Assessing the need for resuscitative endovascular balloon occlusion of the aorta (REBOA) for management of non-compressible hemorrhage at a Canadian urban trauma centre

Resuscitative endovascular balloon occlusion of the aorta (REBOA) is a less invasive alternative to resuscitative thoracotomy (RT) for life threatening, infra-diaphragmatic, non-compressible hemorrhage from trauma. Existing evidence surrounding the ecacy of REBOA is conicting; nevertheless, expert consensus suggests that REBOA should be considered in select trauma patients. There has been a paucity of studies that evaluate the potential utility of REBOA in the Canadian setting. The study objective was to evaluate the percentage of trauma patients presenting to a Level 1 Canadian trauma centre that would have met criteria for REBOA. Methods We conducted a retrospective chart review of patients in the British Columbia Trauma Registry who warranted a trauma team activation (TTA) at our institution. We identied REBOA candidates using pre-dened criteria based on published guidelines. Each TTA case was screened by a reviewer, and then each Potential Candidate was reviewed by a panel of trauma physicians for determination of nal candidacy.


Introduction
Resuscitative endovascular balloon occlusion of the aorta (REBOA) is a less invasive alternative to resuscitative thoracotomy (RT) for life threatening, infra-diaphragmatic, non-compressible hemorrhage from trauma. Existing evidence surrounding the e cacy of REBOA is con icting; nevertheless, expert consensus suggests that REBOA should be considered in select trauma patients. There has been a paucity of studies that evaluate the potential utility of REBOA in the Canadian setting. The study objective was to evaluate the percentage of trauma patients presenting to a Level 1 Canadian trauma centre that would have met criteria for REBOA.

Methods
We conducted a retrospective chart review of patients recorded in the British Columbia Trauma Registry who warranted a trauma team activation (TTA) at our institution. We identi ed REBOA candidates using pre-de ned criteria based on published guidelines. Each TTA case was screened by a reviewer, and then each Potential Candidate was reviewed by a panel of trauma physicians for determination of nal candidacy.

Introduction
Massive exsanguination resulting in circulatory collapse is one of the leading causes of preventable death in trauma. [1][2][3][4][5][6][7] The cornerstone of the emergency management of massive hemorrhage is obtaining control of the bleeding. However, source control becomes a challenge when the source of major hemorrhage involves non-compressible regions such as the trunk and ilio-junctional regions. 8,9 Consequently, noncompressible hemorrhage represents the majority of traumatic exsanguination fatalities accounting for 45-60% of deaths. 10,11 Historically, the only means of attempting to control this type of bleeding was to perform a resuscitative thoracotomy (RT) as a bridge to either de nitive surgical management in the operating room (OR) or arterial embolization by interventional radiology (IR). 8,9 This highly invasive procedure is associated with high levels of mortality and complications. 12 Within the last 10 years, innovation has led to the development and popularization of a second option for the management of life threatening non-compressible hemorrhage. Resuscitative endovascular balloon occlusion of the aorta (REBOA) involves in ation of a balloon in the aorta as a means of reducing blood ow to the distal hemorrhage site as well as preferentially directing blood ow to critical organs including the heart and brain. 8,9, 13 -19 Although REBOA is less invasive than RT, the existing evidence surrounding the e cacy of REBOA is con icted; 16,20−27 nevertheless, expert consensus suggests that REBOA should be considered in select severely injured patients. 28,29 Many US centres that see high volumes of trauma and perform regular RTs have already adopted REBOA at their institutions. However, the pattern and volume of trauma is different in Canada than in the USA. [30][31][32] Although some Canadian institutions are currently performing REBOA, there has been a paucity of studies evaluating the potential utility of this tool in the Canadian context 33 , and it has yet to be adopted as standard of care in Canada. A recent editorial has highlighted the need for further research in the Canadian setting prior to widespread adoption of REBOA 34 . Our study is a Canadian-based trauma population-driven assessment for the need for acquisition of REBOA at a tertiary care, Level 1 trauma centre. The objective of this study was to evaluate the percentage of trauma patients at our Canadian institution that sustained injury resulting in life-threatening hemorrhage below the diaphragm where REBOA would have been indicated as a component of the emergency department (ED) resuscitation.

