SYSTEMATIC REVIEW
Extracorporeal life support in trauma: Worth the risks? A systematic review of published series Kareem Bedeir, MBChB, MS, Raghu Seethala, MD, MSc, and Edward Kelly, MD, Boston, Massachusetts
Extracorporeal life support (ECLS) is a potentially life-saving procedure for trauma patients with severe respiratory failure. Despite this, only a limited number of publications report series of trauma patients who underwent ECLS. The performance and safety of this technology in trauma patients is not fully understood. We described the efficacy and complications of ECLS in trauma patients with respiratory failure. METHODS: A systematic review of published reports was performed utilizing the Medical Literature Analysis and Retrieval System Online (MEDLINE). Studies reporting ECLS in five trauma patients or more were examined for eligibility. Eligible trials were examined for patient characteristics, trauma characteristics, and anticoagulation management. Outcomes were examined for survival, causes of overall mortality, and the incidence of bleeding-related mortality. RESULTS: Six hundred five studies were identified in the initial literature search. Of these, 12 studies met the inclusion and exclusion criteria with a total of 215 patients. The overall survival to discharge ranged from 50% to 79%. Survival to discharge after venovenousECLS and venoarterial-ECLS were different, ranging from 56% to 89% and 42% to 63%, respectively. The leading cause of mortality was sepsis, which was the cause in 48% of all deaths. Bleeding-related mortality ranged between 0% and 38%, and was consistently lower than 15% in studies after 1995. Reported patients with traumatic intracranial hemorrhage (ICH) had survival ranging from 60% to 93%, with no mortalities related to their ICH. There was a recent trend towards initially heparin-free circuitry followed by a lower-than-standard activated clotting time target range. CONCLUSION: The use of ECLS in trauma patients when needed may provide survival benefits that significantly overweigh the feared risk of bleeding associated. (J Trauma Acute Care Surg. 2017;82: 400–406. Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Systematic review, level III. KEY WORDS: Extracorporeal support; trauma; bleeding; mortality. BACKGROUND:
C
hest injury is present in almost half of all multiple trauma cases.1 In trauma patients developing multisystem organ failure (MSOF), respiratory failure is usually the first to occur, the most common and the organ failure with the highest mortality.2,3 Acute respiratory distress syndrome (ARDS) occurs in 5% to 10% of trauma admissions and in up to 20% of patients developing shock or requiring emergency surgery.2,4,5 In addition, ARDS occurs in 20% to 30% of patients with traumatic brain injuries (TBI).6,7 The overall hospital mortality in patients with ARDS in the era of lung protective ventilation strategies is reported to be around 40%.8,9 Evidence suggests that trauma-related ARDS has distinct characteristics that result in a different outcome profile than that with other causes of ARDS. Reports seem to suggest that trauma-related acute respiratory distress syndrome (ARDS) carries a more favorable prognosis compared to other causes of ARDS.5,10 Unfortunately, however, the analysis of three multicenter randomized trials from the ARDS Network (n = 2,451 mechanically ventilated patients) between 1996 and 2005 showed a significant improvement in ARDS-related mortality, except in trauma-related ARDS.5
Submitted: April 3, 2016, Revised: August 25, 2016, Accepted: August 30, 2016, Published online: October 25, 2016. From the Brigham and Women’s Hospital, Boston, Massachusetts. Address for reprints: Edward Kelly, MD, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115; email: [email protected]. DOI: 10.1097/TA.0000000000001292
Despite this reversible nature of the syndrome, still 10% to 20% of trauma-related ARDS patients die despite our best ventilatory management. Extracorporeal life support (ECLS) has been proposed as a viable option to bypass the acute phase as a bridge to lung recovery, allowing for oxygenation while maintaining a lung protective ventilatory strategy. Extracorporeal circulatory support is known to cause a combination of platelet destruction, dysfunction, and activation, predisposing to both clotting and bleeding. The former mandates anticoagulation (AC), which accentuates the risk of bleeding. This has led to the traditional consideration of intracranial bleeding as an absolute contraindication to ECLS. Other contraindications for AC are also considered as contraindications for ECLS in most instances. For obvious reasons, these concerns are more pronounced in trauma patients. Several reports of trauma patients undergoing ECLS have been published.11–13 Because of the small number mostly being case reports, meaningful conclusions were difficult to draw. We conducted a systematic review to examine the outcomes of ECLS in trauma patients with respiratory failure and determine the rates of survival and adverse events with ECLS in this particular population.
