The use of extracorporeal membrane oxygenation (ECMO) in critically ill adults is increasing with no guidelines for antimicrobial prophylaxis. Patients on ECMO are at a high risk for infections, 6.1% of neonates and 20.5% of adults. An Extracorporeal Life Support Organization (ELSO) Infectious Disease Task Force statement concludes that no additional antibiotic coverage is needed for patients on ECMO. Since patients on ECMO are severely ill, providers tend to prescribe empiric antibiotics. To guide rational antibiotic therapy we introduce an ECMO antimicrobial protocol on 7/1/2014 and report it's impact.
We conducted a retrospective review of 294 patients on ECMO between 7/1/2011 and 7/1/2017. The ECMO antimicrobial protocol was introduced on 7/1/2014. We had a cohort of 133 patients before and 161 patients after the implementation of protocol. We evaluated days of antimicrobial use, antibiotic-free days and days of individual antimicrobial use, adjusted for APACHE scores and ECMO duration.
Total days of antimicrobial use after the protocol decreased from 2508 to 2186 days (p=0.01) with statistically significant reduction of individual antimicrobials; vancomycin (407 to 266 p<0.03), cefepime (196 to 165 p<0.06), along with reduced days of anidulafungin, caspofungin, fluconazole, meropenem, and daptomycin. However, when adjusted for mean days on ECMO 7 (4-14) before as compared to 5 (3-9.5) after (p<0.0119), “antimicrobial free days” actually reduced after implementation of the protocol. Early trends of improved stewardship were off-set when time frame and number of patients were increased. Despite this, no difference was seen in rate of nosocomial infections, with increased rates seen for Clostridium difficile (0 vs 4, p<0.06).
“Protocolization” and standardization of antimicrobial recommendations for patients on ECMO led to reduction in use of specific antibiotics but paradoxically increased overall antibiotic use. We are in the process of emphasizing compliance with this protocol, which will be followed by implementation of a more restrictive protocol. We will do a step wedge randomized control prospective analysis to evaluate compliance differences between the medical and surgical critical care services, and the impact on patient outcomes.
P. Sampathkumar, None
J. K. Bohman, None
J. C. O'horo, None