Study Design, Time Period, and Setting
This retrospective descriptive study was conducted at the Royal Columbian Hospital, which is a Level 1 Canadian trauma centre that sees approximately 1000 trauma consults annually. We used medical record data for trauma patients captured in the British Columbia Trauma Registry (BCTR) who presented to our ED between January 1, 2016 and December 31, 2018. The BCTR is a comprehensive and organized provincial trauma surveillance and data collection system that is regularly quality checked.

Population
All BCTR cases that were trauma team activations (TTAs) were screened for inclusion in this study. TTAs are called for patients when there is concern for severe injury based on a set of criteria established by the Health Authority (Appendix A).

Page 4/24
The inclusion and exclusion criteria for REBOA candidacy were chosen based on suggested REBOA protocols published in the literature [35][36][37][38]  College of Surgeons and American College of Emergency Physicians guidelines. 28,29 Cases must have had one of the following inclusion criteria to be considered a REBOA candidate: 1. Patients that sustained blunt trauma with a suspected infra-diaphragmatic injury that presented to the ED with a pulse and a systolic blood pressure less than 90mmHg that were transient or nonresponders to resuscitation 2. Patients that sustained penetrating trauma to the torso below the diaphragm that presented to the ED with a pulse and a systolic blood pressure less than 90mmHg that were transient or non-responders to resuscitation 3. Patients that sustained blunt trauma with a suspected infra-diaphragmatic injury that presented pulseless to the ED but had an organized rhythm on ECG or organized cardiac contraction seen on point-of-care ultrasound 4. Patients that sustained penetrating trauma to the torso below the diaphragm that presented pulseless to the ED and had less than 15 minutes of CPR.
Transient response to resuscitation was de ned as a blood pressure that increased above 90mmHg with two units of blood product resuscitation for less than 15 minutes. Non-response to resuscitation was de ned as a blood pressure that failed to increase above 90mmHg with two units of blood product resuscitation or patients that required inotropic support.
The infra-diaphragmatic torso included the abdomen, pelvis, and ilio-junctional areas. Suspected infradiaphragmatic injury in blunt trauma patients was de ned as a patient with a: 3. Patients that sustained penetrating trauma but had been pulseless for greater than 15 minutes 4. Trauma patients that also had evidence of a major thoracic injury including a cardiac injury, a supradiaphragmatic aortic injury, or other major thoracic vascular injury 5. Trauma patients in which their main source of hemorrhage was from an extremity that was controllable with tourniquet or direct pressure 6. Trauma patients that also sustained a devastating head injury 7. Trauma patients that also had a major neck injury.

Outcome Measures
The primary outcome was identi cation of REBOA candidacy according to inclusion and exclusion criteria. Additionally, patient characteristics, clinical variables, and details on the traumatic event were collected for each case. These included patient sex and age, comorbidities, mechanism of injury, injury severity score (ISS), and presenting vitals. Speci c management variables were also collected: blood products transfused within 1, 2, 4, and 24-hours after initiation of resuscitation; activation of massive transfusion protocol; tourniquet placement; pelvic binder placement; and use of resuscitative thoracotomy and aortic cross-clamping. The details collected contribute to the clinical decision to use REBOA during ED resuscitation and characterize the types of patients that are REBOA candidates. The main source of hemorrhage identi ed was categorized as pelvic fractures, abdominal solid organ injury (liver, spleen, kidney, pancreas), lower extremity amputation, or major vascular injury.

