Secondary Logo

Journal Logo


A three-tier Rescue stent improves outcomes over balloon occlusion in a porcine model of noncompressible hemorrhage

Go, Catherine MD; Elsisy, Moataz BS; Chun, Youngjae PhD; Thirumala, Parthasarathy D. MD; Clark, William W. PhD; Cho, Sung Kwon PhD; Demetris, Anthony J. MD; Tillman, Bryan W. MD, PhD

Author Information
Journal of Trauma and Acute Care Surgery: August 2020 - Volume 89 - Issue 2 - p 320-328
doi: 10.1097/TA.0000000000002715


Noncompressible torso hemorrhage remains one of the leading, yet potentially treatable, causes of death after penetrating trauma. Logistics are magnified on the battlefield where multiple injuries, austere environment and transport delays represent challenges to prompt and definitive vascular repair. Traditional hemorrhage control by thoracotomy with aortic cross-clamping has being increasingly replaced by the advent of less invasive balloon occlusion, known as resuscitative endovascular balloon occlusion of the aorta (REBOA).1,2 Despite the advantages of this minimally invasive approach, a survival benefit has not yet been apparent. In a large clinical series comparing open and endovascular aortic occlusion (AO), over 75% of REBOA were placed in zone I of the aorta3 (the descending thoracic aorta between the subclavian and celiac branches) and mortality of over 70% after REBOA was not improved relative to open cross-clamp. With balloon occlusion times in excess of 20 minutes in over half of those patients, malperfusion of the viscera, spinal cord, and lower extremities is well above the threshold of warm ischemic injury for both kidneys and the liver.4–7 To this end, cautionary recommendations when using REBOA have been highlighted by the Basic Endovascular Skills for Trauma Study Group.8 In summary, ischemic injury may be an important barrier to survival after AO, and distal malperfusion may be curtailing improvements in survival after noncompressible vascular injuries. In addition, although REBOA provides effective antegrade AO, concerns of ongoing retrograde hemorrhage have not been addressed by previous porcine studies, as those studies lacked an actively hemorrhaging injury.9,10

Stent grafts would provide complete coverage of a vascular injury with preserved distal flow, but obstacles include the need to preserve flow to the viscera, uncertainty about the location of the injury in an emergent setting, challenges of wire catheter exchanges for nonvascular physicians, heterogeneity in vascular dimensions, and the permanent nature of currently available stents. We have previously reported on a retrievable stent design for control of both aortic and caval hemorrhage11,12 with preserved distal aortic flow and venous return, respectively. The design allows for emergent use and yet retrieval at the time of permanent repair by sheath advancement over a permanently affixed delivery wire. A limitation of the single-tier stent is that the location of the injury must be known, which is often not clear in exsanguinating, multi-injured patients. A strategy was developed to cover the most “at risk” aorta while preserving critical organ blood flow. A three-tier retrievable stent was designed with covered stent sections to provide simultaneous hemorrhage control in both thoracic and abdominal regions yet an interval bare metal section to preserve visceral perfusion (Fig. 1).

Figure 1
Figure 1:
Three-tier Rescue stent. (A) In the setting of either thoracic or infra-renal aortic injury, femoral access is achieved. (B) A sheathed stent is positioned relative to the visceral vessels. (C) The sheath is withdrawn to deploy the stent with coverage of the thoracic and infrarenal aorta, with preserved visceral and distal flow. (D) At the time of permanent repair, the stent is recaptured by sheath advancement.

The goals of this study were to compare a three-tier Rescue Stent to REBOA in a porcine model with an active aortic injury and specific assessment of retrograde hemorrhage, spinal cord ischemia, and visceral ischemic injury. The stages of the study included aortic injury, followed by damage control for 1 hour (to simulate transport of an injured soldier from the battlefield to definitive care) and concluded with definitive aortic repair.

We hypothesized that continued retrograde hemorrhage and malperfusion are contributors to the persistent high mortality after traumatic aortic injury after REBOA. We further reasoned that a three-tier Rescue stent could achieve improved hemorrhage control while preserving critical perfusion.


