Extracorporeal membrane oxygenation (ECMO) plays a critical role in the intensive care setting.1 The use of ECMO is increasing globally, and this modality is considered suitable for the treatment of acute respiratory failure, cardiac arrest, and cardiogenic shock.2 Furthermore, the potential indications for ECMO continue to increase.3 Technological advances have resulted in ECMO becoming less invasive and easier to use. However, despite its advantages and growing clinician experience, ECMO is still associated with high rates of mortality and complications. Infection during ECMO support remains a major concern due to its association with high morbidity and mortality rates.4,5
Vascular devices can become infected by microorganisms through insertion sites or by bacteremia seeded from a primary infection site. Various intravascular devices have been shown to be prone to microbial biofilm formation, which is a prerequisite for the persistence of many severe or recurrent infections.6 Bacteria can adhere to ECMO catheters, proliferate, and form biofilms that subsequently release bacteria into the circulation and cause rapid septic deterioration, which is a critical clinical problem.7 Previously, we reported that ECMO catheters may be contaminated in patients with a bloodstream infection (BSI).8 However, whether there is a causal relationship between BSI and catheter colonization remains unclear, and questions regarding biofilm formation on ECMO catheters, which are maintained at higher blood flow rates than other intravascular catheters, remain to be answered.
This study was performed to confirm the presence of biofilms on the surfaces of ECMO catheters from patients with acute cardiorespiratory failure and to evaluate the impact of infection during ECMO on biofilm formation. In addition, clinical factors associated with biofilm formation were investigated.
Materials and Methods
Study Design and Population
Between January 2015 and October 2016, 122 adult patients underwent ECMO at our center for respiratory or cardiac support. Of these, 81 consecutive patients were prospectively enrolled in this study. The patients were divided into two groups, an infection group (51 patients) and a control group (30 patients), based on the presence of a systemic infection during ECMO support (Figure 1). The study was approved by our institutional review board, and all subjects provided written informed consent before enrollment.
Systemic infection was defined as the confirmed presence of organisms in one or more blood, bronchoalveolar lavage, or urine culture in the presence of clinical signs of infection, including fever, tachycardia (heart rate > 100/min), hypotension, hyperventilation, increased numbers of mature and immature circulating neutrophils, thrombocytopenia, and organ failure. Blood, sputum, and urine cultures were obtained regularly twice weekly from all patients. In addition, all cultures were repeated whenever a new infection episode was suspected clinically.
Localized catheter infection.
Localized catheter infection was defined as the presence of any microbiological agent and biofilm on an extracted catheter, which presented as a culture-positive biofilm.
Catheter-related infection (CRI) was assumed if the same microbiological agent was found on an ECMO catheter and by blood culture.
The management protocol for ECMO was conducted in accordance with Extracorporeal Life Support Organization (ELSO) guidelines,9 and bedside care for patients undergoing ECMO was delivered by a multidisciplinary intensive care team. All catheters were managed using a standardized protocol and were inserted peripherally by an experienced cardiothoracic surgeon using the Seldinger approach under ultrasound guidance. For venovenous ECMO, the return cannula was placed in the internal jugular vein (IJV) and the drainage cannula was placed in the femoral vein (FV). For venoarterial ECMO, the return cannula was placed in the femoral artery (FA) and the drainage cannula was placed in the FV. Aseptic precautions for all device insertions included the use of a full-size drape, mask, cap, gown, and sterile gloves. Chlorhexidine (2%) was used for skin antisepsis. The ECMO system consisted of a polymethylpentene fiber oxygenator system (QUADROXPLS; Maquet Inc., Hirrlingen, Germany) with simplified Bioline-coated circuits (Maquet Inc.). All patients were supported using centrifugal pumps (Maquet Inc.). No defined regimen was used for antimicrobial prophylaxis. Standardized empirical antibiotic therapy with broad-spectrum drugs was initiated according to anti-infection guidelines for suspected or confirmed infections when deemed necessary by the attending intensivist. During ECMO support, antibiotics were adjusted according to the clinical course or culture results.
