Epidemiology of bacterial endocarditis in The Netherlands. I. Patient characteristics

BACKGROUND: Studies of the epidemiology of bacterial endocarditis are usually based on a retrospective review of medical records from referral centers serving diverse patient populations. These studies are therefore likely to suffer from selection bias. We conducted a nationwide prospective epidemiologic study of endocarditis in the Netherlands. METHODS: During a 2-year period, all cases of consecutively hospitalized patients with suspected endocarditis in the Netherlands were reported to us. While hospitalized, patients were visited for an in-person interview and a review of the medical record. RESULTS: Of 559 episodes, 438 met the criteria for endocarditis; these included 89 episodes of prosthetic valve endocarditis and 349 episodes of native valve endocarditis. Adjusted for age- and sex-specific population figures, the incidence was 19 per million person-years. The incidence increased significantly with age, and men were more often affected than women (266 and 172 cases, respectively). Rheumatic and congenital cardiac lesions formed most of the underlying heart diseases. Mitral valve prolapse was present in only 29 patients with native valve endocarditis (8.3%). A history of intravenous drug abuse was present in 32 patients (7.3%). Viridans streptococci, staphylococci, and enterococci together constituted 86% of the isolated bacterial strains. Only 1.1% of the patients had culture-negative endocarditis. Overall case fatality was 19.7% and varied widely according to causative microorganism. CONCLUSION: The distribution of causal microorganisms, the case fatality rate, and the incidence rate of endocarditis are age related. Therefore, a meaningful comparison of data is only possible between population-based cohorts of patients with endocarditis.