Nosocomial bacteremia. An epidemiologic overview.

Each year nosocomial bacteremia develops in approximately 194,000 patients in U.S. hospitals (5/1,000); 75,000 die. These infections portend $.28 to $.86 billion added costs to health care. Most nosocomial bacteremias occur endemically and are secondary bacteremias, caused by postoperative wound or intra-abdominal infections, urinary tract infections or pneumonia; primary bacteremias most frequently originate from intravascular devices, but the source is unrecognized. Between 1965 and 1978, 97 epidemics of nosocomial bacteremia, including 11 of "pseudobacteremia," were reported. In contrast to endemic bacteremias, 78 percent of the epidemics involved primary bacteremias: 33 outbreaks stemmed from infusion therapy in some form, including seven epidemics traced to a contaminated commercial product. Two thirds of endemic nosocomial bacteremias and 79 percent of epidemics are caused by aerobic gram-negative bacilli. Pseudomonas cepacia, Pseudomonas maltophilia, Flavobacterium and Enterobacter agglomerans rarely cause endemic bacteremia and when encountered often signal an epidemic. Whereas predisposing host conditions greatly increase the risk of bacteremia endemically nosocomial epidemics occur mainly in immunocompetent patients and are related to what therapeutic measures have been taken: segregation in a special care unit (58 percent of outbreaks) or exposure to infusion therapy or other invasive procedures involving the bloodstream (65 percent). At present only about one fourth of endemic nosocomial bacteremias are in theory preventable by more consistent application of existent knowledge of asepsis. The potential for prevention seems greatest for epidemic bacteremias, most of which are related to exposure to invasive devices, to a common source of contamination, or both.