Severe hemorrhage remains a leading cause of preventable mortality in trauma patients. Approximately 40% of trauma-related deaths are due to hemorrhage or its related consequences [1,2]. Non-compressible truncal hemorrhage (NCTH) represents a unique clinical challenge as it is a condition characterized by severe bleeding from within the trunk of the body that is unable to be managed through traditional compression methods. NCTH has mortality rates as high as 85% in military settings and approaching 50% in civilian patients [3,4].

Multiple studies have shown that complete aortic occlusion with devices such as REBOA is a viable resuscitative adjunct for NCTH as it mitigates hemorrhage and enhances cerebral blood flow, thus acting as an interim measure before achieving definitive hemorrhage control [5–8]. While REBOA is effective in controlling bleeding, it induces ischemia downstream from the site of occlusion resulting in severe ischemia reperfusion injury and/or irreversible organ damage. To lessen the ischemic effects induced by full aortic occlusion, techniques such as partial REBOA have evolved, allowing for partial or variable occlusion of the aorta. Partial REBOA has the potential to maintain perfusion above the level of occlusion while simultaneously establishing a permissive state of regional hypoperfusion to areas of uncontrolled hemorrhage [9–11]. As such, these devices are hypothesized to have a more favorable complication profile but the clinical data has not yet answered this question. Intermittent REBOA (iREBOA) is an additional technique that involves periods of full occlusion and periods of deflation, while partial REBOA aims to maintain hemodynamics with reduced distal flow to help mitigate the supraphysiologic pressures created during times of full occlusion [6,12]. Ultimately, there are still many uncertainties about how to utilize these techniques, including the optimal timing, patient population, and titration strategy for achieving better overall outcomes.

In an effort to understand the relative benefits of alternative methods of balloon management for patients receiving REBOA, we compared the morbidity and mortality of partial REBOA and complete REBOA using the American Association for the Surgery of Trauma (AAST) AORTA Registry. We hypothesized that partial REBOA would be associated with better outcomes than complete REBOA.

The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines and the Declaration of Helsinki [13]. All collected data was retrieved from the Aortic Occlusion for Resuscitation in Trauma and Acute Care Surgery (AORTA) registry from 2020 to 2022. The AORTA Registry is a multi-institutional initiative designed to collect prospective data on adult patients (aged 18 or older) who undergo resuscitative aortic occlusion using both open and endovascular techniques during the acute phases of injury. This data is sourced from hospitals across the United States that are verified by the American College of Surgeons as Level I or Level II trauma centers. Designated registrars at each participating center are responsible for entering the data into the online portal developed by the AAST. This data includes patient demographics, clinical characteristics, intervention characteristics, and outcomes. All adult patients (18 years or older) registered in the database who received a complete or partial REBOA to aid in the management of a traumatic injury were considered for inclusion. Patients were excluded if they suffered a non-traumatic hemorrhage, underwent intermittent REBOA, had a head Abbreviated Injury Scale (AIS) ≥3, as the majority of these patients have a dismal prognosis or complications not related to REBOA, or an AIS of 6 in any region of the body, since these injuries are generally not considered survivable.

The primary outcome of interest was any complication (myocardial infarction, stroke, paraplegia, acute kidney injury requiring dialysis, acute lung injury/acute respiratory distress syndrome, distal embolism, need for amputation, bacteremia, pneumonia, sepsis, and multiorgan dysfunction). Secondary outcome measures included in-hospital mortality, discharge Glasgow Coma Scale (GCS), Glasgow Outcome Scale Extended (GOSE), intensive care unit (ICU) as well as hospital length of stay, and time to death. For the adjusted analyses, discharge GCS was dichotomized as ≤8 and >8 while discharge GOSE was dichotomized as ≤4 and >4. Distal pressure targets and balloon titration strategy were not included in the analysis as this data is not reported in the AAST AORTA Registry.

After applying the inclusion and exclusion criteria, 164 patients were deemed suitable for further analysis (Figure 1). In total, 64% (N = 105) were managed using a complete aortic occlusion, while 36% (N = 59) were subjected to a partial aortic occlusion. Patients managed using a complete occlusion were older (40 vs 33 years, p = 0.017), had a higher presenting GCS (14 [8–15] vs 12 [3–14], p = 0.025), and were more likely to be hemorrhaging from the pelvis (18.1% vs 16.9%, p = 0.030) as well as less likely to be hemorrhaging from the head or neck (0% vs 8.5%, p = 0.030). Those who underwent complete occlusion were also less likely to be undergoing CPR on admission (6.7% vs 18.6%, p = 0.035) as well as more likely to be admitted to a level I trauma center (99% vs 91.5%, p = 0.023). There were no statistically significant differences in sex, injury severity, or admission vitals (Table 1).

For patients with Zone 3 placement, a greater percentage of them had complete REBOA (39% vs 18.6%, p = 0.007). Additionally, more patients who were treated with complete REBOA later received a pelvic external fixator (14.3% vs 3.4%, p = 0.032). There were no significant differences in technique for arterial access, final catheter sheath diameter, rate of successful arterial access, survival to removal of access sheath, hemodynamic stability, time to hemodynamic stability, or other interventions performed (Table 2).

In the univariate analysis, there were no statistically significant differences in the rate of complications, ICU or hospital length of stay, in-hospital mortality, or time to death (Table 3). There was a statistically significant difference in median discharge GCS (5.0 [3.0–15] vs 15 [6.0–15], p = 0.011) but no statistically significant difference in discharge GOSE (Table 3). After adjusting for potential confounding in the Poisson regression analysis, no statistically significant difference in complications was detected when comparing partial to complete REBOA [adjusted IRR (95% CI): 1.11 (0.54–2.27), p = 0.777]. This was also the case for all secondary outcomes (Table 4).