ECMO Catheter Collection and Preparation
Catheter samples were harvested aseptically, collected prospectively, and cultured after discontinuation of ECMO support. After removal, sections (5 cm) were cut from the intravascular region, split longitudinally, and transported immediately to the laboratory.
Bacterial Culture and Identification
Catheter pieces were rolled across the surfaces of standard 5% sheep blood agar plates and MacConkey agar plates immediately after collection. After roll-plating, both plates were incubated for 48 hours at 35°C in a 5% CO2/air atmosphere. An automated VITEK 2 system (BioMérieux, St. Louis, MO) was used to identify isolates and for susceptibility testing. Clinical breakpoints recommended by the Clinical and Laboratory Standards Institute (CLSI) were used to define susceptibility and resistance. Carbapenem-resistant Acinetobacter baumannii (CRAB) was defined as an A. baumannii isolate with a minimum inhibitory concentration (MIC) ≥ 16 μg/ml for imipenem. Methicillin-resistant Staphylococcus aureus (MRSA) was defined as an S. aureus isolate with an MIC ≥ 4 μg/mL for oxacillin. Methicillin-resistant Staphylococcus epidermidis was defined as an S. epidermidis isolate with an MIC ≥ 0.5 μg/mL for oxacillin.
Fluorescence Microscopy and Confocal Laser Scanning Microscopy
Catheters were fixed in 4% formaldehyde, embedded in paraffin, and sliced. Biofilms were examined after staining with SYTO 9 nucleic acid stain and Concanavalin A (Con A) extracellular matrix stain. Live bacteria were stained with SYTO 9 (green) and extracellular polymeric substances with Con A (red).10 Samples were observed under a fluorescence microscope (Axio Observer Z1; Carl Zeiss, Oberkochen, Germany) equipped with a 488 nm excitation filter and a 594 nm filter, and by Confocal Laser Scanning Microscopy (Olympus Fluoview FV 1000; Olympus, Tokyo, Japan).
Transmission Electron Microscopy
Catheter samples were divided into small pieces (ca. 1 mm3) and fixed in 2.5% glutaraldehyde (pH 7.4) in 0.1 M phosphate buffer for 2 hours at room temperature. After washing three times with buffer, specimens were postfixed in 1% osmium tetroxide for 1 hour and dehydrated through a graded alcohol series (50%, 60%, 70%, 80%, 90%, and 100% ethanol). Tissues were then cleared in propylene oxide and embedded in Araldite epoxy resin. After polymerization in an oven at 56°C for 48 hours, several semi-thin sections were prepared and stained with toluidine blue. Sections were examined under a light microscope (BH2; Olympus). Thin sections (0.5 μm) were stained with uranyl acetate and lead citrate, and examined by transmission electron microscopy (TEM; H-7600; Hitachi Science System Ltd., Hitachi, Japan). All laboratory examinations were carried out in a blinded manner.
Statistical analyses were performed using R software version 3.0.1 (http://www.R-project.org). Continuous variables are described as means ± standard deviations and group comparisons were performed using the t test. Categorical variables are presented as numbers (%) and were compared using the χ2 test. A binary logistic regression analysis was used to identify individual predictors of biofilm formation. Two-tailed p values <0.05 were considered to reflect statistical significance. Logistic regression results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). Variables significant by the univariate analysis were selected for further analysis; particular attention was paid to clinical and procedural factors rendering biofilm formation more likely. Variables with a p value <0.05 were subjected to a multivariate analysis, which was conducted using the backward stepwise (likelihood ratio) method with entry and removal p values set at 0.05.
