Role of Intercostal Artery Embolization in Management of Traumatic Hemothorax.

Intercostal artery bleeding from trauma can result in potentially fatal massive hemothorax. Traumatic hemothorax has traditionally been treated with tube thoracostomy, video-assisted thoracoscopic surgery, or thoracotomy. Transcatheter arterial embolization (TAE), a well-established treatment option for a variety of acute hemorrhage is not widely practiced for the management of traumatic hemothorax. We present 2 cases of delayed massive hemothorax following chest trauma which were successfully managed by transarterial embolization of intercostal arteries. The published studies are reviewed and a systematic approach to the selection of patients for TAE versus emergency thoracotomy is proposed. Copyright:

INTRODUCTION

Massive hemothorax from intercostal artery bleeding is an infrequent but potentially life-threatening consequence of chest trauma. Exploratory thoracotomy has long been regarded as the therapeutic gold standard for further management of massive hemothorax.[1] Transcatheter arterial embolization (TAE), an effective and less invasive treatment alternative to surgery for hemorrhage control, is an established endovascular therapeutic option for a variety of indications including iatrogenic hemothorax and control of hemorrhage in abdominopelvic trauma. However, despite being described over three decades, the role of TAE in the management of traumatic hemothorax is not as well defined. We present two cases of successful intercostal artery embolization performed for delayed hemothorax secondary to trauma, and the review of the available literature. We discuss the indications, technical aspects, and clinical efficacy of TAE with regards to the management of traumatic hemothorax.

CASE REPORTS

Case 1

A 63-year-old male patient presented to the emergency department (ED), sustaining stab wound to the right lower chest wall. Upon arrival, he was mildly tachypneic with blood pressure of 146/100 mm Hg, heart rate of 90 beats/min, and oxygen saturation of 99%. Physical examination revealed a 1.5 cm stab wound and small chest wall hematoma. The chest radiograph showed no rib fracture nor any significant effusion. Contrast-enhanced computed tomography (CECT) demonstrated a small 7 mm pseudoaneurysm likely arising from the 6th anterior intercostal artery [Figure 1a and b]. The patient was initially planned for percutaneous thrombin injection of the intercostal artery pseudoaneurysm to prevent further expansion of the chest wall hematoma. However, upon arrival to the angiography suite, the pseudoaneurysm had spontaneously thrombosed [Figure 1c and d]. A pressure bandage was applied, and the patient was managed conservatively in the intensive care unit. After 24 h, the patient complained of severe right chest pain and became tachycardic. Hemoglobin showed a significant drop from 14.2 to 9.9 g/dl. The repeat chest radiograph revealed moderate right hemothorax [Figure 1e], which was subsequently confirmed on a repeat CECT [Figure 1f]. A chest tube was inserted and 700 ml of blood was initially drained. Within 24 h, the amount of blood drained from the chest tube increased to 1.5 L, and a decision was made to proceed with angiography and trans-arterial embolization.

Figure 1

Stabbing injury. Arterial (a) and delayed phase (b) images from a CT demonstrate small pseudoaneurysm (white arrow) arising from the anterior intercostal artery. The pseudoaneurysm was also localized on ultrasound (c), however, after 2 h (d) it had spontaneously thrombosed. Follow up chest radiograph (e) and CT (f) obtained 24 h after the injury, show large right hemothorax. Selective angiography (g) of the anterior intercostal artery shows small focus of contrast extravasation. Angiographic image (h) following embolization with mixture of n butyl cyanoacrylate and lipiodol

Through the right femoral access, the right internal mammary artery was cannulated. Selective arteriography showed contrast extravasation from the right 6th anterior intercostal artery [Figure 1g]. Superselective embolization of the 6th intercostal artery was performed with n-butyl-2-cyanoacrylate (NBCA) glue (Histoacryl®) [Figure 1h]. The intercostal arteries above (5th) and below (7th) this level were embolized with gelfoam slurry. Postprocedure, the drain output reduced to 70 ml (at 3-h postprocedure). The hemoglobin level stabilized at 12.0 g/dl with two units of packed red blood cell transfusion. Follow-up radiograph 48 h later showed complete resolution of hemothorax. The chest drain was removed, and he was discharged on the third postprocedural day.

