The use of extracorporeal membrane oxygenation (ECMO) in critically ill adults has increased considerably in the past several years. Controversy remains regarding antimicrobial prophylaxis for patients receiving ECMO support. Patients receiving ECMO are unquestionably at higher risk for nosocomial infection, with 6.1% of neonates and 20.5% of adults acquiring a culture-proven infection during ECMO.1 Much of this risk most likely stems from the multiple other interventions patients simultaneously receive during ECMO, including mechanical ventilation, central venous access, surgical intervention, chest thoracostomy, wound drains, urinary catheterization, and arterial catheterization.2 A nosocomial infection has significant clinical consequences, including prolonged hospitalization, increased complications, and death.1 However, particularly in the adult population, the efficacy of specific antimicrobial prophylactic regimens is not fully established. Currently, the Extracorporeal Life Support Organization (ELSO) Infectious Disease Task Force does not recommend routine antimicrobial prophylaxis during ECMO,3 although to date, no systematic reviews of prophylaxis in ECMO have been performed. We sought to conduct a systematic review of the literature pertaining to antimicrobial prophylaxis during ECMO to form a basis for rational recommendations.
Data Sources and Search Strategies
We conducted a comprehensive search of several databases from each database’s earliest inception to January 6, 2014. The databases included Ovid Medline In-Process & Other Non-Indexed Citations, Ovid MEDLINE, Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus. The search strategy was designed and conducted by an experienced librarian with input from the study investigators. Controlled vocabulary supplemented with keywords was used to search for studies of antibiotic prophylaxis during ECMO. The detailed search strategy is found in the Supplementary Material (see Supplemental Digital Content 1, http://links.lww.com/ASAIO/A76).
Studies included in the review were human trials or reports evaluating newly acquired infections in patients receiving ECMO in intensive care unit settings. Any publication language was considered. References of all relevant articles, including reviews and editorials, were manually searched for other potentially relevant studies. The search strategy followed that proposed in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.4
Data abstracted from each study included the study design and location, modality of ECMO (venoarterial or venovenous), patient population (adult, child, or neonate), patient characteristics, study inclusion criteria, definitions for each type of infection, prophylactic regimens used, and specific isolates identified. If necessary, study authors were contacted for additional information. All articles were independently reviewed by at least two authors to verify appropriate study inclusion and abstraction.
The purpose of this study was to evaluate available evidence for antimicrobial prophylaxis for patients receiving ECMO support. As such, two study end-points of interest were identified. The first end-point was the relative incidence of all infections reported, to provide a rational basis for agent selection for prophylaxis. This was determined by pooling the observed specific isolates obtained from ECMO infections relative to the total time at risk, in days, while receiving ECMO. The second end-point was the overall efficacy of preventing acquisition of a nosocomial infection among studies that compared prophylactic regimens.
The search strategy yielded 1,280 records, which were screened by reviewing each abstract. Manual abstraction identified 7 additional potential articles. After removing duplicates, 1,259 unique articles remained, 42 of which were identified as potentially relevant and underwent full-text review. Ten were excluded for addressing only community-acquired infections, 12 for addressing only pharmacokinetics of antimicrobials during ECMO, seven for unrelated study questions, three for containing insufficient information, and one for being an animal study, leaving nine studies for inclusion in the systematic review. This information is summarized in the flow diagram (Figure 1).
One study meeting the inclusion criteria5 reported preliminary data from a subsequently published study.6 For the purposes of the current study, only the final version was used6; however, neither is included in the nine studies used for systematic review.
Characteristics of the studies included in the systematic review are shown in Table 1. One article was an ELSO registry study of all patients in the registry who received ECMO between 1998 and 2008.1 Eight were single-center retrospective studies.2,7–13 Four of these studies were ELSO centers and may have analyzed data included in the ELSO registry study.6–9 The other studies also took place during the ELSO period and may be, at least in part, included in the registry study. Six provided information regarding the total days supported on ECMO and were used to calculate rates of infections.2,7,8,11–13 All nine studies provided information on infection prevalence and were used in that analysis.1,2,7–13 The patient populations included different combinations of adults, children, and neonates, and study sizes ranged from 112 to 20,741 patients. Prophylactic drug use varied substantially.