The 81 patients included in the study consisted of 51 in the infection group and 30 in the control group. The baseline characteristics of the patients are described in Table 1. In total, 54 (66.7%) patients were male, and the mean overall patient age was 52.7 years. Overall, 61 (75.3%) patients required ECMO for acute respiratory distress syndrome, while 20 (24.7%) required ECMO for cardiogenic shock. The mean the ratio of arterial oxygen partial pressure to fractional inspired oxygen (PaO2/FiO2) (PF ratio) before initiation of ECMO was 64.5, and the value was significantly lower in the infection group than in the control group (57.6 ± 17.1 vs. 75.8 ± 34.4, respectively, p = 0.010). The mean duration of ECMO support was 14.8 days, and the period was significantly longer in the infection group than in the control group (19.6 ± 14.5 vs. 6.6 ± 3.6, respectively, p < 0.001). The mean average blood flow rate for ECMO was 3.0 L/min, and was also significantly higher in the infection group than in the control group (3.2 ± 0.6 vs. 2.7 ± 0.5, respectively; p < 0.001).
Clinical Characteristics and Microbiology of Infections
In the infection group (n = 51), 42 (82.4%) patients had pneumonia and 28 (54.9%) had a BSI. The infection categories and causative microorganisms are listed in Table 2. The mean time to documented infection after initiation of ECMO in these 51 patients was 9.7 days, and the mean duration of ECMO support after a documented infection was 9.3 days.
The predominant causative organism was CRAB (n = 19, 37.3%). Also dominant were Gram-positive pathogens such as S. aureus and S. epidermidis. However, Gram-negative pathogens, including Pseudomonas aeruginosa, Chryseobacterium indologenes, Escherichia coli, Klebsiella pneumonia, Stenotrophomonas maltophilia, Burkholderia cepacia, Serratia marcescens, Saccharomyces cerevisiae, and Candida species, were isolated from cultures.
Detection and Imaging of Biofilms
Biofilms were observed on the subcutaneous sections of ECMO catheters by fluorescence microscopy in 28 patients (34.6%). Biofilms were detected in 43.1% (22/51) of patients in the infection group and in 20.0% (6/30) of controls (p = 0.034; Figure 1).
Biofilms were only recovered from catheter lumens, on which they showed a heterogeneous distribution. On SYTO 9 staining, most were green and therefore viable. The bacteria in the biofilms were generally clustered in small groups, as shown in Figure 2. Most of the bacteria detected were rod-like, and CLSM showed that some were surrounded by matrix-like substances (Figure 3). Several bacteria found in the cultures had a cocci-like appearance, suggesting Staphylococcus species. The bacteria were strongly adherent to the catheter surfaces, which appeared (grossly) to be clean. However, TEM revealed amorphous material attached to the catheters (Figure 4).
Microorganisms in the Biofilms
Microorganisms were identified from the biofilms by traditional culture methods in a total of 16 patients (15 FV catheters, 8 IJV catheters, and 1 FA catheter; Table 3). The culture-positive biofilm rates were significantly higher in the infection group than in the control group (27.5% [14/51] vs. 6.7% [2/30], respectively; p = 0.023).
Carbapenem-resistant A. baumannii was the most common microorganism identified (n = 9, 56.3%); other organisms included Gram-positive methicillin-resistant pathogens (S. epidermidis, n = 3, 18.8%; and S. aureus, n = 2) and Candida species (n = 2). Of the nine patients in whom CRAB was identified, seven showed the same pathogen in blood, BAL, and catheter cultures, and two patients showed the same pathogen in BAL and catheter cultures. Three patients in whom methicillin-resistant S. epidermidis was identified showed the same pathogen in blood and catheter cultures. Of the two patients with MRSA, one patient showed the same pathogen in BAL and catheter cultures, and the other patient showed MRSA only in the catheter without a systemic infection. Candida parapsilosis was identified only from the catheter in one patient; the other patient showed Candida tropicalis only in the catheter without systemic infection.