Case 2

A 90-year-old male patient presented to ED with left chest pain and tenderness following an accidental fall at home. He had a past medical history of hypertension, diabetes mellitus, ischemic heart disease, old stroke, epilepsy, and chronic thrombocytopenic purpura. His blood pressure was 136/84 mm Hg, with a heart rate of 84 beats/min. The chest radiograph did not show any significant pleural effusion or pneumothorax [Figure 2a]. He was treated symptomatically and discharged home after overnight observation. He presented 9 days later to ED with severe dyspnea. The oxygen saturation was 66% and repeat radiograph showed complete whiteout of the left hemithorax [Figure 2b]. A tube thoracostomy was performed and 1500 ml of blood was drained. He became drowsy and blood gas analysis revealed severe metabolic acidosis. Hemoglobin dropped from a baseline of 12 g/dl to 7.3 g/dl. He was intubated and a contrast-enhanced CT study was performed.

Figure 2

Delayed hemothorax following fall. Initial chest radiograph (a) did not show any pleural fluid. Chest radiograph (b) obtained 9 days later, shows almost complete white out of the left hemithorax from large hemothorax. Contrast-enhanced CT images (c and d) demonstrate contrast extravasation from the posterior intercostal artery and multiple displaced rib fractures. Selective angiography images before (e) and after (f) embolization with n-butyl cyanoacrylate -lipiodol mixture

Urgent CECT demonstrated displaced left 9th to 11th rib fractures and identified contrast extravasation from left 10th posterior intercostal artery as cause of massive hemothorax [Figure 2c and d]. In view of the underlying comorbidities and CECT findings, a primary TAE of the culprit intercostal artery was performed. Via right femoral access, the origin of the left 10th posterior intercostal artery was cannulated with a reverse curve catheter and a 2.2F microcatheter was advanced into the intercostal artery. Selective arteriography of the intercostal artery re-demonstrated the contrast extravasation [Figure 2e]. Super-selective embolization was performed using NBCA glue (Histoacryl®) [Figure 2f]. The drain output gradually reduced, and no additional procedures were required for bleeding control. The drain was removed on day 7. His hospital stay was complicated by urinary tract infection and renal impairment. On day 15, the patient was transferred to a step-down rehabilitation center for further care.

DISCUSSION

The chest radiograph is often inadequate for the evaluation of hemothorax. CECT should be performed if the radiograph demonstrates displaced rib fractures and the initial chest drainage confirms hemothorax. Hagiwara et al. found that the patients with >200 ml/h of hemothorax and greater parallel displacement of rib fractures on chest radiograph demonstrated contrast extravasation on CECT.[2] The threshold for CECT examination should be even lower in cases of penetrating injuries and elderly patients with co-morbidities. In patients with significantly displaced rib fractures and persistent bleeding, where a definitive diagnosis of intercostal artery injury represented by contrast extravasation or pseudoaneurysm on CECT can be made, TAE should be considered primarily.[2] In contrast, if the CECT demonstrates parenchymal lacerations, aortic rupture, or pulmonary arterial injuries to be the source of hemothorax, TAE may unnecessarily delay the thoracotomy. Besides imaging findings, the clinical symptoms and laboratory findings play an important role in decision-making. TAE should be considered in settings of significant hemoglobin drop (>2 g/dl) or signs of shock with hypotension or tachycardia.[3] Persistent bleeding from posterior intercostal arteries close to the spine can be difficult to control at thoracotomy. The limited exposure at the posterior rib space makes hemostasis challenging.[4] Transarterial embolization provides a minimally invasive adjunct to surgical methods in such difficult-to-control intrathoracic hemorrhage.[4] A systematic approach to the selection of patients for TAE is outlined in the algorithm [Figure 3].

