S Saint, Ann Arbor VA Medical Center and University of Michigan; RL Lark, University of Michigan; JK Zemencuk, Ann Arbor VA Medical Center and University of Michigan; C Chenoweth, University of Michigan; BA Lipsky, Seattle VA Medical Center and University of Washington; JJ Plorde, Seattle VA Medical Center and University of Washington

Objectives: Over 250,000 patients in the United States develop bacteremia or fungemia each year, with an associated mortality of 10 to 30% per episode. Septicemia is the twelfth leading cause of death in the United States. Though many recent studies of nosocomial bacteremia have been reported, there are limited data focusing on community-acquired bacteremia. Given the morbidity, mortality, and economic consequences of community-acquired bacteremia, we decided to: (1) describe the epidemiology and microbiology of community-acquired bacteremia; (2) determine the crude mortality associated with such infections; and (3) identify independent predictors of mortality.

Methods: This prospective study was conducted at the Seattle Division of the Veterans Affairs Puget Sound Healthcare System from January 1, 1994 to December 31, 1997. All patients with clinically significant community-acquired bacteremia or fungemia were evaluated. Data were collected on demographics, co-morbid conditions, clinical parameters, microorganisms, source of infection, and patient outcome.

Results: During the study period, 387 bacteremic episodes occurred in 334 patients. Staphylococcus aureus (18%), Escherichia coli (15%), and coagulase-negative staphylococci (12%) were the most commonly isolated organisms. The most frequent sources were the urinary tract, intravascular catheters, and pneumonia. Overall, almost one-third of bacteremia cases were directly related to indwelling catheters: either intravascular (20%) or urinary (10%).

Approximately 14% of patients died. Patient characteristics independently associated with increased mortality included shock (OR 3.7, p=0.02), renal failure (OR 4.0, p=0.003), and a "Do Not Attempt Resuscitation" order (OR 21.7, p< 0.001). The risk of death was also higher in those whose source was pneumonia (OR 6.3, p=0.03) or an intra-abdominal site (OR 10.7, p=0.02), or if multiple sources were identified (OR 13.4, p=0.003). The presence of fever (temperature > 38.0 C) was associated with a decreased risk of death (OR 0.4; p=0.005).

Conclusions: This study provides an in-depth analysis of the epidemiology and microbiology of community-acquired bacteremia in the 1990s. S. aureus, E. coli, and coagulase-negative staphylococci are now the leading pathogens of community-acquired bacteremia. Vascular and urinary catheters were implicated in a substantial proportion of infections, emphasizing the need for appropriate and judicious use of such devices. In addition, we will likely observe an increase in the incidence of such device-related infection as healthcare continues to expand into the outpatient setting.

Impact: Our findings suggest that community-acquired bacteremia is often device-related. Fortunately, many of these device-related infections are potentially avoidable through adherence to the established infection control practices used to prevent hospital-acquired infections. Thus, strategies that have been useful in preventing nosocomial device-related bacteremia could prove successful if adapted into the outpatient setting. If this is borne out in future studies, community-acquired bacteremia should increasingly become viewed as a preventable disease.