Biofilm Formation and Risk Factors
A number of variables, including systemic infection, duration of ECMO, ECMO flow, and etiologic organisms (CRAB, MRSA, and methicillin-resistant S. epidermidis), were evaluated to identify independent risk factors for biofilm formation in ECMO catheters (Table 4). A univariate analysis indicated that ECMO flow, systemic infection, and CRAB were associated with biofilm formation (OR: 1.00, 95% CI: 1.00–1.00, p = 0.007; OR: 3.03, 95% CI: 1.06–8.69, p = 0.039; OR: 9.60, 95% CI: 2.94–31.30, p < 0.001, respectively). All these variables were included in a binary logistic regression analysis model. However, a multivariate analysis showed that only CRAB was independently associated with ECMO catheter biofilm formation (OR: 9.60, 95% CI: 2.94–31.30; p < 0.001).
In this study, biofilms were found on ECMO catheters retrieved from 28 of 81 (34.6%) patients on ECMO. The prevalence of a biofilm was higher in the infection group than in the noninfected control group. The culture-positive rate of biofilm was 57.1% (16/28), which was also higher in the infection group than in the control group (27.5% vs. 6.7%, respectively, p = 0.023). Although the causal relationship between infection and biofilm formation is unclear, infection during ECMO support was significantly associated with biofilm formation and ECMO catheter infection. In addition, systemic CRAB infection was independently associated with the presence of a biofilm in ECMO catheters.
Biofilm formation has been proposed to be a major cause of CRI.11,12 When microbes form mature biofilms on catheters in patients, the infections become resistant to antibiotics and can develop into refractory or systemic conditions because biofilms act as pathogen reservoirs.13,14 In this study, catheter and blood cultures from 10 of 16 patients with a culture-positive biofilm were positive for the same pathogens (three methicillin-resistant S. epidermidis and seven CRAB). These 10 patients were assumed to have a CRI, and the other six were taken to have a localized catheter infection. Thus, a biofilm may be a major contributor to an ECMO catheter infection and represent a potential conduit for systemic infection and disease. Catheter removal is the conventional fundamental treatment for catheter-associated bacteremia, but the removal of a compromised ECMO catheter is not possible in patients with unstable cardiopulmonary status.15 Therefore, infection remains a difficult problem in patients on ECMO support.
In previous studies, the duration of ECMO support was the most important risk factor for acquired infection development, and the rate of BSI was found to increase with the duration of support.16 However, in the current study, prolonged catheterization did not increase the prevalence of a biofilm. One explanation is that biofilm formation may depend on the functions of host-produced conditioning films by platelets, plasma, and tissue proteins in addition to the exposure duration.17 Therefore, systemic infection can affect biofilm formation regardless of ECMO duration because systemic infection can affect the condition of the blood.18
In the current study, A. baumannii was the pathogen most frequently isolated from catheters, which was independently associated with biofilm formation. Acinetobacter baumannii is increasingly becoming a cause of bacterial infections in intensive care units.19–23 Although little is known about the pathogeneses of diseases caused by A. baumannii, its ability to persist in the environment on abiotic surfaces has been linked to biofilm formation.24,25 The nature of the substrate contributes to the ability of A. baumannii to adhere to and form a surface biofilm,26 and because ECMO oxygenators, circuits, and catheters are mainly composed of polycarbonate and polypropylene, A. baumannii cells may adhere and persist on these inanimate surfaces.27
Interestingly, high blood flow rates in ECMO tended to produce denser, more tenacious biofilms, and the biofilms formed were not apparent on visual inspection.28 Furthermore, biofilms (e.g., Candida species and MRSA) were observed in the control group without evidence of systemic infection. Potential contamination of the catheter during sampling cannot be entirely excluded, but there is a possibility of localized catheter infection without systemic infection. Bacteria can be embedded within the biofilm on a catheter, and the bacteria in these sessile biofilms cannot be detected in the blood until the biofilm has detached or dispersed from the catheter. Thus, bacterial biofilm formation and a localized catheter infection may proceed without any clinical signs of systemic infection. The defense of biofilms in ECMO catheters is crucial to successful infection management. Further study regarding the use of preemptive antibiotics during ECMO support to prevent a biofilm-related infection is required. Investigations are required to identify effective mechanisms for the prevention and control of biofilm formation on ECMO catheters.
