Nosocomial Infections in Trauma Patients Receiving Extracorporeal Membrane Oxygenation

Abstract

Background:

The risk of infections in acute trauma patients receiving extracorporeal membrane oxygenation (ECMO) has not been well defined, but this population is among the sickest in the hospital. This study characterizes the blood and respiratory pathogens in trauma patients receiving ECMO and tests the hypothesis that trauma patients receiving ECMO pose a unique risk for nosocomial infections.

Methods:

All trauma patients (n = 50) who required ECMO at a level 1 trauma center between July 2014 and September 2023 were retrospectively reviewed. Blood and respiratory samples were examined for pathogens. Burn injuries were excluded from this study.

Results:

Most were male (88%) and sustained blunt injuries (60%), of which the most common mechanism was motor vehicle collision (37%). The median intensive care unit (ICU) length of stay was 26 days (interquartile range [IQR]: 12–54), median ventilator days was 22 (IQR: 9–51), and median length of ECMO treatment was eight days (IQR: 4–18). The most common ECMO type was veno-venous (80%). Blood stream and/or respiratory infections were detected in 22 (44%) patients. The most common was respiratory infection (70.2%). Gram-negative pathogens were most common in both blood stream (53.3%) and respiratory infections (75%). Among the blood stream pathogens, 20% were multi-drug resistant. Among the respiratory pathogens, 41.7% were multi-drug resistant. The median number of days from ECMO cannulation to blood stream infection was 10 days (IQR: 4–16) and the median number of days to respiratory infection was seven days (IQR: 3–11). Mortality was 22.7% (n = 5) among patients who had a respiratory and/or blood stream infection and 60% (n = 17) in ECMO patients who had no infection (p = 0.007).

Conclusion:

Almost half of trauma patients receiving ECMO had nosocomial infections, which is comparable with the infection rate of trauma patients in the ICU not on ECMO. Larger studies are needed to further assess infection risk in these patients and provide guidelines to mitigate this risk.

Introduction

Trauma is a leading cause of death in the United States.¹ Exsanguination and traumatic brain injury are the two most common primary causes of death, whereas respiratory failure, hemorrhagic shock, and multi-organ failure are the most common secondary causes of death.² The utilization of extracorporeal membrane oxygenation (ECMO) in trauma has historically been low because of the risk of bleeding secondary to the anticoagulation required to sustain the ECMO circuit.³ However, the adoption of an ECMO practice with the use of low anticoagulation or no systemic anticoagulation in trauma patients has become feasible.^4–7 As a result, the use of ECMO as a salvage therapy in trauma has increased in recent years with reported survival rates exceeding 60%.^8,9 The most common indication for ECMO in trauma patients is acute respiratory distress syndrome (ARDS) and the most common complications are renal, infectious, and thrombotic.^9,10

Given the recent expansion of ECMO in trauma care, there is still a need to further define the complications associated with its use in the trauma population, especially the infection risk.^11–18 The hypothesis that trauma patients receiving ECMO pose a unique risk for nosocomial infections is tested here.

Methods

This retrospective study was approved by the Institutional Review Board with waiver of informed consent. Electronic medical records were reviewed from all trauma patients who received ECMO at a level-1 trauma center from July 2014 to September 2023, excluding burn patients. Pathogens in blood or respiratory samples taken from the day after ECMO cannulation to the day of ECMO decannulation were recorded. Infections were defined as positive cultures with suspicious organisms isolated for the first time or recurred after ECMO cannulation. These cultures were included in our analysis regardless of the quantitative measure of their organisms, which aligns with institutional policy. If the culture taken during the ECMO period contained a pathogen present in earlier samples before ECMO cannulation, this pathogen was not included in the study analysis. However, if a pathogen acquired multi-drug resistance after ECMO cannulation or recurred in later samples after having been presumably treated during ECMO cannulation, it was included in the study. Urine culture data were not explored as the rate of urinary tract infections in this population was thought to not be impacted by ECMO use and the rates are exceedingly low. Multi-drug resistant was defined as resistance to three or more classes of antibiotic agents.¹⁹ Patient demographics such as age, gender, and race along with injury type, ECMO modality, ECMO cannulation site and location, and outcomes such as length of intensive care unit (ICU) stay, length of ventilator use, length of ECMO treatment, and number of days from ECMO cannulation to infection were identified. Respiratory or blood samples with pathogens commonly considered contaminants were excluded from this study.