Three-tier Retrievable Rescue Stents

The three-tier Rescue stent was laser welded with 0.0155-inch thick nitinol wire (Confluent, Fremont, CA) in a stem and petal design onto a tapered 55-cm-long mandrel.13 Tips of a dilator and floppy guidewire were affixed to the forward end as an integrated guiding nosecone. Distally, the stent was permanently attached to a delivery wire (Lunderquist; Cook Medical, Bloomington, IN) for both deployment and recapture. Polytetrafluoroethylene (PTFE) tubing (Zeus, Orangeburg, SC) (diameter, 10 mm; thickness, 0.127 mm) was balloon dilated to the dimensions of the nitinol scaffold.13 The PTFE was affixed to top tier of the stent for 30 cm (thoracic tier) and the bottom tier for 11 cm, while the middle 12 cm of the stent was left as bare metal for visceral perfusion, as shown in Figure 2. The final diameter of the deployed self-expanding stent was 21 mm proximally and 14 mm distally, corresponding to a 50% oversizing relative to the aortic measurements in the 80-kg porcine model. Tantalum markers (Goodfellow Corp, Coraopolis, PA) were affixed at tier junctions to facilitate fluoroscopic placement. The completed stent was ethylene oxide sterilized after collapse into an 80-cm-long 10-Fr sheath (Oscor, Palm Harbor, FL).

Figure 2
Figure 2:
(A) Nitinol scaffold of a three-tier Rescue Stent constructed with a stem-and-petal design. (B) Completed stent with PTFE lining in zones 1 and 3 for coverage of injuries to the descending and infrarenal aorta, respectively. Zone 2 is bare metal to allow for continuous flow to the visceral branches. A nosecone with guidewire tip was integrated into the rest of the stent. A 100-cm delivery long wire was permanently affixed to the stent scaffold. (C) The stent was collapsed into a 10-Fr sheath with the integrated nosecone/guidewire visible at the leading edge.

Animals and Monitoring

This study approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. Yorkshire cross swine (77–87 kg) were sedated followed by endotracheal anesthesia with isoflurane. Amiodarone 10 mg/kg was administered to prevent fatal ventricular arrhythmias common in the porcine model.14,15 Hemodynamic monitoring was initiated by carotid and femoral arterial lines, as well as a pulmonary artery catheter (Edwards Lifesciences, Irvine, CA). Baseline angiograms, neurologic monitoring, mean arterial pressure (MAP), cardiac output (CO), and central venous pressure (CVP) were recorded.

Neurophysiologic Testing

Somatosensory-Evoked Potentials (SSEP) were obtained by nerve stimulation in all limbs (Rhythmlink International LLC, Columbia SC). Spinal cord ischemia was predicted by an over 50% reduction in SSEP response compared with baseline values11 and with irreversible loss of SSEPs from both hindlimbs indicative of paraplegia.12 Interpretation was performed by a blinded neurophysiologist. Motor-evoked potential (MEP) electrodes were placed as previously described8 with stimulation intensity ranged from 200 V to 600 V, with a train of five pulses, and interstimulus interval of 5.0 ms, and duration 0.05 ms (Protector IOM, XLTEK; Natus Medical, Pleasanton, CA). Irreversible loss of lower extremity MEP is a predictor of paraplegia in porcine9 and human studies.10 To prevent the known impact of inhaled anesthetics on neurophysiologic monitoring during the damage-control phase (DCP), isoflurane was transitioned to dexmedetomidine and propofol during assessment. Neurophysiologic testing was confirmed by clinical postoperative neurologic examination included a modified Tarlov physical examination score performed at baseline and then daily postoperatively.16

Porcine Vascular Injury Model

Animals were divided into three groups for REBOA, Rescue stent after either thoracic aortic injury (Rescue-T), or abdominal aortic injury (Rescue-A). Midline laparotomy allowed for visceral flow measurements among selected animals. The animals were heparinized with 100 units/kg of heparin to simulate posttraumatic coagulopathy and as required by the porcine model due to rapid procedural thrombosis known in this model. The REBOA and Rescue-T groups additionally underwent a left-sided thoracotomy with exposure of the descending thoracic aorta. Aortic injury was created with a 22-Fr dilator of the thoracic aorta (REBOA and Rescue-T) or infrarenal aorta (Rescue-A). Following aortic injury, either REBOA or the Rescue stent was immediately deployed to initiate a DCP. A REBOA balloon (n = 7; Prytime Medical, Boerne, TX) from a femoral access was inflated in the mid-descending thoracic aorta, proximal to the aortic injury. Rescue-T (n = 6) and Rescue-A (n = 6) groups each received the same 55-cm-long three-tier Rescue stent, deployed from a femoral access by sheath withdrawal from just distal to the left subclavian and down to the aortic bifurcation. Radiopaque stent markers on the stent were positioned distal to the left subclavian and matched with external landmarks, fluoroscopically.

