Methods: We conducted a case control study (1 BSI: 3 controls) of ECMO cases from 1/1/08 to 10/1/13. Cases were included if ECMO was required for > 48 hours and excluded if a blood culture was positive within 24 hours of ECMO start. Differences in percentages were compared using Fisher’s exact test and duration of therapy was compared using the Wilcoxon rank-sum test.
Results: 12 of 207 subjects had a BSI. Indications for ECMO were: congenital heart disease (58%), respiratory failure (40%), and sepsis (10%). VA ECMO was employed in 96%. Among BSI cases, common pathogens were: Candida sp.-4, Pseudomonas sp.-2, and Staphylococcal sp.-2. The median time on ECMO prior to positive cultures was 210.5 hours (IQR 70, 231). Sex, age, ECMO duration, indication, circuit type, and use of additional invasive equipment were not risks for BSI. Prophylactic cefazolin was used in 58% of BSI cases (median duration 7 days) and 56% of controls (median duration 4 days). Prophylactic fluconazole was used in 17% of BSI cases and 11% of controls. Median total antimicrobial exposure was greater in BSI patients (31 vs 18 days; p=0.014). Commonly used antimicrobials were vancomycin (79%), cefepime (42%), gentamicin (31%), meropenem (23%), ampicillin (21%), and fluconazole (19%). Death occurred in 8 BSI cases and 16 controls (p=0.318).
Conclusion: No risk factors for developing ECMO associated BSI were identified. Prophylactic antimicrobials were used in over 50% of patients and overall antimicrobial use was greater in patients with a BSI. More data is needed to understand the risks associated with BSI and if prophylactic antimicrobial use is beneficial.
E. Molitor-Kirsch, None
J. Newland, Pfizer/Joint Commission: Grant Investigator , Research grant
RPSdiagnostics: Consultant , Consulting fee