpura and glomerulonephritis.

The Bartonella spp., which are represented by B. quin-tana and B. henselae, have become appreciated as an in-creasingly important etiologic agent of infective endocar-ditis [3] . In particular, B. quintana endocarditis often oc-curs in patients without any preexisting valvular disease and is associated with homelessness, alcoholism and

body louse infestations [3, 6] . The patient was neither a heavy drinker nor a homeless person; however, having ordinarily worked in an unsanitary environment, e.g. a sewage system, may have been related to the source of his infection. Because Bartonella endocarditis often requires surgical intervention involving a heart valve replacement and additive antibiotic agents such as doxycycline, an early diagnosis is of great importance. However, the de-tection and identification of the Bartonella spp. is often a difficult task, since cultures of the clinical material ob-tained from patients are usually negative for bacteria due to their fastidious and slow-growing nature. Therefore, to precisely diagnose Bartonella endocarditis, other modal-ities than blood cultures are required for detecting the bacteria. Serological techniques are the most widely used methods for the diagnosis of a Bartonella infection, but the methodology has some shortcomings linked to the well-known cross-reactions among the Bartonella spp. or between the Bartonella spp. and other organisms such as C. burnetii and Chlamydia spp. [7] . PCR amplification of the Bartonella DNA (or RNA) obtained from homoge-nates of valvular vegetations, if available, is commonly performed for the diagnosis, and direct sequencing of the 16S ribosomal RNA is currently used as the most accu-rate tool to detect Bartonella [8] . However, the direct se-quencing method cannot always be performed in most laboratories. Instead of a direct sequencing method, a PCR-RFLP assay using DNA extracted from tissue and species-specific primers might be an alternative method to diagnose Bartonella endocarditis, as used in the pres-ent case. In fact, it has recently been reported that the PCR-RFLP analysis with the Tac I-digested ribC amplifi-cation products can successfully discriminate between the known Bartonella spp. causing endocarditis [9] . Al-though the possibility of the emergence of a new Barton-ella sp. must be taken into consideration when evaluating the results from the PCR-RFLP analysis, the PCR-RFLP approach concurrently analyzing multiple genes, which was applied in the present case, may be helpful in per-forming a validated diagnosis of endocarditis caused by a particular Bartonella sp.

It is also noteworthy that our patient presented with immunologic phenomena related to small vessel vasculi-tis such as cutaneous purpura and acute glomerulone-phritis with high titers of the C-ANCA. To the best of our knowledge, this is the first report of B. quintana endocar-ditis with positive C-ANCA tests and clinical findings masquerading as an ANCA-associated small vessel vas-culitis. The presence of C-ANCA is generally thought to be highly specific for Wegener’s granulomatosis [4] . In

Dow

nloa

ded

by:

Uni

vers

ity o

f Hon

g K

ong

14

7.8.

204.

164

- 9/

9/20

13 1

2:34

:01

AM

Bartonella Endocarditis Simulating ANCA-Associated Vasculitis

Cardiology 2009;114:208–211 211

contrast, a variety of infectious diseases including infec-tive endocarditis sometimes result in a positive ANCA test and, in particular, ANCA-positive SBE is known to mimic the clinical manifestations of an ANCA-associ-ated small vessel vasculitis, as seen in the present case [5, 10] . However, the reported cases of ANCA-positive SBE have indicated that the small vessel involvement, limited to the skin and kidneys, was due to an immune complex-mediated complement deposition rather than to a pauci-immune inflammation that occurs in ANCA-associated idiopathic vasculitis [5, 10] . Thus, the implication of the presence of ANCAs in infectious diseases remains un-clear. Given that there is no evidence that ANCAs play a role in the pathogenesis of a small vessel disease in SBE, the presence of ANCAs in infective endocarditis may be falsely positive and the infectious process may induce the production of ANCAs, possibly through a nonspecific B cell activation or autoimmunization after the release of proteinase-3 from neutrophils. Conversely, ANCA-asso-ciated idiopathic vasculitis can masquerade as SBE with noninfectious endocardial involvement, possibly through valvulitis inducing a following valve distortion [5, 10, 11] . Although a great degree of overlap in the clinical and lab-oratory manifestations seems to occur between ANCA-associated idiopathic vasculitis with endocardial com-promise and ANCA-positive SBE, potentially leading to a misdiagnosis, it is crucial to distinguish between both of these disease conditions in order to avoid an inappro-priate therapy such as immunosuppressive treatment in patients with SBE, resulting in life-threatening conse-quences. By taking the appropriate clinical steps, the dif-