DCP and Permanent Repair

The Damage Control Phase (DCP) represented the time from initiation of hemorrhage damage control (REBOA or Rescue stent) until definitive repair. Sixty minutes was chosen to simulate the multiple potential barriers to definitive repair, particularly on the battlefield, such as transport delays, adverse terrain, and enemy activity. During the DCP, hemodynamics of were monitored. Volume flow rates of the visceral arteries were recorded (Transonic Systems, Ithaca, NY and AD Instruments, Colorado Springs, CO). All shed blood was recovered with a cellsaver (BRAT2; Sorin Group, Arvada, CO) and returned immediately to the animal. During a 60-minute DCP, postinjury hypotension was managed with epinephrine, phenylephrine, norepinephrine, and crystalloid for a target MAP of 50 mm Hg. Angiography interrogated hemorrhage and visceral perfusion. After the DCP, the REBOA was deflated or the Rescue stent recaptured by sheath advancement, to allow for permanent suture repair of the injury. Of importance, the REBOA was deflated gradually in anticipation of reported aortic hyperemia.17 Laboratory values were drawn at baseline and after both injury and resuscitation. Despite a planned 1-hour DCP for all animals, death of four 60 minute REBOA animals due to profound hypotension and arrest prompted the intervention time to be shortened to 30 minutes for the last three REBOA animals. Rescue stent groups all completed a strict 1 hour DCP.


Following planned necropsy at 48 hours postinjury, core biopsies of saline flushed kidneys and liver or rings of bile duct were formalin-fixed and paraffin-embedded. Hematoxylin and eosin slides for each organ were graded by a blinded transplant pathologist using the following scale: greater than 1, negligible injury; 1 to 1.9, mild injury; 2 to 2.9, moderate injury; 3 to 3.9, severe injury; 4 or greater, necrosis. Specific criteria for each organ included vacuolization in the kidney, centrilobular necrosis in the liver, and loss of biliary epithelium.

Statistical Analysis

A sample size of six animals per group was calculated to provide 80% power (alpha = 0.05) for detection of difference in physiologic measures and the differences detected at this power level were considered clinically significant for this study. Data were examined using STATA 15 software (StataCorp, College Station, TX). For categorical data, χ2 analysis was used. For each continuous, dependent variable, a repeated-measures analysis of variance was computed. Each analysis of variance was then followed by post hoc testing with Scheffe adjustment. Animals with incomplete laboratory data at specific timepoints were excluded from analysis. Averages of physiologic, fluid, and laboratory values are shown with standard deviations or standard error.



Following the DCP and gradual balloon deflation, the first four animals of the REBOA group (REBOA-60) developed profound, refractory hypotension leading to arrest with a single animal surviving only to 5 hours postoperatively. As a result of this universal mortality from 60 minutes of balloon occlusion, the last three animals underwent REBOA occlusion for 30 minutes (REBOA-30). Despite only half the occlusion DCP time, the last three REBOA-30 animals also died from refractory hypotension with cardiac arrest despite defibrillation and pressors. Within the Rescue groups, the stent was deployed in under 2 minutes from the time of arterial access for all animals. All Rescue stent–treated animals survived the DCP, stent recapture, aortic repair, and emergence from anesthesia. Two animals in the Rescue-T group expired from cardiac arrhythmias 2 to 3 hours postoperatively.


Following injury of descending thoracic aorta (Fig. 3A), antegrade extravasation was observed angiographically (Fig. 3B). After REBOA inflation, there was successful AO proximal to the injury (Fig. 3C) but ongoing retrograde extravasation was observed on imaging from the distal aorta (Fig. 3D). Among the Rescue-T animals, thoracic hemorrhage was well controlled by the Rescue Stent (Figs. 4A, B), and the same stent design also controlled abdominal hemorrhage among animals of the Rescue-A group (Figs. 4D, E). Mesenteric and lower-extremity perfusion were both preserved as a result of the bare metal visceral and distal aspects of the stent, respectively (Figs. 4C, F). Operative images of the aortic injuries are depicted in Supplemental Digital Content 1, Figure 1, The average blood loss of the REBOA, Rescue-T, and Rescue-A groups were 3.46 ± 0.93 L, 1.39 ± 0.36 L, and 0.53 ± 0.48 L, respectively (p < 0.001). Baseline hemoglobins were similar and there was no significant change between groups after resuscitation.