All patients received positive pressure ventilation while on ECMO except for one patient who sustained a bronchial injury and was extubated after repair while on ECMO. There is a practice of early extubation in the non-trauma ECMO population that is not feasible in our cohort because of the polytrauma state and the fact that intubation in trauma patients is not solely a respiratory indication. In our institution, ventilated patients are automatically ordered a “VAP prevention bundle,” which includes elevation of the head of the bed and daily oral care. The institutional protocol regarding ECMO cannulation site care includes the sites being dressed with central line dressings that are changed every seven days or when loose or soiled. There was no institutional protocol for culture acquisition during the study period. It is common practice to evaluate patients with broad cultures when there is leukocytosis or significant clinical change. Cultures are drawn with low thresholds given the difficulty in evaluating temperature curves on ECMO.

Statistical analysis was performed using IBM SPSS Statistics version 28 (International Business Machines Corp, Armonk, New York). Data were reported as mean ± standard deviation (SD) if normally distributed or median (interquartile range [IQR]) if non-parametric. Categorical variables were compared with chi-squared test. Population distribution was assessed using the standard error of skewness and kurtosis, with normal distribution occurring when both values fell between −2 and 2. Statistical significance was defined at p < 0.05.

Results

Patient data

The study population comprised 50 trauma patients. Most were male (88%) who sustained a blunt injury (60%) with an average age of 31 ± 12.8 (range 13–66) years (Table 1). The median Injury Severity Score (ISS) was 29 [IQR:18–38] and median Glasgow Coma Scale (GCS) was 12 [IQR: 6–15]. The most common indication for ECMO was ARDS/respiratory failure (86%) and the most common ECMO type was veno-venous (VV 80%), although eight patients (16%) underwent revisions or were re-cannulated later in their hospital course.

Table 1.

Characteristics of Extracorporeal Membrane Oxygenation Patients (n = 50)

Age, mean ± SD	31 ± 12.8
Sex (male)	44 (88%)
Race
White	11 (22%)
Black	23 (46%)
Hispanic	16 (32%)
Median GCS [IQR]	12 [6–15]
Median ISS [IQR]	29 [18–38]
ICU length of stay, d, median [IQR]	26 [12–54]
Total vent days, median [IQR]	22 [9–51]
Hospital length of stay, d, median [IQR]	30 [13–68]
ECMO days, median [IQR]	8 [4–18]
Range of highest temperature in all patients 24 h before culture collection	35.6°C–38.3°C
Mechanism of injury
Blunt	30 (60%)
Penetrating	20 (40%)
Injury category
ATV	2 (4%)
Bike hit by car	1 (2%)
Crush injury	2 (4%)
Fall	3 (6%)
Jet ski accident	1 (2%)
MCC	7 (14%)
MVC	11 (22%)
PHBC	3 (6%)
Stab wound	4 (8%)
GSW	16 (32%)
Type of ECMO^a
VV	40 (80%)
VA	10 (20%)
ECMO indications
Respiratory failure	43 (86%)
Cardiac failure	7 (14%)
ECMO revisions/re-cannulations	8 (16%)

Values are reported as n (%) unless otherwise specified.

First extracorporeal membrane oxygenation (ECMO) type during patient’s hospital stay.

ATV= all-terrain vehicle, MCC= motorcycle collision, MVC= motor vehicle collision, SD = standard deviation, IQR = interquartile range, ICU = intensive care unit, VV = veno-venous, VA = veno-arterial.

