Abstract
Purpose
Uncontrolled pelvic hemorrhage from trauma is associated with mortality rates above 30%. The ability of an intervention to reduce blood loss from pelvic trauma is paramount to its success. The objective of this study was to determine if computed tomography volumetric analysis could be used to quantify blood loss in a porcine endovascular pelvic hemorrhage model.
Materials and Methods
Yorkshire swine under general anesthesia underwent balloon dilation and rupture of the profunda femoris artery, which was confirmed by digital subtraction angiography. Computed tomography angiography and postprocessing segmentation were performed to quantify pelvic hemorrhage volume at 5 and 30 minutes after injury. Continuous hemodynamic and iliofemoral flow data were obtained. Baseline and postinjury hemoglobin, hematocrit and lactate were collected.
Results
Of 6 animals enrolled, 5 survived the 30-minute post-injury period. One animal died at 15 minutes. Median volume of pelvic hemorrhage was 141±106 cm3 at 5 minutes and 302±79 cm3 at 30 minutes with a 114% median increase in hematoma volume over 25 minutes (p=0.040). There was a significant decrease in mean arterial pressure (107 to 71 mm Hg, p=0.030) and iliofemoral flow (561 to 122 mL/min, p=0.014) at 30 minutes postinjury, but no significant changes in hemoglobin, hematocrit, or heart rate.
Conclusion
Computed tomography volumetric analysis can be used to quantify rate and volume of blood loss in a porcine endovascular pelvic hemorrhage model. Future studies can incorporate this approach when evaluating the effect of hemorrhage control interventions associated with pelvic fractures.
Keywords
Introduction
Noncompressible hemorrhage constitutes the most common cause of potentially preventable death in the civilian and military setting. 1 Hemodynamically unstable pelvic fractures are a particularly lethal form of noncompressible hemorrhage, with mortality rates above 30%.2,3 These injuries present a management dilemma due to multiple sources of potential blood loss and require timely and multidisciplinary care to improve chances of survival. Pelvic fracture patterns with the highest associated mortality risk are anteroposterior compression type fractures because they are often associated with arterial hemorrhage. 4
The past decade has seen significant improvements in pelvic arterial hemorrhage control techniques such as the development of resuscitative endovascular balloon occlusion of the aorta (REBOA) and endovascular angioembolization.5,6 Despite the development of these adjuncts to arrest pelvic arterial hemorrhage, their efficacy in reducing pelvic hemorrhage and increasing survival remain unclear. In human contexts, the use of multiple hemorrhage control interventions at once clouds the efficacy of each individual intervention. In order to adequately characterize the impact of each intervention in reducing pelvic hemorrhage, clinically relevant animal models are of paramount importance.
Previous animal models of pelvic hemorrhage have utilized an open approach to injury creation, which results in uncontrolled bleeding and disruption of tissue planes.7,8 This leads to rapid exsanguination mimicking the hemodynamic instability associated with severe pelvic fractures but makes quantification of blood loss challenging. Due to the open technique, it does not accurately simulate retroperitoneal hemorrhage. The ideal model would maintain a closed space while achieving retroperitoneal hemorrhage and allow for accurate quantification of blood loss.
The aim of this study was to produce a clinically relevant endovascular model of pelvic arterial hemorrhage where the rate and volume can be quantified in order to serve as a platform for pelvic hemorrhage control research.
Materials and Methods
Overview
A total of 6 healthy, adult, male Yorkshire swine weighing between 50 and 70 kg were enrolled in experiments that consisted of 4 phases: animal preparation, endovascular access and instrumentation, endovascular injury, and postinjury observation. Baseline imaging, complete laboratory analysis, and vital signs were compared with postinjury measures. Following the end of the protocol the animals were euthanized. The experiments were conducted at the University of Maryland School of Medicine hybrid animal operating room. This study was approved by the Institutional Animal Care and Use Committee of the host facility (protocol no. 0220014) and all procedures were accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International under the supervision of the institution’s veterinarian support staff.