Figure 3
Figure 3:
Retrograde hemorrhage after REBOA. (A) Injury of the descending thoracic aorta with a 22-Fr dilator (B) Hemorrhage (dashed circle) from the injured aorta (C) Antegrade hemorrhage control of by the REBOA balloon (arrow). The injury access wire was left in place as a reference of the injury location (black arrowhead). (D) Persistent retrograde hemorrhage (dashed circle) below the REBOA (arrow).
Figure 4
Figure 4:
A three-tier Rescue stent angiographically controled hemorrhage of the thoracic and abdominal aorta (A) Injury of the descending thoracic aorta (arrow) revealed hemorrhage (dashed line). (B) The Rescue stent achieved hemorrhage control in the thoracic aorta. (C) Mesenteric perfusion was preserved in the bare metal section. (D) Injury of the infra-renal aorta (arrow) with hemorrhage (dashed line). (E) The lower covered section of the Rescue stent mediated hemorrhage control of the abdominal aortic injury, while distal flow to the iliac arteries was preserved (F).


The lower baseline in MAP of thoracic as compared with abdominal injury animals appears to be related to increased isoflurane anesthetic necessary for muscle relaxation during exposure for the thoracic injury, this finding has previously been reported by others.18 Despite similar baseline pressures, MAP fell by an average of 54.6 ± 18 mm Hg in REBOA yet only 19.5 ± 17 mm Hg in the Rescue-T group (p = 0.0047) after injury and amidst damage control. Despite replacement of all shed blood, resuscitation, and ongoing antegrade hemorrhage control, MAP of the REBOA group never recovered to baseline, while the Rescue-T group recovered with resuscitation and the Rescue-A group exhibited no significant drop in MAP throughout the DCP. Both Rescue groups maintained a MAP significantly greater than that of the REBOA group postinjury (Fig. 5A). Additionally, the Rescue-A group required significantly less vasopressor support than the REBOA group (p = 0.036). After injury, the CO of both REBOA and Rescue-T animals declined significantly from their respective baselines, while Rescue-A animals did not deviate from baseline (Fig. 5B). The CVP of the Rescue groups remained significantly higher than that of the REBOA animals (p = 0.007, Fig. 5C). By the end of the DCP, the decline of right ventricular end diastolic volume of REBOA (Fig. 5D) was significantly greater than that of the combined Rescue animals (124 ± 63.6 mL vs. 6 ± 45.2 mL; p = 0.007).

Figure 5
Figure 5:
Improved hemodynamics of Rescue stent vs. REBOA animals. Following aortic injury RESCUE stent–treated animals maintained improved (A) MAP, (B) CO, and (C) CVP as compared with REBOA animals. Although REBOA animals maintained RVEDV early, this declined precipitously over the DCP. Statistical significance was demonstrated between: ‡ Rescue groups vs. REBOA, * Change from baseline Rescue-T and REBOA, and † Change from baseline in combined Rescue animals compared with REBOA.

Neurophysiologic Spinal Cord Ischemia

Preoperatively, all animals were ambulatory with normal baseline MEP and SSEP waveforms. Among REBOA animals during the DCP, SSEP, and MEP were permanently lost within 20 minutes of REBOA (Supplemental Digital Content 2, Figure 2, with the exception of only a single animal of seven recovering these waveforms prior to death. Among both Rescue groups, two animals had no changes and the remainder has dampened SSEP during the DCP, but these returned to normal in all animals. Motor-evoked potentials were dampened during the DCP, but recovered upon stent removal in all animals (except one Rescue-T animal where the cranial electrode became dislodged). All Rescue stent animals demonstrated movement of all extremities upon emergence from anesthesia, with further improvement to baseline after full recovery.