Data Collection
Records for each TTA were accessed through the Health Authority's Patient Care Information System (PCIS). Each case was screened by a reviewer for inclusion and exclusion criteria and data from each identi ed case was extracted using a standardized data collection form (Appendix B). The data was subsequently copied into a secure electronic database.
Based on the de ned indications for REBOA, the reviewers classi ed each case as whether or not the patient would have been a Potential or Likely REBOA Candidate. Potential Candidates were de ned as those that met inclusion criteria without any exclusion criteria. Likely Candidates were those that met criteria for REBOA candidacy and received four or more units of packed red blood cells. All cases that were identi ed as a Likely REBOA Candidate were then subsequently reviewed by a panel of trauma physicians for nal determination of candidacy. Additionally, the two reviewers assessed a sample of 60 cases (9.4% of charts) to ensure inter-rater reliability of case analysis.

Statistical Analysis
Descriptive statistics were utilized. Parametric continuous data was expressed using means and standard deviation (STD), non-parametric continuous data was expressed using median and interquartile ranges (IQR), and categorical data was expressed using n values (%). The Kappa statistic was used to express inter-rater agreement between chart reviewers.

Results
The BCTR recorded 1237 consults to the trauma service at our centre from January 1, 2016 to December 31, 2018. In total, 635 of these consults were TTAs. Following review, 21 patients were classi ed as Potential REBOA Candidates (3.3% of TTAs) and 14 patients were classi ed as Likely REBOA Candidates (2.2% of TTAs; 1.1% of total trauma consults) (Fig. 1). Table 1 provides descriptions of each case that met candidacy criteria. The inclusion criteria met by the 14 Likely Candidates were as follows: ten were blunt trauma with a pulse and SBP < 90mmHg (71.4%), three were penetrating trauma with a pulse and SBP < 90mmHg (21.4%), and one was a blunt trauma and pulseless (7.2%). No patients met the penetrating trauma and pulseless criteria in our cohort. The mechanisms of injury sustained by Likely Candidates are reported in Table 2.  Table 3). Two of the Likely Candidates received CPR (14.3%), one received resuscitative thoracotomy with aortic cross clamping (7.1%), four underwent angiography (28.6%), and four underwent laparotomy (28.6%). Eight patients immediately went to either the OR or IR suite from the ED (57.1%). In total, three of the Likely Candidates died from their injuries (21.4%) ( Table 4).  The main sources of infra-diaphragmatic hemorrhage in the Likely Candidates were secondary to an abdominal source in ten patients (71.4%) (including four spleen injuries, nine liver injuries, and two kidney injuries), pelvic fracture in six patients (42.9%), and major vascular injury in three patients (21.4%). Ten patients had multiple sources of hemorrhage (71.4%) ( Table 2).
Sixty of the cases were assessed independently by both reviewers to assess for inter-rater reliability. Good reliability was found between reviewers for identi cation of patient candidacy (kappa = 0.659).