Fluoroscopy (OEC 9800 plus, General Electric, Boston, MA, USA) was used to facilitate catheter placement and acquire digital subtraction angiography (DSA). Where pelvic angiography was performed, the pigtail catheter (Omniflush, AngioDynamics, Inc, Latham, NY, USA) was used to deliver iodinated contrast at a rate of 5 mL/s for 10 mL, with images acquired at 15 frames per second.
Computed tomography angiography (CTA) was also used throughout the experiments. A 16-slice portable machine (OmniTom, Samsung Neurologica Corporation, Danvers, MA, USA) used a helical acquisition, with a slice thickness of 0.625 mm, during an intra-arterial contrast injection using the pigtail catheter at a rate 2 mL/s for 20 mL. Images were stored in a patient archive and communication system for postprocessing.
Animal Preparation
All animals were acclimatized for 48 hours prior to experimental use. The night before surgery animals were fasted with adequate access to water. Each pig was sedated using intramuscular injection of telazol (4–5 mg/kg) and xylazine (2 mg/kg). Following sedation, the animal was placed in dorsal recumbency on a radiolucent operating table. Endotracheal intubation was performed and each pig was ventilated using positive pressure. Anesthesia was maintained with isoflurane (1%–3%) with 30% oxygen using a volume controlled mode on a Fabius GS ventilator (Dräger, Lübeck, Germany). Various physiological monitors, including blood pressure (BP) (Transonic Systems Inc, Ithaca, NY, USA), electrocardiogram (ECG)/heart rate (HR), and oxygen saturation monitors were applied to the animal.
Endovascular Access and Instrumentation
Arterial and venous access was then obtained by ultrasound-guided Seldinger technique: Under ultrasound guidance, a 21-gauge needle was used to access the target vessel, a 0.018-inch wire was inserted to allow insertion of a 5-Fr micropuncture sheath (Cook Group, Bloomington, IN, USA), and then the 0.018-inch wire was replaced with a 0.035-inch wire to enable replacement of the 5-Fr sheath for a 7-Fr sheath (Terumo Medical Corporation, Elkton, MD, USA). The left carotid artery was accessed via a 7-Fr sheath for insertion of an endovascular balloon (12.0 × 40.0 mm Mustang, Boston Scientific, Marlborough, MA, USA) for injury. The right carotid artery was accessed for insertion of the pigtail catheter to facilitate DSA. A 5-Fr Destination sheath (Terumo Medical Corporation, Elkton, MD, USA) was placed in the left brachial artery for the facilitation of lab draws and the insertion of an aortic root pressure probe.
A laparotomy was performed to facilitate the dissection of the proximal common iliac artery and the placement of a 6-mm ultrasonic flow probe (Transonic Systems Inc, Ithaca, NY, USA) in order to measure vessel flow rate. Though this is not essential to the model, it did allow us to gather additional data in order to further characterize the model. Care was taken to place this flow probe far away from the planned injury site and to minimize dissection as not to significantly disrupt the retroperitoneal space. The retroperitoneum was then tightly closed with running suture. A cystostomy was also performed allowing for drainage of the urinary bladder with a 16-Fr Foley catheter.
Following completion of instrumentation, baseline DSA and CTA were performed as described above. Baseline blood samples were collected for hematology (i-STAT, Abbott Laboratories, Chicago, IL, USA), which included hemoglobin and hematocrit concentration (Hg/Hct), and arterial blood gas (ABL800 Flex, Radiometer America Inc, Brea, CA, USA), which included pH and lactate.
Endovascular Injury
Swine have a different arterial anatomy from human patients (Figure 1), with the terminal aorta trifurcating into a left and right common iliac artery and a midline common trunk that leads to the internal iliac arteries. The common iliac arteries branch in the swine haunch, with a vessel analogous to the human profunda femoris supplying the muscle of the thigh. This vessel is still located in the retroperitoneal space due to the quadrupedal nature of the pig. We selected this vessel for endovascular balloon rupture to simulate a source of retroperitoneal hemorrhage. The average size of this vessel is 6 mm.
