Abstract
BACKGROUND: Blood cultures (BC) are critical for the diagnosis of bloodstream infections, pathogen identification, and resistance testing. Guidelines recommend a blood volume of 8-10 mL per bottle as lower volumes result in decreased sensitivity. We aimed to evaluate factors for non-adherence to recommended volumes and assess the effects on diagnostic performance. METHODS: From February to April 2020, we measured collected blood volumes by weighing all BC containers from inpatient samples at the University Hospital Basel. Information on BC volumes was merged with clinical and microbiological data, as well as nursing staff schedules. We analyzed factors associated with (i) BC sampling volume, (ii) reaching recommended volumes (≥8 mL), (iii) BC positivity, and (iv) time to positivity using linear and generalized linear mixed effect models. RESULTS: We evaluated a total of 4'118 BC bottles collected from 686 patients. A total of 1'495 (36.3%) of all bottles contained the recommended filling volume of ≥8 mL. Using a central venous and arterial catheter for drawing blood resulted in an increase of filling volume by 0.26 mL (95% CI 0.10, 0.41) and 0.50 mL (95% CI 0.31, 0.69) compared to peripheral venipuncture, respectively. Each additional nursing staff working at the time of blood drawing was associated with 6% higher odds of achieving the recommended filling volume. We found no significant correlation between the filling volume and the positivity rate. CONCLUSION: Our results indicate critical pre-analytical quality markers linked to BC collection procedures to reach recommended collection volumes. No significant impact on the positivity rate was found.