Rates of Infection
For the first end-point, the relative incidence of all infections reported, the ELSO registry found an overall prevalence of infection of 11.7%, ranging from 7.6% in neonates to 20.9% in adults, with little variation during the 11-year span of the registry data.1 The rate of infection acquisition per 1,000 days of ECMO was 10.1 for neonates, 20.8 for children, and 30.6 for adults. Among two single-center studies, in which all subjects received prophylaxis, 1 showed a 16.1% prevalence rate7 and the other reported an overall infection rate of 18.4%.9
Bloodstream infections were predominant in most studies that reported the site of infection, followed by surgical site infections, urinary tract infections, and respiratory tract infections. Bloodstream infections ranged from a 2.6%2 to a 19.5%10 prevalence (Table 2), with no clear relationship between studies that used antimicrobial prophylaxis and those that did not. Rates of respiratory tract infections ranged from 1.4%8,12 to 15.8%12 (Table 2), again without clear effects of antimicrobial prophylaxis.
Gardner et al.6 evaluated the prevalence of fungal infections in patients receiving ECMO primarily for cardiac support. In that study, a fungal infection developed in 12% of patients, with surgical site infections reported most commonly.
Of note, most studies did not report the time frame of infections during the course of ECMO, and thus we cannot comment on the prevalence of early infection (possibly related to insertion practices) versus late infection. The largest study to examine this was the registry study of Bizzarro et al.,1 which found a marked increase in infections in patients supported on ECMO for more than 14 days, although the study did not provide information on the types of infection or microbiology associated with early versus late infections.
Effect of Prophylaxis
For the second end-point, the overall efficacy of preventing acquisition of a nosocomial infection, Hsu et al.2 retrospectively evaluated the effect of antimicrobial prophylaxis. They found no correlation between patients receiving prophylaxis and subsequent nosocomial infection. This finding was echoed by another retrospective study by Nagappala et al.,11 which compared infections in those routinely getting prophylaxis at their institution with infections in patients in the ELSO registry during a time when prophylaxis was not routine. Again, no reduced risk of infection was found related to prophylaxis.
The ELSO registry study found coagulase-negative Staphylococci and Candida spp. to be the most common causative organisms overall, followed by Pseudomonas aeruginosa and Staphylococcus aureus.1 Similar predominances of coagulase-negative Staphylococci and Candida were found in the individual center studies by O’Neill et al.9 and Schultze and Heulitt.7 Brown et al.10 and Nagappala et al.11 also noted a coagulase-negative Staphylococci predominance but reported few cases of Candida infection.
One study that only included adults found a predominance of Gram-negative organisms in bloodstream infections, surgical site infections, and respiratory tract infections, with fungal and Gram-positive organisms having a less prominent role.12
Bizzaro et al.1 found a significantly increased rate of death in those who acquired infection during ECMO (odds ratio, 1.91; 95% confidence interval, 1.75–2.08), as well as a prolonged duration of post-ECMO mechanical ventilation (112.24 vs. 77.81 h). These findings were mirrored in every single-center study reporting on these outcomes.2,7–9,11,12
The data previously described indicate that the risk of health care–acquired infection is high for patients receiving ECMO support. Health care–acquired infections are associated with increased mortality and morbidity. Therefore, infection prevention is critical to optimizing patient outcomes. There remains a lack of evidence supporting or refuting routine antimicrobial prophylaxis as part of a comprehensive management algorithm.2,13 These findings have resulted in an ELSO Infectious Disease Task Force statement concluding “There is NO data to support the routine use of antibiotics for patients on ECMO support, simply for prophylaxis, without specific culture or physiologic evidence of ongoing infection…. Prophylactic antibiotics for cannulation should follow standard principles of surgical prophylaxis, and a single dose [of antibiotics], or at the most 24 hours of coverage can be justified with either open or percutaneous cannulation techniques. Additional doses are not supported by any literature.”3
The subset of patients with open chests seems to be at higher risk for infection, and some authors advocate for prophylaxis in this setting.10 This, however, remains an expert opinion–based recommendation and cannot be graded as a strong recommendation. The ELSO Infectious Disease Task Force concurred with this, stating that the presence of mediastinitis is associated with high morbidity and mortality rates, and clinical judgment and patient factors should be considered when deciding whether prophylaxis is appropriate for a given patient with an open chest.3 Likewise, patients with other independent indications for antimicrobials, including active infection, prophylaxis for other specific conditions (e.g., endocarditis), or preoperative/postoperative prophylaxis, should have their therapies continued regardless of ECMO.