Digital subtraction angiography (A) and computed tomography angiography (CTA) (B) demonstrating porcine iliac vascular anatomy.
The injury commenced with the administration of 25,000 units of heparin intravenously to simulate a bleeding diathesis. A 0.035-inch angled stiff Glidewire (Terumo Medical Corporation, Elkton, MD, USA) was inserted through the 7-Fr carotid sheath and advanced under fluoroscopic guidance into the left common iliac artery branch vessel. Once in position, the endovascular angioplasty balloon on a 135 cm catheter was advanced over the wire into the vessel. The balloon was then inflated using an endoflater to the nominal pressure rupturing the vessel using a purely endovascular technique to create the injury.
Postinjury Observation
Following injury creation, the disrupted vessel was evaluated by both DSA and CTA within 5 minutes. Blood sampling was also obtained at this time and animals were then observed for a period of 30 minutes followed by a second set of images and blood draws. Animals were then euthanized using an intravenous injection of potassium chloride.
Continuous physiological records of BP, HR, end tidal carbon dioxide (ETCO2), core temperature (rectal), and urine output (UOP) were made throughout the protocol. Data were collected using ADInstruments (Sydney, Australia) probes and the associated PowerLab system and recorded using their integrated life science data acquisition software LabChart Version 8.
Study Endpoints
The primary outcome was the creation of an extravasating pelvic vascular injury with hematoma formation in the retroperitoneal space. DSA was used to confirm successful vessel rupture, whereas CTA permitted volumetric assessment of hemorrhage. CTA images underwent postprocessing using Horos, a free, open source medical image viewer. 9
CTA was used to confirm the anatomical location and size of the injury while adjunctive data from the common iliac flow probe relayed flow rate. Secondary outcomes included hemodynamic and metabolic indices including BP, HR, Hg/Hct, iliac flow rate (mL/min), and measures of acidosis and tissue ischemia (pH, lactate).
Numerical data were collected and stored in a Microsoft Excel (Redmond, WA, USA) file and analyzed using GraphPad Prism v8.0 (GraphPad Software Inc, La Jolla, CA, USA). Numerical data are presented as mean with standard deviations for parametric values and as median with interquartile (IQR) range for nonparametric values. The paired t test was used to compare non-parametric data and the Wilcoxon signed rank test was used to compare paired nonparametric data. A p value of less than 0.05 was considered significant. CTA images were analyzed using Horos during which the retroperitoneal hemorrhage was volumetrically assessed with the region of interest (ROI) functionality. Using the uninjured side as a visual control, the manual ROI tool was used to compute volume in the axial plane for all regions that fell within Hounsfield unit values consistent with blood. Three-dimensional volumetric assessment was performed using Horos and overlapping regions were merged to provide total volume of pelvic hematoma.
Results
Baseline Data
In total, 6 male swine with a mean weight of 57.9±3.0 kg were enrolled in model development. There were no differences between baseline physiologic and biochemical measurements between animals. Baseline imaging showed no variations in vasculature and no iatrogenic hemorrhage. Mean physiology and baseline laboratory characteristics for all 6 animals are shown in Table 1.
Hemodynamic, Flow, and Laboratory Characteristics Pre- and Postinjury. a
Abbreviation: MAP, mean arterial pressure.
Reported as mean with standard deviation in parenthesis.
At 30 minutes postinjury.