Outcomes

The median ICU length of stay was 26 days (IQR:12–54), median ventilator days was 22 (IQR: 9–51), and median length of ECMO treatment was eight days (IQR: 4–18). The median number of ventilator days before ECMO cannulation was three days, ranging from 0 to 40 days. Out of 50 patients, 18 (36%) were receiving antibiotic agents before ECMO cannulation. To identify infections during ECMO, cultures with pathogens that were isolated for the first time after ECMO cannulation or were found to be multi-drug resistant or recurred after ECMO cannulation were included. Blood stream and/or respiratory infections were detected in 22 (44%) patients (Table 2). Concomitant respiratory and blood stream infections with the same pathogen were detected in five (22.3%) patients. The highest recorded temperature in the 24 hours before culture collection among all patients with infection ranged from 35.6°C to 38.3°C. The median number of days from ECMO cannulation to blood stream infection was 10 days (IQR: 4–16) and the median number of days to respiratory infection was seven days (IQR: 3–11). Mortality was 22.7% (n = 5) among patients who had a respiratory and/or blood stream infection and 60% (n = 17) in ECMO patients who had no infection (p = 0.007). This result has high frailty given the small cohort size and number of deaths. It is possible that the lower mortality in the non-infection group was because of patients dying before developing an infection.

Table 2.

Pathogens in Blood Stream and Respiratory Samples^a

Blood stream pathogens (n = 15)
Gram-negative pathogens	No. of pathogens (n = 8)
Acinetobacter baumannii	1 (12.5%)
Escherichia coli	1 (12.5%)
Klebsiella pneumoniae	1 (12.5%)
Pseudomonas aeruginosa	4 (50%)
Pseudomonas aeruginosa MDR	1 (12.5%)

Gram-positive pathogens	No. of pathogens (n = 5)
Enterococcus faecalis	3 (60%)
Methicillin-resistant Staphylococcus aureus MDR	1 (20%)
Viridans streptococci MDR	1 (20%)

Fungus	No. of pathogens (n = 2)
Candida parapsilosis	2 (100%)
Respiratory pathogens

Gram-negative pathogens	No. of pathogens (n = 27)
Achromobacter denitrificans	1 (3.7%)
Achromobacter xylosoxidans MDR	1 (3.7%)
Enterobacter aerogenes MDR	2 (7.4%)
Enterobacter cloacae	4 (14.8%)
Escherichia coli	1 (3.7%)
Haemophilus influenzae	1 (3.7%)
Klebsiella pneumoniae	1 (3.7%)
Pseudomonas aeruginosa	4 (14.8%)
Pseudomonas aeruginosa MDR	5 (18.5%)
Pseudomonas putida	1 (3.7%)
Sphingomonas paucimobilis MDR	1 (3.7%)
Stenotrophomonas maltophilia	3 (11.1%)
Stenotrophomonas maltophilia MDR	2 (7.4%)

Gram-positive pathogens	No. of pathogens (n = 9)
Enterococcus faecalis	1 (11.1%)
Enterococcus faecium MDR	1 (11.1%)
MRSA MDR	1 (11.1%)
Staphylococcus aureus	3 (33.3%)
Staphylococcus epidermidis MDR	2 (22.2%)
Streptococcus group C	1 (11.1%)

Pathogens considered contaminants: Staphylococcus epidermidis found in blood cultures, Staphylococcus coagulase negative, Candida found in respiratory cultures, Cupriavidus pauculus, Prevotella denticola, Bacillus species, and Lactobacillus casei. One patient was found to have no positive cultures during his hospital admission except for one respiratory culture that found light growth of Proteus mirabilis, Enterobacter aerogenes, and Candida albicans. This culture was considered colonization.

Most infections were respiratory (N = 33, 70.2%). The incidence rate of respiratory infections was 43.9/1000 ECMO days. Among these positive respiratory cultures, there were 36 causative pathogens, of which 27 (75%) were gram negative and 15 (41.7%) were multi-drug resistant.

The most common pathogen was multi-drug-resistant Pseudomonas aeruginosa (N = 5, 13.9%).

There were 14 blood stream infections and 15 pathogens that were isolated from positive cultures. The incidence rate of blood stream infections in this study was 18.6/1000 ECMO days. Among the 15 pathogens, eight (53.3%) were gram negative and five were gram positive (33.3%). The most common pathogen was P. aeruginosa (n = 4, 26.7%). There were three (20%) multi-drug resistant pathogens.

