OBJECTIVES: We aimed to improve the microbial diagnosis of first episodes of febrile neutropenia (FEFNs) since <30% of episodes are microbiologically documented. Consequently patients are usually treated by empirical antibiotic therapy.
METHODS: A prospective study evaluated a new workflow combining: (i) one 40 mL blood culture (BC) sampled from the central venous catheter; (ii) immediate incubation in an automated BC system on the ward; (iii) direct detection of microbial DNA in blood; and (iv) identification and susceptibility testing using rapid methods performed directly on positive BC bottles. Patients were also sampled for the standard workflow with two BC sets incubated in the central laboratory and assessed by classical procedures.
RESULTS: One hundred and twenty consecutive FEFNs were included (February 2008-March 2009). The new workflow was as sensitive as the standard workflow, with BC positivity rates of 30% (36/120) and 28% (34/120), respectively (McNemar's χ(2) =0.67, P=0.41). Direct DNA detection was positive in nine episodes (7.5%) that were also positive in BC. The new workflow provided microbiological results significantly earlier than the standard workflow, with a shorter time to BC positivity (median 12 h 31 min, range 7 h 55 min-25 h 37 min versus median 13 h 01 min, range 9 h 31 min-43 h 33 min, P=0.004) and shorter turnaround times for identification and susceptibility testing, with most of the results obtained <24 h after BC sampling. We retrospectively estimated that the new workflow would lead to earlier adequacy of antimicrobial therapy in 30% of documented cases.
CONCLUSIONS: Our new process improved the microbiological diagnosis in FEFNs. Cost effectiveness needs to be tested.