Adults are well known to have the highest rate of infections during ECMO, but the reason for this is not immediately clear. Host factors would seem to favor higher rates of infection in neonates with immature immune systems and compromised passive immunity, but this is clearly not what is observed. Factors that may influence this include the duration of ECMO support, modes of access, need for component changes, development of thrombus within the circuit, number of interfaces with ECMO fluid pathway, and the smaller bore of the catheter in children providing less surface area for contamination. Further data are needed to identify the reasons for lower rates of infection in children and neonates.
An effective approach to preventing ECMO-related infections will need a multifaceted, multidisciplinary approach, similar to the “bundle” practices used to prevent ventilator- and central line–associated infections. One of the included studies, Brown et al.,10 described applying a combination of education, electively preprimed circuits, and preferential neck site cannulation, in combination with prophylactic antibiotics for only the first 24 hours, which resulted in marked decreases in infection rates.
Future ECMO infection-prevention efforts may benefit from following the example of the central-line “bundle.” Although the microbiology of organisms may be different, and ELSO states that most infections during ECMO are secondary nonbloodstream infections,14 the principles for prevention of bloodstream infection should be similar. Examples of bundle elements not yet examined in ECMO that have proved useful in central lines include maximal barrier precautions, chlorhexidine skin preparation before cannulation,15 daily chlorhexidine bathing,16 and use of chlorhexidine-impregnated dressings around cannulae.17 A candidate “ECMO bundle” based on these existing guidelines is summarized in the appendix. As of yet, none of the associated professional societies have provided guidelines for care of this population.
Elements specific to ECMO meriting investigation could include understanding the relationship of infection to the number of circuit connectors or interventions, development of circuit thrombus, determining the optimum interval for changing oxygenators, adequacy of oxygen delivery, influence of open wounds, and site of cannulation. Another unresolved issue is whether linking other organ support systems, such as dialysis, to ECMO lines reduces or increases the risk of infection.
The most important limitation of this analysis is the paucity of available studies to synthesize. The relatively few studies, limited by retrospective and noncomparative designs, significantly decrease the strength of any recommendation. Moreover, the largest study was a registry study,1 which may have included patients analyzed in several of the single-center studies described. However, the lack of evidence supporting antimicrobial prophylaxis suggests that this should not be routine practice at this time, and equipoise exists for future high-quality studies examining the effects of this intervention.
In conclusion, current, albeit limited, evidence does not support routine prophylactic antimicrobial use in patients receiving ECMO support who do not have another indication for prophylaxis, such as perioperative prophylaxis, open chest coverage, or neutropenia. Identification of effective methods to prevent ECMO infection will require prospective studies and, potentially, the development of targeted, ECMO-specific, bundles. In the era of increasing antibiotic resistance and rising rates of Clostridium difficile infection, surveillance for and early treatment of infections should be considered as an alternative to routine broad-spectrum antibiotic prophylaxis in patients during ECMO.
The authors thank Larry J. Prokop, Mayo Clinic Libraries, for performing the literature search.