The combined incidence of respiratory and blood stream infections was 62.5/1000 ECMO days.

The most common location for ECMO cannulation and/or revision in patients with blood stream infections was at the ICU bedside (n = 7, 77.8%), and of these, five (71.4%) were bifemoral cannulations. Among the eight patients who had blood stream infection, two patients had ECMO revisions; one patient had a veno-arterial ECMO cannulation that was revised the following day for a cold pulseless leg and one patient initially had VV ECMO cannulation that was later revised to dual drainage with two devnous cannulas.

Discussion

This is the largest study to characterize infection risk in trauma patients receiving ECMO. Since the use of ECMO in trauma is limited, trauma patients are not well represented in the current literature on infection risk in ECMO, and therefore, this study addresses this gap in knowledge.^11–18,20

This study identified an infection in almost half of trauma patients receiving ECMO, which is comparable with what has been reported on the infection rate of trauma patients in the ICU.²¹ This suggests that ECMO does not necessarily pose an added risk of infection in trauma patients. However, the incidence rate of infection in this sample (62.5/1000 ECMO days) is much higher than the incidence rate of infection reported in all ECMO patients with the ELSO registry (15.4/1000 ECMO days), implying that higher infection rates are attributable to the trauma population.¹⁵ Treatment of trauma patients often involves multiple invasive and surgical procedures in addition to often needing mechanical ventilation, which are all risk factors for infection.²¹ Trauma patients comprise around 1% of ECMO patients and may be relatively unique to some ECMO providers.²² This underscores the necessity for ECMO providers to acknowledge the high infection risk in trauma patients receiving ECMO support.

For trauma providers who have limited experience with ECMO, it is important to understand that given the influence of ECMO circuitry on body cooling and inflammatory responses, fevers and other signs of infection are difficult to interpret in ECMO patients.^23,24 Therefore, the threshold to obtain an infectious workup in this cohort should be low and clinical changes or lack of clinical progression should prompt infectious workup.

The lungs were the most common site of infection in this sample. Schweickhardt et al. examined infections in a smaller cohort of trauma patients receiving ECMO and found that most infections in their patient sample were also respiratory.¹² Respiratory infections are also the most common infection in the general population of ECMO patients.^16,17 However, the incidence rate of respiratory infection found in this cohort of trauma patients on ECMO (43.9/1000 ECMO days) is slightly higher than what has been reported in another study with 92 non-specific ECMO patients.¹⁸ Ventilator-associated pneumonia is difficult to diagnose in ECMO patients given the poor reliability of radiographical images or interpretation of temperature curves.²³ Clinical providers may therefore need to rely more on positive respiratory cultures in the setting of clinical change to diagnose pneumonia and initiate treatment in ECMO trauma patients.

Gram-negative bacteria, specifically P. aeruginosa, were the most common type of pathogens found in this present study, in both blood and respiratory cultures. This finding aligns with the results of several studies that have also identified the majority of infections to be gram negative.^12,16 However, high rates of Candidal and Staphylococcal infection also occur in ECMO patients.²⁵ P. aeruginosa is a leading cause of infection in the ICU and is often found in patients who are on ventilator support or have catheters.^26,27 It is commonly spread from person-to-person contact or contact with contaminated surfaces. As evidenced by the results in this study, ECMO trauma patients are high risk for colonization with P. aeruginosa . Thus, maintenance of strict infection control practices is paramount.