Preserved Visceral Flow

Flow to the viscera among Rescue groups was demonstrated angiographically, as well as by ultrasonic flow measurements (Supplemental Digital Content 3, Figure 3, Availability of ultrasonic flow measurements was only possible in three REBOA animals but all of the Rescue stent animals. Baseline flow measurements were similar among all groups. During the DCP, visceral flow decreased more in the REBOA animals as compared with the Rescue animals (Supplemental Digital Content 4, Table 1, Among Rescue animals, the decrease in visceral flow correlated with a decrease in MAP (correlation coefficients range from 0.338 to 0.698; p < 0.05) and suggests that visceral flow reductions were likely related to hypotension in this hemorrhagic model.

Serologic Assessment of Organ Function

Serologic measures of renal and hepatic injury could not be assessed among REBOA animals secondary to early postoperative death. As an assessment of renal function among Rescue animals, creatinine levels on postoperative day 2 averaged 1.8 ± .53 mg/dL which was not significantly different from baseline (1.5 ± 0.15 mg/dL, p = 0.082) and which falls well within the range of normal creatinine for the porcine model.19 Of note, one animal was excluded from creatinine analysis after receiving an excessive contrast dose during stent repositioning. Next, using ALT as the most specific of aminotransferases for liver injury, the ratio of average ALT at 24 hours or 48 hours after injury to baseline was less than 5 which corresponds to mild hepatic injury among all Rescue animals surviving to 24 hours and 48 hours.


Organ histology of all 10 Rescue stent–treated animals surviving to 48 hours postoperatively were examined by a blinded pathologist and compared with the single REBOA animal surviving 5 hours into the postoperative period. Representative images are shown in Figure 6. Among Rescue-treated animals, the scoring revealed negligible ischemic change in the kidneys, liver and bile duct. Conversely, although the single REBOA animal had only 5 hours to manifest ischemic changes, the scoring already revealed mild ischemic change of the kidneys (vacuolization), severe ischemic change of the liver (centrilobular necrosis), and mild ischemic change of the bile duct (loss of epithelium).

Figure 6
Figure 6:
H&E–stained kidney, liver and bile duct. A REBOA animal revealed renal vacuolization (A), hepatic centrilobular necrosis (C) and loss of biliary epithelium (E), features that were not significant among Rescue animals (B, D, F).


The challenging logistics of civilian noncompressible hemorrhage are magnified on the battlefield, where resources and experience are limited, while transport delays to definitive care are often prolonged. New approaches to damage control are crucial in order to impact upon the exceptionally high mortality of noncompressible hemorrhage. This is especially important when delays are expected between levels of care, such as in a battlefield setting or in mass casualties, when patients overwhelm the resources of a hospital.

REBOA has represented a less invasive approach for damage control of noncompressible hemorrhage and yet has demonstrated real-world mortality that remains overall poor and largely similar to open aortic cross-clamp.3 Earlier porcine studies of aortic zone 1 REBOA had suggested more favorable outcomes9,10 that contrast to the dismal outcome of the REBOA group of this study, likely reflective of key differences in the porcine models. Previous studies detailed “controlled” 35% blood loss in the absence of an actual aortic injury.9,10,20 By contrast, an actual aortic injury with “uncontrolled” hemorrhage, as detailed in this study, is more likely to be representative of true traumatic scenarios, where hemorrhage is an ongoing phenomenon. Proximal AO obviously provides antegrade hemorrhage control, yet the untended vascular injury remains a source for significant ongoing retrograde hemorrhage. The average REBOA blood loss in this study was 3.5 L (64% of blood volume) which is 1 L to 2 L greater than reported in controlled hemorrhage models.9,10 Hemodynamically, while prior studies of controlled hemorrhage revealed normalized proximal MAP after balloon occlusion,9,10,20 this study of uncontrolled hemorrhage revealed significant reduction in MAP. Although retrograde blood loss occurs distal to the occlusion balloon, this may yet impact perfusion above the occlusion balloon by virtue of reduced venous return to the heart.

Spinal cord ischemia is common after proximal aortic surgery in humans, and this study examined neurophysiologic monitoring during aortic damage control with findings of marked spinal cord ischemia after balloon occlusion. Although we were unable to corroborate the neurophysiologic evidence of spinal cord ischemia with a postinjury clinical examination among REBOA animals (secondary to death), the study by Long et al. also revealed marked clinical paralysis after REBOA.9 Spinal cord ischemia in the setting of hypotension during thoracoabdominal repair is well established,21 and it is possible that ongoing hypotension magnifies spinal cord ischemia in this REBOA group as well. It is still not clear if prevention of spinal cord ischemia observed with the Rescue stent is by virtue of improved hemodynamics or by preserved distal perfusion, such as pelvic collaterals.

