Study population
This retrospective cohort analysis included 26 consecutive adult trauma patients with blunt or non-thoracic, penetrating injuries and refractory hemorrhagic shock managed at a level 1 trauma center with a dedicated trauma hybrid operating room. Derivation of the study population is illustrated in Supplemental Figure 1. There were 1,527 level 1 (highest acuity) trauma team activations during a 52-month period ending November 2019. These cases were identified in a prospective, institutional trauma registry.
Twenty-six adult (age 18 years or greater) patients with blunt or non-thoracic, penetrating trauma that underwent zone 1 aortic occlusion for refractory hemorrhagic shock, defined as systolic blood pressure less than 90 mmHg with a transient or no response to volume resuscitation. Aortic occlusion was performed by resuscitative thoracotomy and aortic cross-clamping in 13 patients and by REBOA in 13 patients. Cases of penetrating, thoracic injuries with hemorrhagic shock are a contraindication for REBOA and were therefore excluded. In these patients, thoracotomy alone provides direct exposure for operative control of exsanguinating hemorrhage. Institutional Review Board approval was obtained.
Primary outcome and power analysis
The primary outcome was survival to hospital discharge. A power analysis was performed to estimate the number of subjects per cohort that would be necessary to detect a statistically significant difference in survival to hospital discharge. Because the availability of a dedicated trauma hybrid operating room could affect the technical performance, timing, and efficacy of REBOA balloon positioning and other angiographic hemorrhage control procedures, the power analysis was performed using data from the authors’ institution. In a prior study of REBOA use in a mixed population of trauma and non-trauma patients from the author’s institution, 30-day survival was 38%; survival among eight patients undergoing resuscitative thoracotomy during the same period was 0% (9). During internal quality control audits of the REBOA experience at the authors’ institution, it was noted that survival to discharge had reached 54%. Using these proportions and setting α and β to conventional values of 0.05 and 0.80, respectively, the present study would be adequately powered to detect a statistically significant difference in survival to hospital discharge with 13 patients in each cohort (13).
Secondary outcome and data collection
The secondary outcome was discharge GCS among survivors. Discharge GCS and other variables that were not available within the authors’ prospective institutional trauma registry were obtained by manual review of the electronic health records. Variables describing patient characteristics are shown in Table 1. Variables describing patient management are shown in Table 2. Timing of hemorrhage control was defined as attaining systolic blood pressure 100 mmHg or greater without ongoing vasopressor or blood product transfusion requirements or subsequent episodes of hypotension with systolic blood pressure less than 90 mmHg, consistent with consensus recommendations regarding blood pressure targets in damage control resuscitation after trauma (14). Variables describing resuscitation included tranexamic acid administration within four hours and transfusion of red blood cells and plasma within 24 hours. Additional, tertiary outcomes included survival past the emergency department, survival past the operating room, lengths of stay in the hospital and ICU, days on mechanical ventilation, discharge disposition, and complications classified by the Clavien-Dindo system that was adapted for trauma by Naumann et al. (15)
Trauma hybrid operating room, REBOA equipment, technique, and staff training
Our institution built a dedicated trauma hybrid operating room in a repurposed and remodeled angiography suite on the second floor of the hospital directly above the emergency department. The hybrid operating room contains a ceiling-mounted C-arm and a fluoroscopy-compatible table with tilt functions. The hybrid operating room is immediately adjacent to a fluoroscopy control room.
According to institutional protocols, REBOA catheters were placed by trauma surgeons for patients with blunt or non-thoracic penetrating trauma, hemorrhagic shock (i.e., systolic blood pressure <90 mmHg), and a transient response or no response to volume resuscitation (9). The REBOA procedure was either initiated in the emergency department for some patients or in the operating room for others, depending on their clinical trajectory and hemodynamic response to resuscitation. Balloon inflation was initially performed in Zone 1 for all patients. Subsequent balloon deflation and relocation to Zone 3 was at the discretion of the attending trauma surgeon based on hemodynamic response, completion of operative hemorrhage control techniques, and injury patterns. Initially, REBOA was performed with a 12-Fr introducer and aortic occlusion balloon designed by Cook Medical (Bloomington, Indiana). Subsequently, REBOA was performed using a 7-Fr introducer and aortic occlusion balloon designed by Prytime Medical (Boerne, Texas). Notably, both aortic occlusion balloon catheters must be deflated to obtain distal aortic blood flow around the balloon, in contrast to newer catheters that allow a more controlled amount of distal aortic blood flow through channels within the balloon. The senior author trained trauma surgeons and senior residents (i.e., PGY4 and PGY5 residents) in REBOA concepts and techniques with a series of 90-minute audiovisual presentations followed by hands-on simulation training (9).
Statistical analysis
To determine whether REBOA would be associated with a higher rate of neurologically intact survival, the primary outcome (hospital discharge alive) and secondary outcome (GCS at hospital discharge) were directly compared between resuscitative thoracotomy and REBOA cohorts. To understand whether potential differences in outcomes were attributable to baseline patient characteristics and other hemorrhage control and resuscitation parameters, these factors were also directly compared between resuscitative thoracotomy and REBOA cohorts. Continuous variables were compared by the non-parametric Kruskal-Wallis test and reported as median values with interquartile ranges. Binary variables were compared by Fisher’s Exact test and reported as raw numbers with percentages. Significance level was set at 5%.