The present study identified a high rate of multi-drug-resistant bacteria (35.3%), which corresponds with previous reports in ECMO patients.¹⁸ ECMO patients are a high-risk group for colonization with multi-drug-resistant bacteria given long ICU stays and ventilator use, often needing broad antibiotic agent exposure.²⁸

This study is limited by the fact that it is a retrospective review at a single center and has a small sample size. Because ECMO is not commonly used, it is difficult to find a large sample size of ECMO trauma patients at a single institution. The analysis is also limited by missing data in charts and varying practice patterns by alternating providers. This cohort size does not allow for yearly comparison. However, there have been no notable changes in institutional practice regarding ECMO care for this study period. The longitudinal impact of progressive improvement in ICU care has the potential to improve infectious complications and is another topic to be explored with a larger cohort. Bacteremia causation is unable to be determined in this study. This should be explored in another study using genomics to identify the exact causes of the species of bacteria found in blood. The retrospective nature of the study does not allow for clarity on which antibiotic agents were chosen to cover which infection. Given this limitation of missing or unclear information in medical records on antibiotic agent use, this variable was excluded. Lastly, respiratory infection was defined by positive cultures given the difficulty in diagnosing pneumonia in ECMO patients from radiological evidence or interpretation of temperature curves.²³

Conclusion

In the largest study to characterize infection risk in trauma patients receiving ECMO, infection was identified in almost half of the patient sample, with respiratory being the most common. Most pathogens causing infection were gram negative in both blood and respiratory samples, specifically P. aeruginosa. This finding underscores the importance of practicing strict infection control when treating ECMO trauma patients.

Footnotes

Authors’ Contributions

L.S. contributed to the study conception and design, acquisition of data, analysis and interpretation of data, drafting of article and critical revision of article. L.T.B. contributed to the acquisition of data, analysis and interpretation of data, and critical revision of the article. J.M.D. contributed to the critical revision of the article. M.D.C.-L. contributed to critical revision of the article. N.B.L. contributed to critical revision of the article. B.L.C. contributed to critical revision of the article. O.P. contributed to critical revision of the article. M.L. contributed to acquisition of data. E.B.L. contributed to critical revision of the article. V.H. contributed to critical revision of the article. J.P.M. contributed to critical revision of the article. T.M. contributed to critical revision of the article. N.N. contributed to critical revision of the article. K.G.P. contributed to critical revision of the article. B.M.P. contributed to analysis and interpretation of data, drafting of article, and critical revision of the article.

Author Disclosure Statement

The authors have no other conflicts of interest.

Funding Information

This study was supported by internal funds from the Daughtry Family Department of Surgery.

References

Wisqars leading causes of Death Visualization Tool. Centers for Disease Control and Prevention. Available from: https://wisqars.cdc.gov/lcd/?o=LCD&y1=2022&y2=2022&ct=10&cc=ALL&g=00&s=0&r=0&ry=0&e=0&ar=lcd1age&at=groups&ag=lcd1age&a1=0&a2=199

Callcut

, Kornblith

, Conroy

, et al.; Western Trauma Association Multicenter Study Group. The why and how our trauma patients die: A prospective Multicenter Western Trauma Association study. J Trauma Acute Care Surg, 2019; 86(5):864–870.

Gray

, Haft

, Hirsch

, et al. Extracorporeal life support: Experience with 2,000 patients. Asaio J, 2015; 61(1):2–7.

Bedeir

, Seethala

, Kelly

. Extracorporeal life support in trauma: Worth the risks? A systematic review of published series. J Trauma Acute Care Surg, 2017; 82(2):400–406.

Grant

, Hart

, Lineen

, et al. The impact of an advanced ECMO program on traumatically injured patients. Artif Organs, 2018; 42(11):1043–1051.

Muellenbach

, Kredel

, Kunze

, et al. Prolonged heparin-free extracorporeal membrane oxygenation in multiple injured acute respiratory distress syndrome patients with traumatic brain injury. J Trauma Acute Care Surg, 2012; 72(5):1444–1447.

Arlt

, Philipp

, Voelkel

, et al. Extracorporeal membrane oxygenation in severe trauma patients with bleeding shock. Resuscitation, 2010; 81(7):804–809.

Guttman

, Tillmann

, Pannell

, et al. Extracorporeal membrane oxygenation use in Trauma Quality Improvement Program centers: Temporal trends and future directions. J Trauma Acute Care Surg, 2020; 89(2):351–357.

Swol

, Brodie

, Napolitano

, et al.; Extracorporeal Life Support Organization (ELSO). Indications and outcomes of extracorporeal life support in trauma patients. J Trauma Acute Care Surg, 2018; 84(6):831–837.