PATIENTS AND METHODS Search Strategy and Selection Criteria Published reports of trauma patients requiring ECLS were reviewed utilizing the Medical Literature Analysis and Retrieval J Trauma Acute Care Surg Volume 82, Number 2
400
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 82, Number 2
Bedeir et al.
Figure 1. Searching, screening, and inclusion process of published reports.
System Online (MEDLINE). A protocol for the search, selection, and data extraction processes was followed. The search was conducted using the combination of keywords “trauma” and “injury” plus “extracorporeal support” and “cardiopulmonary support.” All published abstracts in English language were screened for relevance to ECLS in trauma patients (last searched May 2016). Full texts of screening results were examined, and study reports were eligible if they included five or more trauma patients receiving ECLS. All types of studies were included. Reports were excluded if they were (1) case reports or series with less than five patients reported, (2) reports of trauma patients undergoing cardiopulmonary bypass in the operating room for cardiovascular or pulmonary injury repair (only reports of trauma patients receiving ECLS in the intensive care unit were eligible), (3) reports of ECLS caused by both traumatic and nontraumatic causes where characteristics and outcomes of the trauma population were not specifically reported, (4) reports of ECLS after cardiac arrest as a bridge to organ donation, and (5) reports with the same patients were reported again in a subsequent publication by the authors. Reports with less than five patients were excluded because they were likely to originate from centers of low experience. In the cases when they originated from centers with larger experience, the same patients were reported again in a subsequent larger series published.
Data Extraction, Risk of Bias, and Data Synthesis A predefined data extraction form was developed a priori and key data were collected by reviewers independently. The risk of bias was assessed at the study level using Cochrane’s Collaboration Tool. Through six domains, this tool evaluates the risk of bias and categorizes each study as high risk, low risk, or unclear risk. The following descriptive data were collected: year of publication, country study was conducted in, number of patients, duration of enrollment, patient age, mechanism of injury, and Injury Severity Scores (ISS). The following patient management characteristics were gathered: anticoagulation management (initiation,
monitoring, target ranges, and perioperative management), time from trauma to ECLS initiation, venous versus arterial inflow access, the use of heparin-bonded circuitry, and surgeries performed while on ECLS. Outcome data collected included time on ECLS, overall survival to discharge, and causes of mortality specifically bleeding.
RESULTS Twelve papers published from 1994 to 2015 fit our inclusion criteria and were eligible for examination after screening 605 search results (Fig. 1). Five hundred fifty-one reports were excluded because of nonrelevance to ECLS in trauma. Full texts of the remaining 54 reports were examined for eligibility. Thirty reports included less than five patients. Four reports reported more than five trauma patients, but reported characteristics and outcomes did not specify those for the trauma subset. Four reports described cardiopulmonary bypass in the operating room for operative repair of the heart, lungs, or major vessels. Three results included patients that were later reported by the authors in a publication that was eligible for inclusion, and one report included the majority of patients undergoing ECLS after cardiac arrest as a bridge to organ donation. The 12 included reports originated from 10 centers in the United States, Germany, Taiwan, the United Kingdom, and Israel, in descending order of sample sizes (Tables 1 and 2). After excluding patients reported more than once, a total of 215 patients were reported (Table 3). All included studies were case series with no control groups. Using Cochrane’s Collaboration Tool, all included studies were assessed to have a high risk of bias. The variables most consistently reported and were included in pooled analysis included baseline variables: age, mechanism of trauma, and ISS. Consistently reported variables that are related to ECLS included inflow target (VAvs. VV), duration from trauma, and duration while on ECLS (Table 3).