It is notable that despite the absence of associated injuries and complete replacement of all shed blood, mortality in the REBOA group was 100% whether the balloon occlusion period was 30 minutes or 60 minutes. Universal mortality did not allow for laboratory findings of hepatic or renal ischemia, which can take days to become apparent both serologically and even histologically. Regardless, a single short term REBOA survivor revealed significant histologic organ ischemia after only 5 hours of reperfusion, which contrasted to the negligible changes among Rescue groups. Our study is in agreement with two prior studies8,22 documenting adverse outcomes of complete balloon occlusion of the aorta. In summary of our REBOA findings, this model of uncontrolled hemorrhage concludes that balloon occlusion results in significant retrograde hemorrhage, spinal cord ischemia, hypotension and histologic organ ischemia not observed among Rescue stent–treated injuries.

While providing direct injury coverage and preserved distal flow, there are numerous obvious barriers to current stent designs in exsanguinating noncompressible hemorrhage. These include the urgency of stent placement, absence of endovascular expertise, the need to preserve visceral perfusion, heterogeneity in human vascular dimensions, uncertainty about injury location in an emergent setting, and the permanent nature of current stents. To begin to approach these complex logistics, we have previously reported on a simple (single tier) retrievable stent design that demonstrated rapid hemorrhage control of aortic injury.11 We have further demonstrated improved survival when applied to injuries of the vena cava.12 A notable feature of the “stem and petal” design is rapid retrieval of the stent at the time of a controlled, permanent repair. The stents of this study are created from the shape memory alloy nitinol and are therefore self-expanding, which is especially favorable for use in a variety of vessel diameters. Traditional permanent stents require high radial force to prevent migration, whereas the retrievable stents of this study remain permanently attached to a delivery wire and there is no risk of migration. As a result, the radial force of these retrievable stents was reduced to both facilitate recapture and further minimize, if not eliminate, iatrogenic injury. As further evidence that a single stent could accommodate a spectrum of aortic diameters, our group has previously demonstrated efficacy for hemorrhage control in conduits with diameter as much as 50% smaller than the stent set diameter.23

A limitation of the single-tier Rescue stent11 is that a clinician would need to know the location of the injury to position the stent in either the thoracic or abdominal aorta. This would be a challenge in the multi-injured, unstable, and acutely exsanguinating patient without prior imaging, especially in a battlefield setting. A strategy was adopted to empirically cover both areas of highest risk for injury, while preserving flow to areas of the highest potential risk of ischemic injury. This strategy underscored the development of a three-tier Rescue stent, essentially the length of the entire aorta with covered thoracic and infrarenal abdominal aortic sections but bare metal in the middle section to preserve visceral perfusion. Contrasting to the current use of vascular stents in trauma, the three-tier Rescue stent is not intended as a permanent repair. Rather the stent represents a rapidly delivered damage-control maneuver to bridge between physicians who first encounter the exsanguinating patient until definitive repair of hemorrhage in a more controlled environment. The single piece design in this study with integrated nosecone and attached delivery wire removes the need for exchanges of different wires, dilators and catheters by personnel inexperienced with this sequence of events.

Arguably, exsanguinating hemorrhage is not exclusive to injuries of the great vessels, and, in fact, aortic injuries are far outnumbered by hemorrhage from solid organs and within the pelvis. Proof-of-concept for the Rescue stent in several other hemorrhage scenarios is shown in Supplemental Digital Content 5, Figure 4, In the case of profound hemodynamic instability at initial trauma presentation, partial stent deployment creates functional AO to allow for preserved cerebral and coronary perfusion, with full deployment after initial resuscitation. Alternately, the covered section can be deployed briefly but intentionally low to cover visceral or iliac vessels, thereby reducing solid organ hemorrhage and pelvic hemorrhage, respectively, while improving operative visualization. As a result, further investigation of the Rescue stent seems warranted across a spectrum of hemorrhagic traumatic injuries.

