The performance of blood cultures (BCs) relies on optimal sampling. Sepsis guidelines do not specify which sampling protocol to use but recommend two sets of BC bottles, each set containing one aerobic and one anaerobic bottle. For the single-site sampling (SSS) protocol, only one venipuncture is performed for all four bottles. The predominating multisite sampling (MSS) protocol implies that BC bottles are collected from two separate venipuncture sites. The aim of this study was to compare SSS and MSS. Primary outcomes were number of BC sets collected, sample volume, and diagnostic performance. This was a retrospective clinical study comparing BC results in an emergency department before and after changing the sampling protocol to SSS from MSS. All BC samples were incubated in the BacT/Alert BC system. The analysis included 5,248 patients before and 5,364 patients after the implementation of SSS. There was a significantly higher proportion of positive BCs sampled with SSS compared with MSS, 1,049/5,364 (19.56%) and 932/5,248 (17.76%), respectively (P = 0.018). This difference was due to a higher proportion of solitary BC sets (two BC bottles) in MSS. Analyzing only patients with the recommended four BC bottles, there was no difference in positivity. SSS had a higher proportion of BC bottles with the recommended sample volumes of 8-12 ml than MSS (P < 0.001). Changing the sampling protocol to SSS from MSS resulted in higher positivity rates, higher sample volume and fewer solitary BC sets. These advantages of SSS should be considered in future sepsis guidelines.
Keywords: bacteria; bloodstream infection; contamination; sepsis.