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
401
402
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
University of Case series, Regensburg, Germany high
University of Case series, Regensburg, Germany high
Arlt et al.,20 2010
Ried et al.,21 2013†
Biderman et al.,22 University of Tel 2013 Aviv, Israel
Chang-Gung Hospital, Taiwan
Huang et al.,19 2009
Case series, high
Case series, high
10/5
52/10
10/3
9/3
28/8
Glenfield Hospital, UK
Cordell-Smith et al.,18 2006 Case series, high
30/9
Michaels University of Michigan Case series, high et al.,17 1999* and Oregon Emanuel Health Center, USA
University of Michigan, USA
Senunas et al.,16 1997
6/3
14/6
Case series, high
Oregon Emanuel Health Center, USA
Perchinsky et al.,15 1995
24/5
n/ years
Case series, high
Case series, high
Study Type, Risk of Bias
University of Michigan, USA
Center, Country
Anderson et al.,14 1994
Authors, year
29 (19–42)
32 (16–72)
32 (17–62)
35 (18–47)
27 (18–48)
26 (15–59)
19 (5–47)(14% peds)
— (5 Patients: Characteristics and Outcomes Causes of Death (n/% of all subjects) – In 12 patients, trauma to ECLS 80% – Trauma to ECLS 6 d: 17% survival – Half the patients had some sort of bleeding 15 (50) — – Trauma to ECLS ≤5 d had a survival OR of 7.2 (p = 0.006) 3 (50)
15 (63)Peds: 7 (58) – Major bleeding (9/38%) Adult: 8 (67)
Survival to Discharge n, %
Bedeir et al. J Trauma Acute Care Surg Volume 82, Number 2
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
ISS, Injury Severity Score; Peds, pediatric; IVC, inferior vena cava; MSOF, multisystem organ failure; ICH, intracranial hemorrhage; ECLS, extracorporeal life support; OR, odds ratio; NE ggt, norepinephrine infusion; PRBCs, packed red blood cells; TBI, traumatic brain injury; CVVH, continuous venovenous hemodiafiltration. *Includes 12 adult patients from Anderson et al. (1994). † Patient overlap between Huang et al. (2009) and We et al. (2014), between Senunas et al. (1997) and Anderson et al. (1994), and between Arlt et al. (2010) and Ried et al. (2013). ‡Half the patients on pumpless extracorporeal life support (PECLA).
13 (68) 29 19/5 China University Hospital, Taiwan
Wu et al.,24 2014‡
Wu et al.,25 2015
Case series, high
41 (25–58)
19 (100)
14 (70) 35 20 (100) 20/9 Case series, high
38 (20–72)
26/8
Wake Forest Hospital, LA County Hospital, and University of Southern California, USA Chang-Gung Hospital, Taiwan Guirand et al.,23 2014
Case series, high
33
21 (81)
15 (58)
– Major bleeding (3/15%)
– 35% suffered major bleeding, of which 50% died – Sepsis/MSOF (3/15%) – Trauma to ECLS
Extracorporeal life support in trauma: Worth the risks? A systematic review of published series Kareem Bedeir, MBChB, MS, Raghu Seethala, MD, MSc, and Edward Kelly, MD, Boston, Massachusetts
Extracorporeal life support (ECLS) is a potentially life-saving procedure for trauma patients with severe respiratory failure. Despite this, only a limited number of publications report series of trauma patients who underwent ECLS. The performance and safety of this technology in trauma patients is not fully understood. We described the efficacy and complications of ECLS in trauma patients with respiratory failure. METHODS: A systematic review of published reports was performed utilizing the Medical Literature Analysis and Retrieval System Online (MEDLINE). Studies reporting ECLS in five trauma patients or more were examined for eligibility. Eligible trials were examined for patient characteristics, trauma characteristics, and anticoagulation management. Outcomes were examined for survival, causes of overall mortality, and the incidence of bleeding-related mortality. RESULTS: Six hundred five studies were identified in the initial literature search. Of these, 12 studies met the inclusion and exclusion criteria with a total of 215 patients. The overall survival to discharge ranged from 50% to 79%. Survival to discharge after venovenousECLS and venoarterial-ECLS were different, ranging from 56% to 89% and 42% to 63%, respectively. The leading cause of mortality was sepsis, which was the cause in 48% of all deaths. Bleeding-related mortality ranged between 0% and 38%, and was consistently lower than 15% in studies after 1995. Reported patients with traumatic intracranial hemorrhage (ICH) had survival ranging from 60% to 93%, with no mortalities related to their ICH. There was a recent trend towards initially heparin-free circuitry followed by a lower-than-standard activated clotting time target range. CONCLUSION: The use of ECLS in trauma patients when needed may provide survival benefits that significantly overweigh the feared risk of bleeding associated. (J Trauma Acute Care Surg. 2017;82: 400–406. Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Systematic review, level III. KEY WORDS: Extracorporeal support; trauma; bleeding; mortality. BACKGROUND:
C
