Covered stents for injuries of subclavian and axillary arteries☆

Material and methods 

Data were reviewed for all patients seen at the University of Tennessee Medical Center in Knoxville (UTMCK) with blunt or penetrating injuries (including iatrogenic) to the subclavian and axillary arteries between January 1, 1996 and July 30, 2002. Demographic data, mechanism of trauma, blood pressure, concomitant injuries at presentation, and treatment method were recorded. Blood loss, operative time, and wound-related complications were noted. Angiograms were examined to identify lesions amenable to endovascular treatment. Clinical evaluation and noninvasive examination with duplex ultrasound scanning were performed to evaluate graft patency after discharge. Length of recovery was not compared, because patients in both groups had multiple related injuries or medical comorbid conditions that primarily determined their length of stay. The Mann-Whitney rank-sum test and Fisher exact test were used for statistical evaluation. P < .05 was considered significant. Results are given as mean ± SEM.

Results 

Twenty-seven patients with injury to the subclavian or axillary artery were seen between January 1, 1995 and July 30, 2002 (Table I). Two patients died before intervention, 2 patients underwent nonoperative treatment, and 23 patients underwent angiography resulting in either surgical or endovascular repair (EVAR). Eleven patients had extensive, non-focal segment injury of the vessel, transection or occlusion, and this subset underwent open surgical repair. The remaining 12 patients had focal lesions amenable to endovascular therapy or open repair. Of these, five underwent open surgical repair and seven underwent endovascular repair, depending on surgeon preference. Mechanism and type of injury, associated injuries and medical conditions, trauma score (for patients with trauma), treatment, and 1-year patency are summarized for the open repair group in Table II and for the endovascular repair group in Table III. The three iatrogenic injuries occurred after subclavian vein catheterization attempts, resulting in arteriovenous fistula between the subclavian artery and vein. A Wallgraft endoprosthesis (Boston Scientific, Natick, Mass) was used for all endovascular repairs. The injury was approached percutaneously through the femoral artery in 4 patients and through brachial artery cutdown in 3 patients. Lesions that appeared difficult to cross were accessed through the brachial artery because this provides a direct, shorter, less tortuous approach. The Seldinger technique was used, with 0.035-inch guide wires and 9F to 10F sheaths. Vessel diameter was not a criterion in selecting patients for open repair or covered stent placement. The size of the stent graft to be implanted was determined with intravascular balloons of known length and diameter. The stent graft was oversized by 10% to 20% to ensure seal. Six procedures were performed with the patient under general anesthesia, and one procedure with intravenous sedation and local anesthesia. All patients who received general anesthetic were intubated before the procedure, because of trauma or primary disease. There were no vascular access site–related complications and no endograft infections in the endovascular repair group. Two patients in the endovascular repair group died of complications of the primary disease, at 3 weeks and 8 months, respectively, after stent placement. There was no clinical evidence of graft occlusion at the time of death. One endoprosthesis became occluded 9 months after placement. Arm claudication developed, and the patient underwent carotid artery to axillary artery bypass grafting. In the open repair group, 1 patient underwent primary repair and 4 patients underwent reverse saphenous vein interposition grafting. All interposition grafts were patent at 1 year. There were no wound infections and no limb loss–related complications in either group. Blood loss was significantly less for patients who underwent endovascular repair (70 ± 12.2 mL vs 220 ± 56.1 mL; P = .01). Procedure time was shorter with the endovacular approach compared with open repair (132 ± 15 minutes vs 193 ± 15 minutes; P = . 04). Patency rate at 1 year was not significantly different between the open and endovascular repair groups (P = 1; Fisher exact test).

Table I. Clinical characteristics of study population

		No. of patients
	Gender (M/F)	16/11
	Age (y)
	Average	33.3
	Range	19-74
	Mechanism of injury
	Penetrating	11
	Blunt	13
	Iatrogenic	3
	Vessel injury
	Subclavian	10
	Axillary	17

Table II. Summary of clinical features of patients who underwent open repair

		Patient 1	Patient 2	Patient 3	Patient 4	Patient 5
	Age	74	34	25	48	33
	Gender	Female	Male	Male	Male	Male
	Mechanism of injury	Motor vehicle accident	Gunshot wound	Motor vehicle accident	Logging accident	Gunshot wound
	Trauma score	10	12	12	6	12
	BP < 90 at presentation (uninjured arm)	No	No	No	No	No
	Associated injuries				Rib fractures, pneumothorax	Brachial plexus
	Vessel injury	Axillary artery, branch bleeding	Axillary artery, pseudoaneurysm	Subclavian artery, intimal dissection	Axillary artery, initimal dissection	Subclavian artery, pseudoaneurysm
	Repair	Primary repair	Bypass with reverse saphenous	Bypass with reverse saphenous vein	Bypass with reverse saphenous vein	Bypass with reverse saphenous vein
	One-year patency	Yes	Yes	Yes	Yes	Yes

Table III. Summary of clinical features of patients who underwent endovascular repair

		Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6	Patient 7
	Age	19	32	31	30	37	44	39
	Gender	Male	Female	Male	Male	Female	Female	Female
	Mechanism of injury	Motor vehicle accident	Gunshot wound	Gunshot wound	Stab wound	Iatrogenic	Iatrogenic	Iatrogenic
	Trauma score	6	12	12	7
	BP < 90 at presentation (uninjured arm)	Yes	No	No	No	No	No	No
	Associated injuries/comorbidities	Multiple intrabdominal, pneumothorax, long bone fractures, brachial plexus			Hemothrox, pneumothorax	Pneumothorax	Self-induced diltiazem overdose, ARDS, renal failure	HIV, renal failure, malnutrition, chronic, anemia, pulmonary edema, candidemia
	Vessel injury	Vertebral artery, avulsion/bleeding	Subclavian artery, pseudoaneurysm	Late (4 mo after injury) subclavian artery, pseudoaneurysm	Axillary artery, pseudoaneurysm	AVF	AVF	AVF
	Repair	9 mm Wallgraft	8 mm Wallgraft	10 mm Wallgraft	9 mm Wallgraft	8 mm Wallgraft	Two 7 mm Wallgrafts	7 mm Wallgraft
	Access site	Femoral	Brachial	Brachial	Femoral	Femoral	Brachial	Femoral
	One-year patency	Yes	Yes	Yes	Yes	No; occluded 9 mo after placement	NA; patient died 3 wk after stent placement	NA; patient died 8 mo after stent placement

ARDS, adult respiratory distress syndrome; HIV, human immunodeficiency virus; AVF, arteriovenous fistula; NA, not available.

Discussion 

Subclavian and axillary artery injuries constitute 5% to 10% of artery trauma in civilians.6 The most frequent cause is penetrating trauma,7 and structures concomitantly involved include the brachial plexus,8 aerodigestive tract, sympathetic chain, and spinal cord.1 Surgical stress, frequently combined with concomitant injuries, results in significant morbidity.1 Kalakuntla et al9 reported a postoperative complication rate of 24%. Mortality ranges from 5% to 30% in various studies.10, 11, 12 Endovascular covered stent placement eliminates the acute need for surgical dissection, decreasing the risk for injuring important adjacent structures such as the vagus nerve, recurrent laryngeal nerve, and phrenic nerve, and the innominate vein. Careful patient selection is necessary, and only focal lesions that can safely be traversed with a guide wire can be approached in this fashion. Patients with decreased life expectancy when long-term patency is not a primary concern might benefit from this less invasive procedure. Also, patients who are poor candidates for general anesthesia can undergo an endovascular approach with intravenous sedation and local anesthesia. In our series, two patients (28%) died of complications of the primary disease, that is, diltiazem overdose 3 weeks after stent placement in 1 patient and complications of human immune deficiency virus infection 8 months after placement in the other patient.

A Wallgraft endoprosthesis was used in all of our patients. This was the only covered stent available in our hospital during the study period. The Wallgraft is a self-expandable braided steel stent covered with a Dacron graft. Covered Palmaz stents and expandable polytetrafluoroethylene (ePTFE)–covered stents also have been used to treat subclavian or axillary artery injuries.4, 5 Patency and other outcomes may be different with Dacron- covered stents versus ePTFE-covered stents.

Technically, we must consider shortening of the stent that occurs with deployment. One of our patients required two grafts for complete coverage of the injury. Care also must be taken to not occlude the vertebral artery during stent placement. The femoral percutaneous approach was used in 4 patients, and a brachial artery cutdown in 3 patients. The brachial approach provides better control of the guide wire and a more direct route to a lesion in which a large part of the circumference of the vessel may have been disrupted. This may also be done percutaneously. We did not find vessel size to be a decisive factor in our choice of open repair or endovascular repair.

Phipp et al13 reported 3 patients in whom stents or stent grafts were placed in the subclavian artery or vein to treat thoracic outlet syndrome, and 6 months to 2 years later had stent fracture. Compression of the endoprosthesis between the clavicle and the first rib is most likely the cause of fracture in this location. The pathophysiology of occlusive disease or thoracic outlet syndrome is different from the pathophysiology of traumatic injuries. We have not encountered problems with stent fracture in our patients.

Short-term results of endovascular repair, as reported by Patel et al14 and du Toit et al5 are encouraging, but long-term durability has not been established. Nevertheless, stent thrombosis does not preclude future revascularization, which, if necessary, is done under less emergent circumstances after the acute injury has resolved.

Subclavian and axillary vessel injuries occur infrequently, and it is unlikely that a prospective randomized trial with a sufficient number of patients will be conducted comparing the endovascular and open approaches. A potential bias was introduced in our study, because selection of open repair or endovascular repair was determined by surgeon preference. The purpose of our report is not to indicate superiority of either of these methods. Rather, we believe that, in properly selected patients, covered stents offer an additional way of dealing with these injuries. A limitation of our study is the relatively small number of patients, although the largest retrospective series to date reported 79 patients,11 and most series include 21 to 28 patients.1, 7, 9

In conclusion, endovascular techniques are an alternative approach to treatment of subclavian or axillary injury, and result in shorter operative time and less blood loss. Short-term patency is similar to that with the open approach. Nevertheless, this is a new technique, and long- term results are under evaluation. Therefore periodic patient follow-up and graft assessment are necessary.