Figure 3

Previously published case reports and studies on transarterial embolization of intercostal artery(ies) for traumatic hemothorax

A clear understanding of the anatomy of anterior and posterior intercostal arteries and collateral circulation is crucial to technical success.[5] Wherever feasible “frontdoor–backdoor” technique of embolization should be applied to achieve adequate hemostasis.[6] Only front door embolization with gelfoam or particles may result in pseudoaneurysm at the arterial injury site and manifest as delayed hemothorax.[7] Various embolic materials as gelfoam, Polyvinyl-alcohol particles, NBCA glue, and coils have been used. No one embolic agent has been shown to be superior to the rest. No major immediate procedure-related complication has been reported. Nonetheless, inadvertent spinal artery embolization is always a concern during the embolization of the intercostal arteries. When liquid or particulate embolic materials are to be used in the posterior intercostal arteries, the angiograms should be scrutinized for opacification of the artery of Adamkiewicz [Figure 4]. Microcoils are the most used embolic agents and should be favored if the angiograms demonstrate classic hairpin appearance of spinal artery feeders on angiography and the bleeding source is close to the origin of the radicular arteries.[389]

Figure 4

Ar tery of Adamkiewicz. Posterior intercostal ar tery angiogram (from case 2) demonstrates the characteristic hairpin turn (arrow) at its junction with the anterior spinal artery

The largest case series that described the role of TAE in massive hemothorax reported a technical success rate of 87%.[10] Most other studies with five or more patients have also reported similar results, with higher success rate in traumatic cases. The published case studies are summarized in Table 1. Tamburini et al. found that patients with a single bleeding artery showed a better outcome in terms of bleeding control compared to two or more arteries.[10] Despite the high technical success rate, up to 60% of patients would require additional surgical intervention to remove the clotted hemothorax or manage incontrollable bleeding. Most of these additional surgical procedures are nonurgent video-assisted thoracoscopic surgery (VATS) for retained hemothorax.[10] The average rate of clinical failure (defined as patients requiring further surgeries i.e., VATS or thoracotomy) after TAE to control the bleeding is about 20%.[6] The possible predictors of clinical failure are incomplete embolization (e.g. two or more intercostal arterial injuries),[10] presence of collateral pathways (e.g. internal mammary arteries, inferior phrenic arteries, musculophrenic arteries, etc.),[519] abnormal clotting process (due to severe initial bleeding, chronic liver disease) and failure to recognize alternate sources of bleeding like pulmonary parenchymal lacerations, pulmonary arterial injuries[17] or diaphragmatic rupture.[18]

Table 1

Previously published case reports and studies on transarterial embolization of intercostal artery (ies) for traumatic hemothorax