Extracorporeal membrane oxygenation is of growing importance and is being increasingly used in the critical care field.29 Most patients under ECMO are critically ill, and are therefore predisposed to preexisting or acquired infections. Although the results of the current study do not indicate a clear causal correlation between infection during ECMO support and ECMO catheters, they do suggest that ECMO catheters are a potential source of infection. This study has several limitations due to the small sample size from a single institute. In particular, the high proportion of A. baumannii may be related to the high burden of A. baumannii infection in intensive care units in Asia, especially Korea. However, the increasing global prevalence of intensive care unit–acquired infections caused by A. baumannii highlights the need for stricter infection control during ECMO support.30 In addition, as all patients had been administered antibiotics, bacterial growth in the catheters may have been suppressed. However, this is the first study to evaluate the prevalence of biofilms on catheters in a high-flow ECMO setting and the relationships between clinical and procedural variables and the prevalence of catheter biofilms. This report provides meaningful information regarding the clinical and microbiological features of ECMO CRIs. Therefore, ECMO specialists should be aware of potential ECMO catheter contamination in patients with an infection, reassess the need for ECMO, and seriously consider the timing of weaning patients off ECMO. Additional studies to determine the clinical significance of biofilm formation on ECMO catheters and to prevent ECMO CRIs are required.
1. Yeo HJ, Jeon D, Kim YS, Cho WH, Kim DVeno-veno-arterial extracorporeal membrane oxygenation
treatment in patients with severe acute respiratory distress syndrome and septic shock. Crit Care 2016.20: 28,
2. Del Sorbo L, Cypel M, Fan EExtracorporeal life support for adults with severe acute respiratory failure. Lancet Respir Med 2014.2: 154–164,
3. Smith M, Vukomanovic A, Brodie D, Thiagarajan R, Rycus P, Buscher HDuration of veno-arterial extracorporeal life support (VA ECMO) and outcome: An analysis of the Extracorporeal Life Support Organization (ELSO) registry. Crit Care 2017.21: 45,
4. Zangrillo A, Landoni G, Biondi-Zoccai G, et alA meta-analysis of complications and mortality of extracorporeal membrane oxygenation
. Crit Care Resusc 2013.15: 172–178,
5. Stefano B, Stefano DB, Vittorio S, et alInfections during extracorporeal membrane oxygenation
: Epidemiology, risk factors, pathogenesis and prevention. Int J Antimicrob Agents 2017; 50: 9–16.
6. Mack D, Rohde H, Harris LG, Davies AP, Horstkotte MA, Knobloch JKBiofilm formation in medical device-related infection
. Int J Artif Organs 2006.29: 343–359,
7. Müller T, Lubnow M, Philipp A, et alRisk of circuit infection
in septic patients on extracorporeal membrane oxygenation
: A preliminary study. Artif Organs 2011.35: E84–E90,
8. Kim DW, Yeo HJ, Yoon SH, et alImpact of bloodstream infections on catheter colonization during extracorporeal membrane oxygenation
. J Artif Organs 2016.19: 128–133,
10. Schlafer S, Meyer RLConfocal microscopy imaging of the biofilm
matrix. J Microbiol Methods 2017.138: 50–59,
11. Yousif A, Jamal MA, Raad IBiofilm-based central line-associated bloodstream infections. Adv Exp Med Biol 2015.830: 157–179,
12. Zhang L, Gowardman J, Morrison M, Runnegar N, Rickard CMMicrobial biofilms associated with intravascular catheter-related bloodstream infections in adult intensive care patients. Eur J Clin Microbiol Infect Dis 2016.35: 201–205,