hest injury is present in almost half of all multiple trauma cases.1 In trauma patients developing multisystem organ failure (MSOF), respiratory failure is usually the first to occur, the most common and the organ failure with the highest mortality.2,3 Acute respiratory distress syndrome (ARDS) occurs in 5% to 10% of trauma admissions and in up to 20% of patients developing shock or requiring emergency surgery.2,4,5 In addition, ARDS occurs in 20% to 30% of patients with traumatic brain injuries (TBI).6,7 The overall hospital mortality in patients with ARDS in the era of lung protective ventilation strategies is reported to be around 40%.8,9 Evidence suggests that trauma-related ARDS has distinct characteristics that result in a different outcome profile than that with other causes of ARDS. Reports seem to suggest that trauma-related acute respiratory distress syndrome (ARDS) carries a more favorable prognosis compared to other causes of ARDS.5,10 Unfortunately, however, the analysis of three multicenter randomized trials from the ARDS Network (n = 2,451 mechanically ventilated patients) between 1996 and 2005 showed a significant improvement in ARDS-related mortality, except in trauma-related ARDS.5
Submitted: April 3, 2016, Revised: August 25, 2016, Accepted: August 30, 2016, Published online: October 25, 2016. From the Brigham and Women’s Hospital, Boston, Massachusetts. Address for reprints: Edward Kelly, MD, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115; email: [email protected]. DOI: 10.1097/TA.0000000000001292
Despite this reversible nature of the syndrome, still 10% to 20% of trauma-related ARDS patients die despite our best ventilatory management. Extracorporeal life support (ECLS) has been proposed as a viable option to bypass the acute phase as a bridge to lung recovery, allowing for oxygenation while maintaining a lung protective ventilatory strategy. Extracorporeal circulatory support is known to cause a combination of platelet destruction, dysfunction, and activation, predisposing to both clotting and bleeding. The former mandates anticoagulation (AC), which accentuates the risk of bleeding. This has led to the traditional consideration of intracranial bleeding as an absolute contraindication to ECLS. Other contraindications for AC are also considered as contraindications for ECLS in most instances. For obvious reasons, these concerns are more pronounced in trauma patients. Several reports of trauma patients undergoing ECLS have been published.11–13 Because of the small number mostly being case reports, meaningful conclusions were difficult to draw. We conducted a systematic review to examine the outcomes of ECLS in trauma patients with respiratory failure and determine the rates of survival and adverse events with ECLS in this particular population.
PATIENTS AND METHODS Search Strategy and Selection Criteria Published reports of trauma patients requiring ECLS were reviewed utilizing the Medical Literature Analysis and Retrieval J Trauma Acute Care Surg Volume 82, Number 2
400
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 82, Number 2
Bedeir et al.
Figure 1. Searching, screening, and inclusion process of published reports.
System Online (MEDLINE). A protocol for the search, selection, and data extraction processes was followed. The search was conducted using the combination of keywords “trauma” and “injury” plus “extracorporeal support” and “cardiopulmonary support.” All published abstracts in English language were screened for relevance to ECLS in trauma patients (last searched May 2016). Full texts of screening results were examined, and study reports were eligible if they included five or more trauma patients receiving ECLS. All types of studies were included. Reports were excluded if they were (1) case reports or series with less than five patients reported, (2) reports of trauma patients undergoing cardiopulmonary bypass in the operating room for cardiovascular or pulmonary injury repair (only reports of trauma patients receiving ECLS in the intensive care unit were eligible), (3) reports of ECLS caused by both traumatic and nontraumatic causes where characteristics and outcomes of the trauma population were not specifically reported, (4) reports of ECLS after cardiac arrest as a bridge to organ donation, and (5) reports with the same patients were reported again in a subsequent publication by the authors. Reports with less than five patients were excluded because they were likely to originate from centers of low experience. In the cases when they originated from centers with larger experience, the same patients were reported again in a subsequent larger series published.
Data Extraction, Risk of Bias, and Data Synthesis A predefined data extraction form was developed a priori and key data were collected by reviewers independently. The risk of bias was assessed at the study level using Cochrane’s Collaboration Tool. Through six domains, this tool evaluates the risk of bias and categorizes each study as high risk, low risk, or unclear risk. The following descriptive data were collected: year of publication, country study was conducted in, number of patients, duration of enrollment, patient age, mechanism of injury, and Injury Severity Scores (ISS). The following patient management characteristics were gathered: anticoagulation management (initiation,
monitoring, target ranges, and perioperative management), time from trauma to ECLS initiation, venous versus arterial inflow access, the use of heparin-bonded circuitry, and surgeries performed while on ECLS. Outcome data collected included time on ECLS, overall survival to discharge, and causes of mortality specifically bleeding.
