Blood Culture Results and Overtreatment Associated With the Use of a 1-Hour Sepsis Bundle

Some experts recommend using a 1-h sepsis bundle, but clinical data supporting this strategy are lacking.

We aimed to determine the rate of, and clinical predictors for, bacteremia for patients undergoing a 1-h sepsis bundle, and to determine the percentage of "code sepsis" patients who are ultimately diagnosed with sepsis or a bacterial infection.

This retrospective chart review evaluated code sepsis patients from three emergency departments (EDs) that utilize a 1-h sepsis bundle. The primary outcome was the rate of true-positive blood cultures. Secondarily, we analyzed various clinical factors using logistic regression analysis to determine which are associated with bacteremia.

Of the 544 code sepsis patients analyzed, 33.8% (95% confidence interval [CI] 29.9-38.0%) were ultimately diagnosed with sepsis, and 54.6% (95% CI 50.3-58.8%) were diagnosed with a bacterial infection. Exactly 7.0% (95% CI 5.0-9.5%) of the blood cultures performed were true positives. On multivariate logistic regression analysis, temperature > 38°C (100.4°F) or < 36°C (96.8°F), lactate > 4 mmol/L, and indwelling line/device were found to be positively associated with true-positive blood cultures.

In a group of code sepsis patients from facilities that use a 1-h sepsis bundle, the majority were ultimately not diagnosed with sepsis, and nearly half did not have a bacterial infection. A small minority of patients had bacteremia. Restricting blood culture ordering in patients with possible sepsis to only those who have increased risk for bacteremia could lead to a more judicious use of blood cultures.