10.

Zhang

, Zhang

, Huang

, et al. ECMO in adult patients with severe trauma: A systematic review and meta-analysis. Eur J Med Res, 2023; 28(1):412.

11.

Ait Hssain

, Vahedian-Azimi

, Ibrahim

, et al. Incidence, risk factors and outcomes of nosocomial infection in adult patients supported by extracorporeal membrane oxygenation: A systematic review and meta-analysis. Crit Care, 2024; 28(1):158.

12.

Schweickhardt

, Sams

, Radowsky

, et al. Infections complicating extracorporeal membrane oxygenation in patients with traumatic injuries. Injury, 2023; 54(2):405–408.

13.

Quintana

, Mazzeffi

, Galvagno

, et al. A retrospective study of infection in patients requiring extracorporeal membrane oxygenation support. Ann Thorac Surg, 2021; 112(4):1168–1175.

14.

Menaker

, Galvagno

, Rabinowitz

, et al. Epidemiology of blood stream infection in adult extracorporeal membrane oxygenation patients: A Cohort Study. Heart Lung, 2019; 48(3):236–239.

15.

Bizzarro

, Conrad

, Kaufman

, et al.; Extracorporeal 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(3):277–281.

16.

Juthani

, Macfarlan

, Wu

, et al. Incidence of nosocomial infections in adult patients undergoing extracorporeal membrane oxygenation. Heart Lung, 2018; 47(6):626–630.

17.

Wang

, Ni

, Wang

, et al. Nosocomial infections in adult patients receiving extracorporeal membrane oxygenation in China: A retrospective cohort study. Am J Infect Control, 2023; 51(11):1237–1241.

18.

Grasselli

, Scaravilli

, Di Bella

, et al. Nosocomial infections during extracorporeal membrane oxygenation: Incidence, etiology, and impact on patients’ outcome. Crit Care Med, 2017; 45(10):1726–1733.

19.

Magiorakos

, Srinivasan

, Carey

, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: An international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect, 2012; 18(3):268–281.

20.

Raff

, Kerby

, Reiff

, et al. Use of extracorporeal membranous oxygenation in the management of refractory trauma-related severe acute respiratory distress syndrome: A national survey of the Eastern Association for the Surgery of Trauma. Trauma Surg Acute Care Open, 2019; 4(1):e000341.

21.

Giamberardino

, Cesário

, Carmes

, et al. Risk factors for nosocomial infection in trauma patients. Braz J Infect Dis, 2007; 11(2):285–289.

22.

, Griswold

, Raff

, et al. National estimates of the use and outcomes of extracorporeal membrane oxygenation after acute trauma. Trauma Surg Acute Care Open, 2019; 4(1):e000209.

23.

Franchineau

, Luyt

, Combes

, et al. Ventilator-associated pneumonia in extracorporeal membrane oxygenation-assisted patients. Ann Transl Med, 2018; 6(21):427.

24.

Jackson

, Sobieszczyk

, Aden

, et al. Fever and leukocytosis are poor predictors of bacterial coinfection in patients with COVID-19 and influenza who are receiving extracorporeal membrane oxygenation. Open Forum Infect Dis, 2023; 10(11):ofad501.

25.

Pieri

, Agracheva

, Fumagalli

, et al. Infections occurring in adult patients receiving mechanical circulatory support: The two-year experience of an Italian National Referral Tertiary Care Center. Med Intensiva, 2013; 37(7):468–475.

26.

Blanc

, Francioli

, Zanetti

. Molecular Epidemiology of Pseudomonas aeruginosa in the Intensive Care Units—A Review. Open Microbiol J, 2007; 1:8–11.

27.

About Pseudomonas aeruginosa. Centers for Disease Control and Prevention. Available from: https://www.cdc.gov/pseudomonas-aeruginosa/about/index.html

28.

Trouillet

, Chastre

, Vuagnat

, et al. Ventilator-associated pneumonia caused by potentially drug-resistant bacteria. Am J Respir Crit Care Med, 1998; 157(2):531–539.