Follow-up blood cultures in Staphylococcus aureus bacteremia: a probability-based optimization.

ABSTRACT

Staphylococcus aureus bacteremia (SAB) is a relevant finding which prompts a thorough diagnostic work-up. Follow-up blood cultures (BC) are essential in this work-up. We investigate the probability of detecting an ongoing bacteremia after initiation of active therapy according to the number of BC taken at key time points. A retrospective analysis of all patients with SAB in a 6-year period was performed. Total number of BCs taken and the positivity was registered for each day after start of therapy. A positivity-rate was corrected using a logistic mixed effects model. Observed detection frequencies were applied to calculate detection probabilities using binomial distributions. Three hundred and seventeen cases were withheld for analysis. A BC bottle positivity rate of 66.7% was found 1 day after initiation of active therapy, which decreased to 48.5% on day 4. When using 1 set of FU-BC, 73.4% of persisting SABs are detected. To maintain a probability of detection of ≥ 90%, 2 BC sets should be taken on day 2 and day 4 after start of therapy. In 10 of 109 patients with positive FU-BC, skip phenomena were registered, with a significant higher proportion in patients with < 4 BC bottles taken (14%) than when ≥ 4 BC bottles were taken (4.1%). We recommend taking 2 BC sets on days 2 and 4 after start of therapy in order to detect ≥ 90% of persisting SABs, limiting skip phenomena and blood volume required. We strongly advice against taking a single BC set as follow-up for SAB.