[Prognostic factors of bacteremia: prospective study]

Abstract

Background: To identify the epidemiology and risk factors with influence in the outcome and mortality of a series of bacteriemic patients.

Material and methods: A prospective study of bloodstream infections with clinical significance detected in a secondary hospital of 650 beds over period from May 1998 to May 1999. The true bacteriemia was defined in basis to the criteria both the physician and microbiologist. A total of 16 variables were defined and categorized such as clinical-epidemiologic, intrinsic risk factor, extrinsic risk factor, outcome and survival. We used SPSS statistical package: For cuantitative variables we carried out with the mean with confidence interval of 95%, for cualitative variables: number and %. Univariate analysis of the results was carried out with the X2 test and t Student, the survival was expressed with Kaplan Meyer graphics and the logistic regression model.

Results: A total of 320 positive blood cultures were studied but only 272 blood cultures were considered true bacteriemia in 259 patients. The calculated incidence of significant episodes of bacteriemia per 1000 admissions/year was 13. The overall mortality was 22% whereas death attributable to bacteriemia was 16%. The mean age was 66.9 years (IC 95% 65-69), 59% episodes occurred in men. The 78% episodes occurred in patients hospitalized in medical services. 52% episodes were of nosocomial infection and 48% of community acquired infection. According to the severity of the underlying disease, 15% had fatal diseases and 35% episodes occurred in patients without underlying disease. According to the univariant analysis, the variables which where significantly associated with greater risk death were: etiology (fungus), septic shock, the inadequate antibiotic therapy, presence of extrinsic factors (central intravenous catheter, performance of invasive procedures, previous antimicrobial therapy) and the hospital stay of less than 10 days. According to the multivariable analysis showed that the factors remaining independent predictors of mortality were: septic shock (p < 0.0001, OR: 8), inadequate antimicrobial therapy (p < 0.005, OR: 6.7), existence of two or more extrinsic risk factors (p < 0.04).

Conclusions: The presence of septic shock was the most important variable which influenced in the mortality in our serie, together with inappropriate antimicrobian therapy and the association of various extrinsic risk factors. These variables could be modified partly, for this reason the aggressive hemodynamic control and the early and appropriate antibiotic therapy would be the support of the successful bacteriemia management.