In our registry analysis, we found no statistically significant differences in any complications between patients who received partial REBOA and complete REBOA. This includes complications such as myocardial infarction, stroke, paraplegia, acute kidney injury requiring dialysis, distal embolism, need for amputation, and multi-organ dysfunction. There were also no statistical differences in ICU or hospital length of stay, discharge GCS or GOS, in-hospital mortality, or time to death.

These results were surprising given that several preclinical models have demonstrated that partial REBOA reduces ischemia-reperfusion injury and allows for longer balloon inflation time [14–16]. However, this study is subject to the inherent limitations of a retrospective multicenter registry of time-sensitive, life-saving interventions. One of the main limitations of this study is that important clinical metrics including information on duration and type of partial REBOA were not fully characterized in the AAST AORTA registry. In addition, more than 40% of patient entries were missing time to definitive hemorrhage control data. Taken together, the omission of these key metrics and lack of granularity hinders investigators’ ability to fully interpret and draw conclusions from the registry. This missing data is key for understanding clinical efficacy as it is well established that longer periods of occlusion are associated with increased complications [9,16]. The absence of complete and in-depth data poses additional challenges for researchers, as patient data points are missing from various, inconsistent areas, creating a highly heterogeneous database.

Despite several preclinical studies showing improved outcomes with partial REBOA, there is limited clinical data to advocate its use. In an analysis of the Aortic Balloon Occlusion (ABO) trauma registry, Paran et al. found no difference in mortality among patients who underwent partial vs complete occlusion of the aorta [17]. The results of our study support these findings in that there was no significant difference between partial and complete occlusion groups both for complications and mortality. Further clinical evidence is warranted to define the superiority of partial REBOA over complete REBOA.

Most notably, the aortic occlusion strategy was self-reported by the centers along with type of balloon used (i.e. Prytime ER-REBOA or p-REBOA PRO) and it lacks the granularity to determine how clinicians were implementing partial REBOA. This includes no information regarding their balloon volume titration strategy or distal pressure targets. Further, reported balloon placement was confirmed by plain film, albeit inconsistently, or in rare cases with computer tomography (CT) fluoroscopy. Understanding the method of partial REBOA titration is critical because small changes in balloon volume can cause large changes in flow downstream [18]. Depending on how the balloon is titrated, it is possible to induce an intermittent occlusion phenomenon in which downstream flow is either completely arrested or fully restored. This is in contrast to partial occlusion as is intended, with only 10–20% of downstream flow allowed. The differences between partial and intermittent REBOA can be subtle to the provider at the bedside, but can certainly impact hemodynamics and overall hemorrhage control. Without high-fidelity hemodynamic data, such differences are difficult to tease out.

The above referenced 10–20% of downstream flow allowed is based on pre-clinical research involving the use of partial REBOA [18–21]. While there is no universally fixed definition, this range is a widely accepted target that balances the need for aortic occlusion while preserving some distal perfusion. In preclinical animal models, direct measurement of downstream flow has been achieved using flow probes that are capable of precise flow measurements. In clinical settings, direct measurement of downstream flow is more difficult but can be estimated by distal pressure targets and Doppler ultrasound. There are several preclinical studies that seek to correlate downstream flow to the distal mean arterial pressure (MAP) below the balloon [18,19,22–24]. These studies demonstrate a fairly linear relationship of distal MAP to flow across various states of hypovolemic shock. Given the lack of granularity in the AORTA registry, and the inability to provide direct flow measurements in a clinical context, this data does not exist.

Without a comprehensive understanding of downstream flow and more precise accounting of relevant variables, the conclusions we can draw are limited. This realization should serve as a caution regarding the limitations inherent in this registry, which is particularly relevant given that many REBOA studies utilize the AORTA registry. While some studies may be designed to address the data shortcomings, others may not, especially when describing outcomes directly related to distal ischemia, as in our present paper. Our data sheds light on the need for improving the granularity of the AAST AORTA registry and reaching a consensus on the definition of partial REBOA, which will allow for better analysis and interpretation of REBOA groups.

Finally, while there were no statistically significant differences for indication between the partial REBOA vs the complete REBOA groups, the indication for use is widely varied. These indications include arrival to the emergency department in arrest, hemodynamic stabilization for additional workup with cross-sectional imaging, transport to the operating room or interventional radiology suite, placement of REBOA intraoperatively for emergency surgery, and placement for planned elective surgery. Increasing the sample size in the registry and completeness of the database will allow us to better analyze these vastly different indications to help determine which patient populations might benefit from the use of endovascular hemorrhage control devices.

In conclusion, endovascular technologies such as REBOA have emerged as a valuable tool in the management of NCTH; however, its use remains controversial due to associated morbidity and uncertainty about which patient groups will benefit. While some preclinical studies have demonstrated that partial REBOA can reduce ischemia reperfusion injury, in our registry analysis we found no statistically significant difference in complications between patients who received partial or complete REBOA. These findings may suggest that the observed reduction in ischemic injury in preclinical studies may not necessarily translate to a decrease in patient complications. However, the current body of clinical data falls short in providing the nuanced insights required to address these crucial questions. To understand which patient populations will benefit from these devices and how to best implement them, we need to improve the granularity of the data from which we are studying them. Ultimately, this study is a call for increased enrollment in the database, commitment to data integrity, and attention to detail in recording patient variables.