Can bacteremia be predicted in surgical intensive care unit patients?

To determine which clinical features are associated with bacteremia in a SICU. To determine if infections are identified prior to bacteremia via culturing of other body fluids. To determine if antibiotic regimens are changed after the results of the blood culture were obtained.

A retrospective, unit-based, case control study.

A 10 bed SICU in a 552-bed, tertiary care and Level I Trauma center.

All SICU patients with one or more positive blood cultures over a 2 year period (n = 24) were matched by diagnosis, procedure, and age to SICU patients with negative blood cultures (n = 48).

Bacteremic and control patients had similar APACHE II scores though death was more likely in bacteremic patients (p < 0.05) and they had higher hospital charges (p < 0.02). There was no difference in any of the clinical variables studied (minimum and maximum temperature, maximum white blood cell count, minimum mean arterial blood pressure) between the bacteremic and control groups on the days leading up to and the day of the positive blood culture. Coincident infections of lung, bladder, wound, and central venous catheters were identified in 42% of bacteremic patients. The identification of organisms found in the blood had a direct impact on the antibiotic regimen of 54% of the bacteremic patients.

A better screen for obtaining blood cultures in this SICU was not identified. If antibiotics are begun empirically before the results of blood cultures are known, the results of other body fluid cultures can be used to guide therapy initially. However, the data obtained from positive blood cultures was often helpful in changing empirical therapy. Therefore, blood cultures remain important diagnostic tools.