Changing epidemiology of infective endocarditis: a retrospective survey of 108 cases, 1990-1999.

The aim of this study was to report the experience with infective endocarditis over the past decade, describe the changing clinical and epidemiological features of the disease, and attempt to determine the optimal number of blood culture sets required for diagnosis. All cases diagnosed during a 10-year period were reviewed clinically and microbiologically. In addition, a retrospective assessment of blood culture data was performed. From the period 1990-1999, 108 cases that met the von Reyn or Duke's criteria were recorded. The major underlying cardiac condition was the presence of a prosthetic valve ( n=33 patients, 31%). Among patients with native valves, nonrheumatic valvular heart disease of the elderly was the most common underlying factor ( n=19 patients, 25%). Overall, 13 patients (11%) died. Predictors on admission for increased mortality were shortness of breath, age >60 years, time to defervescence, erythrocyturia, hemoglobin level <10 g/dl, and leukocytosis >15,000 (all P<0.05). Analysis of blood culture data showed that the diagnostic yield among groups from whom either only one or more than six blood culture sets were drawn was reduced compared with that among groups from whom between two and five culture sets had been taken. The outcome of endocarditis in this series from a community hospital was much more favorable compared with that reported in surveys from large tertiary centers. Moreover, obtaining more than two or three blood cultures is neither helpful nor cost-effective in the initial assessment of patients with suspected endocarditis.