Hemodynamic Data
Out of 6 animals used, 1 animal died 15 minutes after injury. Due to this, 30-minute postinjury data were not obtained in this animal. Following balloon disruption, the HR did not change significantly, but the mean arterial pressure (MAP) decreased (107 to 71 mm Hg, p=0.030). Preinjury iliofemoral mean flow rate was 561±170 mL/min, which decreased to a mean rate of 122±72 mL/min after injury (p=0.014). Although there was a trend toward decrease in hemoglobin and hematocrit, this did not reach statistical significance. Markers of shock postinjury are reported in Table 1. Continuous hemodynamic data over the postinjury period demonstrated a decrease in heart rate, iliofemoral flow, and BP (Figure 2).
Continuous hemodynamic monitoring demonstrating changes in heart rate (A), iliofemoral flow (B), and blood pressure (C). Time of injury is marked by the vertical hashed line.
Injury and Hematoma Quantification Data
Iliac and profunda arteries were successfully accessed through an endovascular approach with effective balloon disruption in all cases. DSA demonstrating the injury is shown in Figure 3.
Digital subtraction angiography demonstrating normal iliac vessels (A, left), followed by balloon rupture and resultant contrast extravasation from the profunda femoris artery (A, right). Subsequent open exploration of the retroperitoneum demonstrating large hematoma without intraperitoneal violation (B).
Open retroperitoneal exploration demonstrated no gross intraperitoneal violation. CTA was obtained in 18 instances, all of which were of sufficient diagnostic quality to allow postprocessing and segmentation. At 5 minutes, the median hematoma volume was 141±106 cm3. At 30 minutes, the median hematoma volume was 302±79 cm3. This resulted in a 114% median increase in hematoma volume over 25 minutes (p=0.04). Bleeding rate was calculated by dividing median hematoma volume by the number of minutes elapsed thus far. Bleeding rates were found to be 28.2 mL/min at 5 minutes and 10.1 mL/min at 30 minutes. Changes in retroperitoneal hematoma volume at 5 minutes and 30 minutes for each Yorkshire swine are demonstrated in Figure 4. Postprocessing of the CTA images using segmentation and 3-dimensional reconstruction are shown in Figure 5.
Calculated pelvic hemorrhage volume at 5 and 30 minutes after injury.
Thirty-minute postinjury computed tomography angiography (CTA) (upper left) and postprocessing using segmentation (upper right) and 3-dimensional reconstruction (lower panels).
Discussion
This study presents a method for quantification of hemorrhage with CT volumetric analysis in an endovascular porcine model of pelvic arterial injury. It employs a simple approach for injury creation using endovascular balloon dilation and rupture of the profunda femoris artery, which lies in the retroperitoneal space. Decrease in MAP and iliofemoral flow are representative of the degree of hemorrhage induced by the injury. Postprocessing segmentation demonstrate an increase in pelvic hemorrhage volume from 5 to 30 minutes with a resultant decrease in bleeding rate along the same time interval. These findings validate the use of CT volumetric analysis as a method for quantifying blood loss in a porcine model of pelvic hemorrhage.
There has been an exponential growth in the use of CTA for trauma, ranging from diagnosis of neck, chest, and extremity vascular injury to management of abdominal solid organ injury.10,11 CT imaging is often the next step in management of the hemodynamically stable trauma patient following initial assessment in the trauma bay. 12 CTA is rapid and as sensitive and specific as DSA for the diagnosis of most vascular injuries, and when properly utilized can avoid an unnecessary invasive procedure. 13 The ubiquity in use of CT imaging and CTA for trauma evaluation creates the opportunity to maximize the technology as a tool for information gathering to guide management.
