Analysis of 1,186 episodes of gram-negative bacteremia in non-university hospitals: the effects of antimicrobial therapy.

Over the five-year period 1977-1981, we studied 1,186 episodes of bacteremia due to Enterobacteriaceae and Pseudomonadaceae in the four non-university hospitals of a single metropolitan area. Overall patient mortality was 36.3%, and mortality attributed specifically to infection was 19.0%. The importance of severity of underlying disease, site of infection, microorganism, and age--previously determined to be prognostic factors in studies conducted at tertiary-care centers--was confirmed. Appropriate initial antimicrobial therapy (defined as the administration of an effective agent in adequate dose and by a suitable route of administration on the first calendar day on which blood cultures were positive) did not improve survival compared with the use of an ineffective antimicrobial agent or no therapy at all. However, appropriate antimicrobial therapy subsequent to the first calendar day on which blood cultures were positive clearly affected survival. These findings confirm previous conclusions regarding the frequency and severity of gram-negative bacteremia and the overall impact of antimicrobial therapy on this condition. These studies also suggest the possibility that the definition of optimal initial therapy in some groups of patients should be reconsidered.