Bartonella quintana and Coxiella burnetii as Causes of Endocarditis, India

To the Editor: In industrialized countries, blood culture is negative for 2.5%–31% of infectious endocarditis cases (1). In developing countries such as South Africa (2), Algeria (3), and Pakistan (4), culture is negative for 48% to 56%. Negative cultures delay diagnosis and treatment, which profoundly affects clinical outcome. Negative blood cultures commonly result from previous administration of antimicrobial drugs, right-sided endocarditis, or fastidious or noncultivable pathogens (1). Our aim was to identify fastidious agents of blood culture–negative endocarditis by serology. Because of recent attention to zoonotic microorganisms as agents of this condition in developing countries (1), we investigated the prevalence of Coxiella burnetii, Bartonella spp., and Brucella melitensis among endocarditis patients in India. We cultured blood from 111 patients admitted to the Government General Hospital, Chennai, India, from August 2005 through December 2006, with a diagnosis of infectious endocarditis according to the Duke criteria (5). Informed consent was obtained from all patients. Three blood samples from each patient, collected at hourly intervals, were inoculated into brain–heart infusion broth supplemented with 0.04% sodium polyanethol sulfonate (HiMedia, Mumbai, India). Cultures were incubated at 37°C in a 5% CO2 atmosphere for 14 days and checked each day for turbidity. Subcultures were made on 5% sheep blood and MacConkey agar at 37°C at 24 hours, 48 hours, and when culture broth appeared turbid. Blood cultures were negative for 80 (72%) of the 111 patients. Serum from 63 patients was available for serologic testing. Of these patients, 30 were male and 33 were female; age range was 5–65 years and mean age was 25.5 years. Endocarditis involved the native valve for 60 (95.23%) and a prosthetic valve for 3 (4.76%). The most frequent predisposing factor was rheumatic heart disease, found in 38 (60.31%). Of the 60 with native valve endocarditis, the involved valve was mitral for most (36, 60.0%), followed by aortic (8, 13.33%), tricuspid (7, 11.66%), and pulmonary (1, 1.66%); 8 (13.33%) had both valvular and nonvalvular endocarditis. Of the 3 patients with prosthetic valve endocarditis, the involved valve was mitral for 2 and aortic for 1. Serum samples were screened for Bartonella spp. and C. burnetii by microimmunofl uorescence (6,7). A diagnosis of endocarditis was based on an immunoglobulin (Ig) G titer >800 to phase I C. burnetii; this titer has a positive predictive value of 98% (6). A diagnosis of Bartonella infection was based on the combination of a positive microimmunofl uorescence titer (IgG to B. quintana or B. henselae >200) and a Western blot profi le consistent with endocarditis (8). Identifi cation of the causative species was obtained by Western blot after cross-adsorption with either B. henselae or B. quintana antigens (8).