Segmentation refers to the process of dividing medical images into sections with similar properties such as color, contrast enhancement, or texture. 14 This can be used for a range of purposes, such as quantifying growth of a tumor over time or assessing liver volume in patients undergoing hepatic resection. 15 The use of segmentation for CT volumetric analysis has been validated in several settings, ranging from quantification of pleural effusion volumes in patients undergoing thoracentesis to calculating abdominal cavity fluid volumes to predict need for decompressive laparotomy in patients with intra-abdominal hypertension.16,17
CT volumetric analysis has emerged as a valuable tool that can be used to quantify degree of hemorrhage in patients with abdominopelvic trauma to guide the need for hemorrhage-control interventions. In 2003, a study by Blackmore et al 18 used a manual technique to calculate hematoma volumes from CT imaging in the setting of pelvic fractures, showing increased need for intervention with greater volume of hemorrhage. A recent study by Dreizin et al 19 advanced this further by using a semiautomated technique of CT segmentation to evaluate the volume of pelvic hemorrhage in the setting of unstable pelvic ring fractures. Using this technique, they demonstrated that pelvic hematoma volumes greater than 433 mL had a positive predictive value of requiring angioembolization of 100%. As CT postprocessing technology continues to advance, the rapid availability of information relating to rate and quantity of hemorrhage in the setting of torso trauma could aid decision making.
The current study demonstrates how CT volumetric analysis could be translated to the laboratory and provide an opportunity to evaluate various pelvic hemorrhage control interventions in a monitored setting. There are several advantages to using the approach described. First, the technique of injury creation is simple, easily replicated, and creates no tissue trauma. This ensures that the segmentation process will not be influenced by tissue edema or hemorrhage associated with the dissection.
Second, the use of the profunda femoris artery ensures that the hemorrhage remains in the retroperitoneum. This is representative of pelvic hemorrhage as it occurs in the human population following pelvic trauma during which there is rarely any intraperitoneal bleeding. Open techniques to injury creation can result in an element of intraperitoneal blood loss or external bleeding, which decreases the accuracy of pelvic hemorrhage calculations as not all the blood loss is accounted for.
Having validated the approach to pelvic hemorrhage creation and subsequent segmentation, the next step is to evaluate use of endovascular techniques for pelvic hemorrhage control. The approach described in the current study will allow for a detailed assessment of changes in rate and volume of blood loss when pelvic angioembolization techniques are used. An animal study allows for the evaluation of a single intervention at one time, an approach that cannot be employed easily in a human setting where multiple pelvic hemorrhage control interventions are used at once.
There are some limitations to this study that must be mentioned. The sample size is small, which affects the robustness of the findings as evidenced by the lack of statistical significance in several hemodynamic parameters from pre to post injury creation. This could likely be explained by the variability in the amount of blood loss from the 5-minute to 30-minute postinjury period, with some animals bleeding more than others. Although this is representative of the variability in a human population, the degree of hemorrhagic shock experienced by the animals subsequently was also variable.
Second, as the process of volumetric analysis was not completely automated, the authors cannot determine the extent of human error without a comparison group. The ROI that fell within the range of Hounsfield units consistent with hemorrhage was manually measured, followed by automated summation and removal of overlapping areas using Horos software. The range used is an estimate of density consistent with blood, and it is possible that some areas of actual blood loss were not included and areas that may not have been blood, but other types of fluids, were included.
Third, it is important to mention that this model of pelvic hemorrhage only accounts for arterial blood loss associated with pelvic fractures. Pelvic fractures often have a venous component of blood loss that involves the sacral venous plexus, and a bony component related to the fractures themselves. This model can only be used to study interventions related to control of arterial hemorrhage and not other interventions commonly used such as preperitoneal packing and external fixation that are employed for other types of bleeding.
Conclusion
The current study presents a simple, easily reproducible, and representative animal model of retroperitoneal, arterial hemorrhage using endovascular balloon rupture of the profunda femoris artery. The process of CT volumetric analysis has been validated in several previous human studies and was successfully performed in the current study involving 6 Yorkshire swine. It demonstrated the ability to calculate rate and volume of blood loss from 5 to 30 minutes postinjury, with associated decreases in MAP and iliofemoral flow. This model creates the opportunity for future studies of endovascular interventions to arrest bleeding from pelvic fracture–associated arterial hemorrhage.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by an industry sponsored grant from General Electric (Grant Number CCT 2020-13818).
