To assess the most relevant features of hospital-acquired endocarditis, we conducted a retrospective study of cases of infectious endocarditis at a single university hospital from 1978 through 1992. During this period 248 episodes of infectious endocarditis were documented; 23 (9.3%) of these episodes were hospital-acquired and were not associated with cardiac surgery. (This figure represented a remarkable rise in the frequency of nosocomial endocarditis, only one case of which was identified among 101 cases of endocarditis treated at the same institution between 1960 and 1975.) In each of the 23 nosocomial cases, endocarditis was the result of bacteremia associated with a hospital-based procedure: intravenous catheterization (15 cases), instrumentation of a diseased urogenital tract (seven cases), or liver biopsy (one case). Staphylococcus aureus and Enterococcus faecalis were the predominant organisms isolated from intravenous catheters and the urogenital tract, respectively. Two of seven enterococcal isolates were highly resistant to gentamicin (MIC, > 2,000 micrograms/mL). Overall mortality was 56%. Two subsets of at-risk patients with different anatomic and clinical manifestations were identified. Our results emphasize that infectious endocarditis must be considered a serious nosocomial hazard against which preventive measures must be implemented.