While the stents of this study were placed under fluoroscopic guidance, this modality may not be available in the emergency setting. In a trauma bay or military field hospital, positioning relative to bone landmarks by a digital portable X-ray may represent one approach for positioning the stent. Alternately, a signal emitting tag mounted on the stent as previously described11 may eventually offer portable, real-time positioning of a stent relative to palpable anatomic landmarks. Potential obstacles of the Rescue Stent relate to vascular pathology and are compounded by the anticipated heterogeneity of the human population. In fact, the demographics of most penetrating trauma patients are under age 40,24 and fall outside the range of typical patients with atherosclerotic, tortuous or aneurysmal vessels. While beyond the scope of this present study, a morphometry study is being published separately to address some triage strategies for the critical issue of stent placement when fluoroscopy is not available.

In conclusion, a retrievable, three-tier Rescue stent appears to reduce hemorrhage while preserving visceral flow, improving hemodynamics and preventing spinal cord ischemia. Although many logistical factors remain to be resolved, this study suggests that the Rescue stent represents an improvement over AO for the acute management of exsanguinating torso hemorrhage of the aorta. More generally, this study further suggests that retrievable stents may offer a rapid and flexible approach to temporize a variety of exsanguinating injuries.


C.G. participated in the literature search, data collection, data analysis, data interpretation, writing. M.E. participated in the study design, data collection, writing. Y.C. participated in the study design, data collection, writing. P.T. participated in the study design, data collection, data analysis, writing. W.W.C. participated in the study design. S.K.C. participated in the study design. A.J.D. participated in the data collection, data analysis. B.W.T. participated in the literature search, study design, data collection, data analysis, data interpretation, writing, critical revision.


The authors would like to thank Jenna Kuhn for technical support, Larry Fish, Ph.D. for statistical analysis, the McGowan Center for Preclinical Studies, as well as Erika Fish and Megan Rode for neurologic monitoring.


This work was supported by the Assistant Secretary of Defense for Health Affairs, through the Defense Medical Research and Development Program under Award No. W81XWH-16-2-0062. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the Department of Defense.

The authors declare no conflicts of interest.