Interpretation of Findings
This study contributes to the evidence supporting the population-driven need for acquisition of REBOA at Canadian trauma centres. The majority of studies on the e cacy of REBOA in the management of massive hemorrhage from infra-diaphragmatic injury are from other countries, including the USA and Japan, 16,[23][24][25][26][27] in which REBOA is used more commonly; however, the trauma populations, resources, and patterns of injury are vastly different internationally.
In our study, we identi ed that 14 patients over a three-year period at our institution would have met criteria for use of REBOA during resuscitation. Those that met criteria represented a severely injured group of patients and constituted 2.2% of the TTAs and 1.1% of all trauma consults. This can be extrapolated to indicate that REBOA may be utilized in four to ve patients per year at our institution. Our ndings are consistent with the results of a recent 2021 publication that conducted a similar needs assessment for REBOA at multiple major trauma centres in Edmonton, Canada 33 . This study also found that 1.1% of the study population met eligibility criteria for deployment of REBOA. Although a seemingly small number, this may be clinically signi cant given that trauma patients are often young, previously healthy individuals with the physiologic reserve to survive this procedure. Additionally, these patients stand to gain many potential high-quality years of life.
In comparison, our institution performed between three to six resuscitative thoracotomies per year (1.0% of all trauma consults) during the study time period. Similar to REBOA, RT is a procedure that is performed rarely and requires the need for subsequent immediate de nitive management; however, in the setting of infra-diaphragmatic hemorrhage, RT is substantially more invasive and only attempted in patients that have already progressed to cardiac arrest.
Our data indicate that the most common indication for the use of REBOA in our trauma population was for patients that had sustained blunt trauma. The most common mechanism of injury in the REBOA candidates was being a pedestrian struck by a vehicle. This is consistent with Canadian statistics that show the majority of trauma in our country is secondary to blunt trauma. In comparison, the United States has a much higher rate of penetrating trauma. [30][31][32] Of note, the percentage of cases that met candidacy criteria in our study is greater than that found in a 2019 study at a US trauma centre that used similar inclusion criteria. The authors found that 0.6% of the trauma patients seen in their ED per year may have potentially bene tted from REBOA, but over half (53%) of which had sustained a penetrating traumatic injury. 20 Three patients out of the 14 Likely Candidates died from their injuries. The other 11 patients that met Likely Candidate criteria survived without REBOA. Eight out of the 14 Likely Candidates went directly to the OR or IR suite for attempted embolization. The other six patients did not end up requiring surgery or embolization. This re ects that the physician's decision to place REBOA is made early during patient assessment and sometimes without de nitive imaging. As with any intervention, its availability provides an option to the care provider that can be employed on a case-by-cases basis according to the physician's clinical judgement that it will improve the outcome of the patient.
It is notable that the mean SBP on ED arrival for both the Potential and Likely Candidates was greater than 100mmHg. This demonstrates that a patient's clinical status cannot be re ected by a single value representative of a moment in their post-injury course. In contrast to these blood pressures, the mean number of units of blood product transfused in the Likely Candidates was 5.5 in the rst hour and 39.2 total in the rst 24 hours. This meets criteria for massive transfusion at our institution and is indicative of the critical condition of these patients.

Strengths and Limitations
This study has a number of strengths and limitations. Firstly, it is a retrospective chart review so identi cation of cases that meet REBOA candidacy does not necessarily re ect that the intervention would have changed the patient outcome. This limitation is inherent to the study design but we opted for this approach as an important rst step in the assessment of the possible utility of this tool in our trauma population. Secondly, it is possible that cases were missed for inclusion in our study as we only screened TTAs. Other studies have de ned the indication for REBOA using ICD-9 or 10 codes 20 or ISS 33 ; however, TTA criteria are more clinically relevant as they are used in real time by physicians during active management of trauma cases, which is a strength of our methods. We attempted to minimize missed cases by including missed TTAs. Finally, the generalizability of this study is limited in that it was conducted at a single institution. However, our Level 1 trauma centre serves a large urban catchment area, which likely re ects a similar population at other major Canadian trauma centres.

Conclusion And Future Directions
In conclusion, our study contributes to the assessment of the potential value of REBOA in trauma management in the Canadian setting. We found that 1.1% of our annual Canadian trauma population would meet REBOA candidacy, which is similar to the percentage of RTs we perform yearly. While this is a low-volume procedure, it is one that could be potentially life-saving in the most severely injured patients.
The next step would be to conduct a formal needs assessment for the implementation of REBOA including a cost-bene t analysis and evaluation of the implications to trauma systems at Canadian institutions. The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing Interests
The authors declare they have no competing interests.

Funding
Our study received funding from the Royal Columbian Hospital Foundation Fund.

Author's Contributions
Elizabeth Purssell prepared the initial literature review and was the primary author for the rst draft of the manuscript. Sean Patrick was active in data collection and editing of the manuscript. Joseph Haegert participated in case review for the data analysis. Vesna Ivkov participated in study preparation, data extraction, and editing. John Taylor provided the methodological expertise and was active in editing of the manuscript.