The clinical significance of positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults. II. Clinical observations, with special reference to factors influencing prognosis.

Among 500 patients with bacteremia and fungemia, total mortality was 42%; about half of all deaths were attributable directly to infection. Mortality increased with age, but deaths unrelated to infection itself were responsible in part for this increase. Mortality was 2.6% among obstetric-gynecologic patients, 42% among medical patients, 49% among surgical patients, and 60% among transplant patients. The risk of death was especially high with enterococcal, facultative gram-negative, fungal, polymicrobial, or hospital-acquired sepsis; in the presence of shock, leukopenia, absolute granulocytopenia, or defined predisposing conditions (neoplasia, cirrhosis, and combinations of factors such as surgery and renal failure); and with a primary infected focus in the respiratory tract, the skin, a surgical wound, an abscess, or an unknown site. Body temperature was inversely related to mortality. Survival was increased by the use of appropriate antibiotics and, where applicable, additional therapeutic maneuvers (e.g., drainage). Multivariate analysis defined seven variables that independently influenced outcome: microorganism, blood pressure, body temperature, primary focus of infection, place of acquisition of infection, age, and predisposing factors. Although some adverse prognostic factors are not amenable to intervention, prevention of nosocomial bacteremia and fungemia and early reversal of hypotension may reduce the death rate from sepsis.

Duke Scholars

Author Lyman Barth Reller Pathology