ferential diagnosis is usually obtained without difficulty, because serial blood cultures should allow the clinician to detect streptococcal or enterococcal SBE with a posi-tive ANCA test [5, 10, 12] . However, when serial blood cultures are all negative, as in the present case of B. quin-tana endocarditis, it may become difficult to distinguish between the 2 diseases. Given that the elevated C-ANCA titer normalized with the clinical resolution of the SBE after an appropriate surgical and antibiotic therapy, and that the patient then has remained free of any evidence of systemic vasculitis during the follow-up, it is plausible that the presence of the C-ANCA was also falsely positive in the present case. Notably, the awareness that an ANCA positivity associated with a small vessel vasculitis can also occur in a culture-negative SBE such as B. quintana endocarditis is important.

In conclusion, we encountered a patient with a cul-ture-negative SBE caused by B. quintana accompanied with a positive C-ANCA test and clinical findings mas-querading as ANCA-associated vasculitis of the skin and kidneys. Both the serological Bartonella IgG test and PCR-RFLP analysis for multiple genes, as well as the renal biopsy findings, were helpful for making a correct diag-nosis and appropriate treatment choice. When encoun-tered with a positive ANCA test in patients suspected of having culture-negative endocarditis and systemic vas-culitis, physicians should take an SBE caused by fastidi-ous or slowly growing bacteria such as Bartonella spp. into consideration and take the appropriate diagnostic steps to detect any infectious diseases before the initia-tion of immunosuppressive therapy.

References

1 Hoen B, Selton-Suty C, Lacassin F, Etienne J, Briançon S, Leport C, Canton P: Infective endocarditis in patients with negative blood cultures: analysis of 88 cases from a one-year nationwide survey in France. Clin Infect Dis 1995; 20: 501–506.

2 Breathnach AS, Hoare JM, Eykyn SJ: Culture-negative endocarditis: contribution of Bar-tonella infections. Heart 1997; 77: 474–476.

3 La Scola B, Raoult D: Culture of Bartonella quintana and Bartonella henselae from hu-man samples: a 5-year experience (1993–1998). J Clin Microbiol 1999; 37: 1899–1905.

4 Harper L, Savage CO: Pathogenesis of ANCA-associated systemic vasculitis. J Pathol 2000; 190: 349–359.

5 Chirinos JA, Corrales-Medina VF, Garcia S, Lichtstein DM, Bisno AL, Chakko S: Endo-carditis associated with antineutrophil cyto-

plasmic antibodies: a case report and review of the literature. Clin Rheumatol 2007; 26: 590–595.

6 Fournier PE, Lelievre H, Eykyn SJ, Mainardi JL, Marrie TJ, Bruneel F, Roure C, Nash J, Clave D, James E, Benoit-Lemercier C, De-forges L, Tissot-Dupont H, Raoult D: Epide-miologic and clinical characteristics of Bar-tonella quintana and Bartonella henselae endocarditis: a study of 48 patients. Medi-cine (Baltimore) 2001; 80: 245–251.

7 Maurin M, Birtles R, Raoult D: Current knowledge of Bartonella species. Eur J Clin Microbiol Infect Dis 1997; 16: 487–506.

8 Goldenberger D, Künzli A, Vogt P, Zbinden R, Altwegg M: Molecular diagnosis of bacte-rial endocarditis by broad-range PCR ampli-fication and direct sequencing. J Clin Micro-biol 1997; 35: 2733–2739.

9 Avidor B, Graidy M, Efrat G, Leibowitz C, Shapira G, Schattner A, Zimhony O, Giladi M: Bartonella koehlerae, a new cat-associ-ated agent of culture-negative human en-docarditis. J Clin Microbiol 2004; 42: 3462–3468.

10 Choi HK, Lamprecht P, Niles JL, Gross WL, Merkel PA: Subacute bacterial endocarditis with positive cytoplasmic antineutrophil cy-toplasmic antibodies and anti-proteinase 3 antibodies. Arthritis Rheum 2000; 43: 226–231.

11 Anthony DD, Askari AD, Wolpaw T, Mc-Comsey G: Wegener granulomatosis simu-lating bacterial endocarditis. Arch Intern Med 1999; 159: 1807–1810.

12 Wagner J, Andrassy K, Ritz E: Is vasculitis in subacute bacterial endocarditis associated with ANCA? Lancet 1991; 337: 799–800.

Dow

nloa

ded

by:

Uni

vers

ity o

f Hon

g K

ong

14

7.8.

204.

164

- 9/

9/20

13 1

2:34

:01

AM