13. Donlan RM, Costerton JWBiofilms: Survival mechanisms of clinically relevant microorganisms. Clin Microbiol Rev 2002.15: 167–193,
14. Lambert ML, Suetens C, Savey A, et alClinical outcomes of health-care-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: a cohort study. Lancet Infect Dis 2011.11: 30–38,
15. Baron EJ, Miller JM, Weinstein MP, et alA guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2013 recommendations by the Infectious Diseases Society of America (IDSA) and the American Society for Microbiology (ASM)(a). Clin Infect Dis 2013.57: e22–e121,
16. Kim GS, Lee KS, Park CK, et alNosocomial infection
in adult patients undergoing veno-arterial extracorporeal membrane oxygenation
. J Korean Med Sci 2017.32: 593–598,
17. Raad I, Costerton W, Sabharwal U, Sacilowski M, Anaissie E, Bodey GPUltrastructural analysis of indwelling vascular catheters: A quantitative relationship between luminal colonization and duration of placement. J Infect Dis 1993.168: 400–407,
18. Walborn A, Hoppensteadt D, Syed D, et alBiomarker profile of sepsis-associated coagulopathy using biochip assay for inflammatory cytokines. Clin Appl Thromb Hemost 2017:DOI:10.1177/1076029617709084.
19. Fournier PE, Richet HThe epidemiology and control of Acinetobacter baumannii
in health care facilities. Clin Infect Dis 2006.42: 692–699,
20. Talbot GH, Bradley J, Edwards JE Jr, Gilbert D, Scheld M, Bartlett JGAntimicrobial Availability Task Force of the Infectious Diseases Society of America: Bad bugs need drugs: An update on the development pipeline from the Antimicrobial Availability Task Force of the Infectious Diseases Society of America. Clin Infect Dis 2006.42: 657–668,
21. Weber BS, Harding CM, Feldman MFPathogenic acinetobacter: From the cell surface to infinity and beyond. J Bacteriol 2015.198: 880–887,
22. Greene C, Wu J, Rickard AH, Xi CEvaluation of the ability of Acinetobacter baumannii
to form biofilms on six different biomedical relevant surfaces. Lett Appl Microbiol 2016.63: 233–239,
23. Chung DR, Song JH, Kim SH, et alAsian Network for Surveillance of Resistant Pathogens Study Group: High prevalence of multidrug-resistant nonfermenters in hospital-acquired pneumonia in Asia. Am J Respir Crit Care Med 2011.184: 1409–1417,
24. Longo F, Vuotto C, Donelli GBiofilm formation in Acinetobacter baumannii
. New Microbiol 2014.37: 119–127,
25. Choi AH, Slamti L, Avci FY, Pier GB, Maira-Litrán TThe pgaABCD locus of Acinetobacter baumannii
encodes the production of poly-beta-1-6-N-acetylglucosamine, which is critical for biofilm
formation. J Bacteriol 2009.191: 5953–5963,
26. Brandão GA, Pereira AC, Brandão AM, de Almeida HA, Motta RRDoes the bracket composition material influence initial biofilm
formation? Indian J Dent Res 2015.26: 148–151,
27. Lehle K, Philipp A, Gleich O, et alEfficiency in extracorporeal membrane oxygenation
-cellular deposits on polymethylpentene membranes increase resistance to blood flow and reduce gas exchange capacity. ASAIO J 2008.54: 612–617,
28. Beyenal H, Lewandowski ZInternal and external mass transfer in biofilms grown at various flow velocities. Biotechnol Prog 2002.18: 55–61,
29. Mullany D, Shekar K, Ziegenfuss M, et alThe effects of the introduction of an adult ECMO program on statewide referral patterns, casemix and outcomes in patients with acute respiratory distress syndrome or pneumonia. Intensive Care Med 2017.43: 1065–1066,
30. Antunes LC, Visca P, Towner KJAcinetobacter baumannii: Evolution of a global pathogen. Pathog Dis 2014.71: 292–301,