Author, years	n	Age/mean age	Blunt or penetrating injury	Injured vessel	CT findings	Embolic agents	Technical failure	Clinical failure	Comments
Carrillo et al., 1998[11]	8	ND	6 blunt 2 penetrating	5 ICAs 3 IMAs	ND	Gelfoam	None	None	2 deaths due to severe cerebral injuries 3 TAEs after exploratory thoracotomy
Matsumoto et al., 1998 [12]	1	21	Blunt	Multiple ICAs	ND	Gelfoam	None	None	Embolization after thoracotomy and surgical treatment of diaphragmatic rupture, shattered spleen, and PL
Kazuo et al., 1999 [13]	1	77	Blunt	Single ICA	CE-yes	Gelfoam	None	None	Delayed massive hemothorax while positioning patient in the operating room
Kessel et al., 2004 [14]	1	85	Blunt	Single ICA	CE-yes	Embospheres	None	None	Massive hemothorax, initial drainage 1500 ml. First-line TAE was performed. Presence of significant comorbidities
Sekino et al., 2005 [5]	1	36	Penetrating	Single ICA	CE-yes	Microcoils	None	None	Delayed hemothorax which occurred 9 days after the initial presentation
Hagiwara et al., 2008 [2]	5	39	Blunt	Multiple ICAs	CE-yes PL-none	Gelfoam and microcoils	None	None	TAE as the first line for patients with chest drainage>200 ml/h, CE on CT and ICA as the source of bleeding
Hagiwara and Iwamoto, 2009 [15]	1	25	Blunt	Multiple ICAs	CE-yes PL-none	Gelfoam	None	None	Burst fractures of lower thoracic vertebrae with bilateral ICA injuries which are close to the spine
Chemelli et al., 2009 [3]	11	53	Blunt	Multiple ICAs	CE, pseudoaneurysm-yes PL-none	Coils, PVA particles, and NBCA	None	Yes, 1	Series comparing TAE in blunt traumatic and iatrogenic ICA injuries. 90.9% primary technical success in the traumatic injury group
Behera et al., 2014 [16]	1	52	Penetrating	Single ICA	Pseudoaneurysm	Microcoils	None	Yes	Required additional VATS for retained hemothorax 3 weeks after the injury
Nemoto et al., 2014 [7]	1	67	Blunt	Multiple ICAs	CE-yes Pseudoaneurysm	Initial-gelfoam Repeat-microcoils	None	Yes	Required repeated TAE with microcoils for a ICA pseudoaneurysm on day 5 of the presentation
Stampfl et al., 2014 [6]	2	ND	1 blunt 1 penetrating	ND	ND	Glue and microcoils	None	ND	Series of 19 cases, mostly iatrogenic with clinical success of 80%
Izaaryene et al., 2017 [17]	1	31	Penetrating	Single ICA	CE-yes PL-yes	Microcoils	None	Yes	Required additional thoracotomy and lung parenchymal wound was repaired
Izaaryene et al., 2017 [17]	1	61	Blunt	Multiple ICAs	CE-yes, multiple	Onyx	None	Yes	Required emergency thoracotomy and pulmonary artery injury was detected intraoperatively
Moore et al., 2018 [4]	1	19	Gunshot wound	Multiple ICAs	ND	Microcoils	None	None	Adjunct TAE after thoracotomy. Source of bleeding was close to the posterior rib space with limited exposure on thoracotomy
Ota et al., 2018 [18]	1	83	Blunt	Single ICA	CE-yes	ND	None	Yes	Required emergency thoracotomy 4 h later. Found additional diaphragmatic injury
Tamburini et al., 2019 [10]	18	Mean 70	ND	Single and multiple ICAs	CE-yes, 87% of cases	PVA, coils, and gelfoam	13%	Yes, 2 cases	The largest published series. 2 patients required exploratory thoracotomy. Multiple ICAs and amount of blood transfused were predictors of poor outcome

n: Sample size, CT: Computed tomography, ICAs: Intercostal arteries, IMAs: Internal mammary arteries, CE: Contrast extravasation, PL: Pulmonary laceration, TAEs: Transarterial embolizations, ND: Not described, NBCA: n-butyl-2-cyanoacrylate, VATS: Video-assisted thoracoscopic surgery, PVA: Polyvinyl-alcohol

CONCLUSION

TAE provides an effective and less invasive therapeutic option in the management of traumatic hemothorax in a selected group of patients. Despite a remarkably high technical success and almost no major complication, the decision for TAE versus surgery as a primary treatment is dependent on the identification of the origin of bleeding. Primary TAE is recommended for isolated intercostal artery injuries; while emergency thoracotomy should be performed in unstable patients with pulmonary lacerations, hilar vessel injuries, or diaphragmatic rupture. Adjunct TAE should be considered in cases whereby surgical control of hemorrhage is expected to be tedious and difficult.

Research quality and ethics statement

The authors followed applicable EQUATOR Network () guidelines, notably the CARE guideline, during the conduct of this report.

Declaration of patient consent

The authors certify that they have obtained allappropriate patient consent forms. In the form the patients has given their consent for their images and other clinical information to be reported in thejournal. The patients understand that their names and initials will not be published and due efforts will bemade to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.