1. Moore LJ, Brenner M, Kozar RA, Pasley J, Wade CE, Baraniuk MS, Scalea T, Holcomb JB. Implementation of resuscitative endovascular balloon occlusion of the aorta as an alternative to resuscitative thoracotomy for noncompressible truncal hemorrhage. J Trauma Acute Care Surg. 2015;79(4):523–530; discussion 30-2.
2. Scott DJ, Eliason JL, Villamaria C, Morrison JJ, Houston R, Spencer JR, Rasmussen TE. A novel fluoroscopy-free, resuscitative endovascular aortic balloon occlusion system in a model of hemorrhagic shock. J Trauma Acute Care Surg. 2013;75(1):122–128.
3. DuBose JJ, Scalea TM, Brenner M, et al. The AAST prospective aortic occlusion for resuscitation in trauma and acute care surgery (AORTA) registry: data on contemporary utilization and outcomes of aortic occlusion and resuscitative balloon occlusion of the aorta (REBOA). J Trauma Acute Care Surg. 2016;81(3):409–419.
4. Taner CB, Bulatao IG, Perry DK, Sibulesky L, Willingham DL, Kramer DJ, Nguyen JH. Asystole to cross-clamp period predicts development of biliary complications in liver transplantation using donation after cardiac death donors. Transpl Int. 2012;25(8):838–846.
5. Patel AR, Eggener SE. Warm ischemia less than 30 minutes is not necessarily safe during partial nephrectomy: every minute matters. Urol Oncol. 2011;29(6):826–828.
6. Thompson RH, Lane BR, Lohse CM, Leibovich BC, Fergany A, Frank I, Gill IS, Blute ML, Campbell SC. Every minute counts when the renal hilum is clamped during partial nephrectomy. Eur Urol. 2010;58(3):340–345.
7. Foley DP. Impact of donor warm ischemia time on outcomes after donation after cardiac death liver transplantation. Liver Transpl. 2014;20(5):509–511.
8. Davidson AJ, Russo RM, Reva VA, et al. The pitfalls of resuscitative endovascular balloon occlusion of the aorta: risk factors and mitigation strategies. J Trauma Acute Care Surg. 2018;84(1):192–202.
9. Long KN, Houston R, Watson JD, Morrison JJ, Rasmussen TE, Propper BW, Arthurs ZM. Functional outcome after resuscitative endovascular balloon occlusion of the aorta of the proximal and distal thoracic aorta in a swine model of controlled hemorrhage. Ann Vasc. 2015;29(1):114–121.
10. Markov NP, Percival TJ, Morrison JJ, Ross JD, Scott DJ, Spencer JR, Rasmussen TE. Physiologic tolerance of descending thoracic aortic balloon occlusion in a swine model of hemorrhagic shock. Surgery. 2013;153(6):848–856.
11. Chun Y, Cho SK, Clark WC, Wagner WR, Gu X, Tevar AD, McEnaney RM, Tillman BW. A retrievable rescue stent graft and radiofrequency positioning for rapid control of noncompressible hemorrhage. J Trauma Acute Care Surg. 2017;83(2):249–255.
12. Go C, Chun Y, Kuhn J, Chen Y, Cho SK, Clark WC, Tillman BW. Damage Control of Caval Injuries in a Porcine Model using a Retrievable Rescue stent. J Vasc Surg Venous Lymphat Disord. 2018;6(5):646–656.
13. Chen Y, Tillman B, Go C, Cho SK, Clark WW, Hur TB, Ding Y, Chun Y. A novel customizable stent graft that contains a stretchable ePTFE with a laser-welded nitinol stent. J Biomed Mater Res B Appl Biomater. 2019;107(4):911–923.
14. Bharati S, Levine M, Huang SK, Handler B, Parr GV, Bauernfeind R, Lev M. The conduction system of the swine heart. Chest. 1991;100(1):207–212.
15. Swindle MM. ANESTHESIA & ANALGESIA IN SWINE Available from: Accessed 06-12-2019.
16. Strauch JT, Lauten A, Spielvogel D, Rinke S, Zhang N, Weisz D, Bodian CA, Griepp RB. Mild hypothermia protects the spinal cord from ischemic injury in a chronic porcine model. Eur J Cardiothorac Surg. 2004;25(5):708–715.
17. Williams TK, Tibbits EM, Hoareau GL, Simon MA, Davidson AJ, DeSoucy ES, Faulconer ER, Grayson JK, Neff LP, Johnson MA. Endovascular variable aortic control (EVAC) versus resuscitative endovascular balloon occlusion of the aorta (REBOA) in a swine model of hemorrhage and ischemia reperfusion injury. J Trauma Acute Care Surg. 2018;85(3):519–526.
18. Pagel PS, Fu JL, Damask MC, Davis RF, Samuelson PN, Howie MB, Warltier DC. Desflurane and isoflurane produce similar alterations in systemic and pulmonary hemodynamics and arterial oxygenation in patients undergoing one-lung ventilation during thoracotomy. Anesth Analg. 1998;87(4):800–807.
19. Clinical Pathology: Normal Chemistry Values 2014. Available from: Accessed 05-10-2019.
20. White JM, Cannon JW, Stannard A, Markov NP, Spencer JR, Rasmussen TE. Endovascular balloon occlusion of the aorta is superior to resuscitative thoracotomy with aortic clamping in a porcine model of hemorrhagic shock. Surgery. 2011;150(3):400–409.
21. Coselli JS, Conklin LD, LeMaire SA. Thoracoabdominal aortic aneurysm repair: review and update of current strategies. Ann Thorac Surg. 2002;74(5):S1881–S1884; discussion S92–8.
22. Russo RM, Neff LP, Lamb CM, Cannon JW, Galante JM, Clement NF, Grayson JK, Williams TK. Partial resuscitative endovascular balloon occlusion of the aorta in swine model of hemorrhagic shock. J Am Coll Surg. 2016;223(2):359–368.
23. Elsisy M, Tillman B, Go C, Kuhn J, Cho S, Clark W, Park J, Chun Y. Comprehensive Assessment of Mechanical Behavior of an Extremely Long Stent Graft to Control Hemorrhage in Torso. J Biomed Mater Res B. 2020;in press.
24. Onat S, Ulku R, Avci A, Ates G, Ozcelik C. Urgent thoracotomy for penetrating chest trauma: analysis of 158 patients of a single center. Injury. 2011;42(9):900–904.

Torso hemorrhage; noncompressible; retrievable stent; balloon occlusion; porcine

Supplemental Digital Content

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.