RESULTS Twelve papers published from 1994 to 2015 fit our inclusion criteria and were eligible for examination after screening 605 search results (Fig. 1). Five hundred fifty-one reports were excluded because of nonrelevance to ECLS in trauma. Full texts of the remaining 54 reports were examined for eligibility. Thirty reports included less than five patients. Four reports reported more than five trauma patients, but reported characteristics and outcomes did not specify those for the trauma subset. Four reports described cardiopulmonary bypass in the operating room for operative repair of the heart, lungs, or major vessels. Three results included patients that were later reported by the authors in a publication that was eligible for inclusion, and one report included the majority of patients undergoing ECLS after cardiac arrest as a bridge to organ donation. The 12 included reports originated from 10 centers in the United States, Germany, Taiwan, the United Kingdom, and Israel, in descending order of sample sizes (Tables 1 and 2). After excluding patients reported more than once, a total of 215 patients were reported (Table 3). All included studies were case series with no control groups. Using Cochrane’s Collaboration Tool, all included studies were assessed to have a high risk of bias. The variables most consistently reported and were included in pooled analysis included baseline variables: age, mechanism of trauma, and ISS. Consistently reported variables that are related to ECLS included inflow target (VAvs. VV), duration from trauma, and duration while on ECLS (Table 3).
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
401
402
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
University of Case series, Regensburg, Germany high
University of Case series, Regensburg, Germany high
Arlt et al.,20 2010
Ried et al.,21 2013†
Biderman et al.,22 University of Tel 2013 Aviv, Israel
Chang-Gung Hospital, Taiwan
Huang et al.,19 2009
Case series, high
Case series, high
10/5
52/10
10/3
9/3
28/8
Glenfield Hospital, UK
Cordell-Smith et al.,18 2006 Case series, high
30/9
Michaels University of Michigan Case series, high et al.,17 1999* and Oregon Emanuel Health Center, USA
University of Michigan, USA
Senunas et al.,16 1997
6/3
14/6
Case series, high
Oregon Emanuel Health Center, USA
Perchinsky et al.,15 1995
24/5
n/ years
Case series, high
Case series, high
Study Type, Risk of Bias
University of Michigan, USA
Center, Country
Anderson et al.,14 1994
Authors, year
29 (19–42)
32 (16–72)
32 (17–62)
35 (18–47)
27 (18–48)
26 (15–59)
19 (5–47)(14% peds)
— (5 Patients: Characteristics and Outcomes Causes of Death (n/% of all subjects) – In 12 patients, trauma to ECLS 80% – Trauma to ECLS 6 d: 17% survival – Half the patients had some sort of bleeding 15 (50) — – Trauma to ECLS ≤5 d had a survival OR of 7.2 (p = 0.006) 3 (50)
15 (63)Peds: 7 (58) – Major bleeding (9/38%) Adult: 8 (67)
Survival to Discharge n, %
Bedeir et al. J Trauma Acute Care Surg Volume 82, Number 2
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
ISS, Injury Severity Score; Peds, pediatric; IVC, inferior vena cava; MSOF, multisystem organ failure; ICH, intracranial hemorrhage; ECLS, extracorporeal life support; OR, odds ratio; NE ggt, norepinephrine infusion; PRBCs, packed red blood cells; TBI, traumatic brain injury; CVVH, continuous venovenous hemodiafiltration. *Includes 12 adult patients from Anderson et al. (1994). † Patient overlap between Huang et al. (2009) and We et al. (2014), between Senunas et al. (1997) and Anderson et al. (1994), and between Arlt et al. (2010) and Ried et al. (2013). ‡Half the patients on pumpless extracorporeal life support (PECLA).
13 (68) 29 19/5 China University Hospital, Taiwan
Wu et al.,24 2014‡
Wu et al.,25 2015
Case series, high
41 (25–58)
19 (100)
14 (70) 35 20 (100) 20/9 Case series, high
38 (20–72)
26/8
Wake Forest Hospital, LA County Hospital, and University of Southern California, USA Chang-Gung Hospital, Taiwan Guirand et al.,23 2014
Case series, high
33
21 (81)
15 (58)
– Major bleeding (3/15%)
– 35% suffered major bleeding, of which 50% died – Sepsis/MSOF (3/15%) – Trauma to ECLS