Appendix: The Extracorporeal Membrane Oxygenation Bundle
In all patients:
- Place ECMO cannulae in a sterile setting (e.g., operating room) whenever possible
- Follow proper insertion practices for all lines: hand hygiene, maximal barrier precautions, skin antisepsis, and avoidance of femoral cannulation whenever possible
- Assess daily the need for any invasive lines that may present an alternate nidus for infection, including ECMO cannulae
- Maintain clean, sterile dressings at all line sites
- Use clippers rather than razors for removal of hair at ECMO insertion sites and any other operative sites
- Maintain euglycemia
- Remove urinary catheter as soon as feasible
- Perform dressing changes in an aseptic manner with chlorhexidine-containing swabs every 48 h
- Use maximal barrier precautions for ECMO component changes
- Keep all stopcocks and needleless connectors sterilely covered
In intubated patients:
- Elevate head of bed to at least 30°
- Interrupt sedation daily and assess for readiness to extubate
- Maintain stress ulcer prophylaxis with sucralfate, H2 blocker, or proton-pump inhibitor
- Perform daily oral care with chlorhexidine-containing agent.
1. Bizzarro MJ, Conrad SA, Kaufman DA, Rycus PExtracorporeal Life Support Organization Task Force on Infections, Extracorporeal Membrane Oxygenation. . Infections acquired during extracorporeal membrane oxygenation in neonates, children, and adults. Pediatr Crit Care Med. 2011;12:277–281
2. Hsu MS, Chiu KM, Huang YT, Kao KL, Chu SH, Liao CH. Risk factors for nosocomial infection during extracorporeal membrane oxygenation. J Hosp Infect. 2009;73:210–216
4. Moher D, Altman DG, Liberati A, Tetzlaff J. PRISMA statement. Epidemiology. 2011;22:128, author reply 128
5. Gardner AH, Prodhan P, Stovall SH, et al. Fungal infections and antifungal prophylaxis in pediatric cardiac extracorporeal life support. J Thorac Cardiovasc Surg. 2012;143:689–695
6. Gardner AH, Prodhan P, Stovall SH, et al. Fungal infections and antifungal prophylaxis in pediatric cardiac extracorporeal life support. J Thorac Cardiovasc Surg. 2012;143:689–695
7. Schutze GE, Heulitt MJ. Infections during extracorporeal life support. J Pediatr Surg. 1995;30:809–812
8. Steiner CK, Stewart DL, Bond SJ, Hornung CA, McKay VJ. Predictors of acquiring a nosocomial bloodstream infection on extracorporeal membrane oxygenation. J Pediatr Surg. 2001;36:487–492
9. O’Neill JM, Schutze GE, Heulitt MJ, Simpson PM, Taylor BJ. Nosocomial infections during extracorporeal membrane oxygenation. Intensive Care Med. 2001;27:1247–1253
10. Brown KL, Ridout DA, Shaw M, et al. Healthcare-associated infection in pediatric patients on extracorporeal life support: the role of multidisciplinary surveillance. Pediatr Crit Care Med. 2006;7:546–550
11. Nagappala K, Gonzalez R, Lelli J, Langenburg S.. Pediatric extracorporeal membrane oxygenation: is antibiotic prophylaxis necessary? (Abstract). Crit Care Med. 2010;38(Suppl):A176
12. Sun HY, Ko WJ, Tsai PR, et al. Infections occurring during extracorporeal membrane oxygenation use in adult patients. J Thorac Cardiovasc Surg. 2010;140:1125–1132.e2
13. Aubron C, Cheng AC, Pilcher D, et al. Infections acquired by adults who receive extracorporeal membrane oxygenation: risk factors and outcome. Infect Control Hosp Epidemiol. 2013;34:24–30
14. ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support. 2013 Ann Arbor, MI, USA Extracorporeal Life Support Organization
15. Yilmaz G, Koksal I, Aydin K, Caylan R, Sucu N, Aksoy F. Risk factors of catheter-related bloodstream infections in parenteral nutrition catheterization. JPEN J Parenter Enteral Nutr. 2007;31:284–287
16. O’Horo JC, Silva GL, Munoz-Price LS, Safdar N. The efficacy of daily bathing with chlorhexidine for reducing healthcare-associated bloodstream infections: a meta-analysis. Infect Control Hosp Epidemiol. 2012;33:257–267
17. Safdar N, O’Horo JC, Ghufran A, et al. Chlorhexidine-impregnated dressing for prevention of catheter-related bloodstream infection: a meta-analysis*. Crit Care Med. 2014;42:1703–1713