First described by Shennan in 1934,1 penetrating aortic ulcers are among the least frequent entities causing acute aortic syndrome although asymptomatic cases are increasingly identified in thoracic imaging studies. They are caused by the ulceration of a previous atherosclerotic plaque, penetrating the aortic lumen from the internal elastic lamina to the artery media.2 Usually affecting elderly patients with previous atherosclerosis, penetrating aortic ulcers are located in the distal aortic arch and descending thoracic aorta (Standford type B)3 and may be associated with intramural hematoma or other aortic syndromes.4.

The natural history of aortic ulcers remains unclear and reports on their malignancy are contradictory.5, 6, 7, 8 However, especially in symptomatic patients (associated with acute aortic syndrome), a high risk of progression (to growth, pseudoanaeurysm formation, or aortic dissection) and of aortic rupture (early rupture in ≤38% of symptomatic patients) have been described,5 associated with high mortality.

Endovascular treatment of penetrating aortic ulcers has shown promising results and mortality has been reduced by 0%-11%9, 10, 11 in the treatment of selected cases–even in patients contraindicated for open surgery, which continues to be associated with high mortality (20%).12, 13 However, mean follow-up studies published to date are short-term –usually <3 years 9 , 10 , 13 , 14–and offer little information on mid-term behavior. The objective of the present study is to present short- and mid-term results of endovascular treatment of penetrating aortic ulcers in the last 10 years.

Initial treatment of aortic ulcers in our center (Instituto Clínico del Tórax, Hospital Clínic, Barcelona, Spain) is medical (monitoring, control of blood pressure, and aortic imaging). Only symptomatic patients (acute aortic syndrome) or asymptomatic patients associated with rapidly progressing aortic diameter >60mm in the ulcerated zone, or ulcer diameter and depth >20mm,3 have been referred for surgical repair. Endovascular treatment has been preferred unless the aortic anatomy was unfavorable.

In this context, from June 1999 to August 2010 we treated 22 consecutive patients using aortic endografts for penetrating ulcers in the aortic arch and descending thoracic aorta. Surgery was indicated for acute aortic syndrome in 11 patients (50%): chest pain in 8 (36.4%), aortic rupture in 1 (4.5%), distal embolization with chest pain in 1 (4.5%), and aortobronchial fistula in 1 (4.5%). The remaining 50% were operated because they presented >60mm aortic diameter in the ulcerated zone (5 patients) or >20mm ulcer diameter and depth (6 patients). All preoperative, operative, and follow-up data were recorded prospectively and met EUROSTAR criteria.

Most patients were men (90.1%); mean age was 69.8 years with a high prevalence of comorbidity (Table 1). In all patients, diagnosis and aortic measurements were based on preoperative computed tomography angiographic imaging. Seven patients (31.8%) had associated intramural hematoma, usually small in extent and aortic ulcer-related.

Table 1. Preoperative Characteristics.

ASA, American Society of Anesthesiologists’ classification of physical fitness.
Described according to the EUROSTAR protocol 15 and the Society for Vascular Surgery/International Society for Cardiovascular Surgery clinical practice guidelines. 16
Data are expressed as no. (%) or mean (range).

In the sample, 27.3% (6 patients) required previous supraaortic trunk debranching to obtain an adequate proximal anchor zone in the aortic arch; 2 patients (9.1%) underwent left subclavian artery occlusion without previous revascularization.

All procedures (22) used general anesthesia, orotracheal intubation, and systemic heparinization in a hybrid theater with radiologic control using Siemens Axiom Artis equipment (Siemens, Tarrytown, United States) of recent vintage, and selective transesophageal echocardiography in patients with limited proximal neck (5 patients, although frequency of use is increasing). We used femoral surgical access to advance the device (except in 4 patients who required iliac approach), and 1 left humeral percutaneous or contralateral femoral approach, and advanced a pigtail catheter to the aortic arch or ascending aorta for intraoperative angiography. We implanted 1.3 devices per patient, requiring strict pharmacologic control of blood pressure at deployment (mean blood pressure <80mmhg in proximal anchorage above zones z1 and z2 to the left subclavian artery ostium more marked hypotension mean blood pressure <50mmhg with the aid of an occlusion balloon in inferior vena cava z0 anchor zones proximal to brachiocephalic trunk 17 We used prophylactic control of intra- and postoperative cerebrospinal fluid pressure in 2 patients (9%). Details of intraoperative data are in Table 2.

Table 2. Intraoperative Data.

Data are expressed as no. (%) or mean (range).
*Endograft proximal anchor zone in the aortic arch: Z0 (proximal to the brachiocephalic trunk ostium), Z1 (between the brachiocephalic trunk and the left primitive carotid artery), Z2 (between the left primitive carotid artery and the left subclavian artery), Z3 (aortic arch distal to the left subclavian artery), and Z4 (descending thoracic aorta). 17

We obtained descriptive pre- and in-hospital data (mean±SD, median [range]) and frequency (no. [%]) with SPSS 15.0. We used Kaplan-Meier survival curves to analyze the time until each event during follow-up (complications, reinterventions, and procedure- or aorta-related mortality); we calculated means and 95% confidence intervals and estimated survival at 100 months. Differences between groups were obtained by Kaplan-Meier survival analysis and evaluated with the Mantel-Cox log-rank test (factors possibly associated with poor prognosis: age, acute aortic syndrome, urgent presentation, presence of associated intramural hematoma, aortic diameter). We considered P<.05 statistically significant.

All patients were treated successfully with no complications or intraoperative conversions (Figure 1, Figure 2). The aortic ulcer was excluded in all cases and all patients with chest pain reported that this disappeared postoperatively. However, during hospitalization 27.3% of the sample presented postoperative complications, related to the procedure or to previous comorbidities (Table 3). Only 1 neurologic complication occurred (paraparesis) in a patient without preoperative cerebrospinal fluid drainage, and this was fully resolved spontaneously in-hospital.

Figure 1. Exclusion of an aortic ulcer (pre- and postoperative computed tomographic angiography) in 2 different patients.

Figure 2. Exclusion of an aortic ulcer (pre- and postoperative computed tomographic angiography) in 2 different patients.

Table 3. In-Hospital Data.

Data are expressed as no. (%) or median (range).

Two in-hospital reinterventions were required: embolization of the left subclavian artery for early type II endoleak in a patient receiving a Talent endograft (Medtronic AVE, Santa Rosa, California, United States) with subclavian coverage without previous revascularization, and surgical repair of a humeral pseudoaneurysm following puncture. Both were successful. One in-hospital death occurred due to septic shock and acute respiratory failure caused by postoperative bronchoaspiration. The remaining patients were discharged following a median 3 days hospitalization.

After a mean 52.3 (0.1-122) months follow-up, 2 patients presented new complications (for type I proximal leak and type III leak due to displacement of 2 previous endografts). Both required repeat endovascular reintervention (deployment of new endografts to seal the leak). No conversion to open surgery occurred nor were there any new procedure- or aorta infection-related deaths during follow-up. Two patients required endovascular repair of an abdominal aorta aneurysm and 1 required endovascular repair of a new aortic ulcer proximal to the previous repair (Table 4).

Table 4. Short- and Mid-Term Follow-up Data.

Data are expressed as no. (%) or mean (range).
*Cumulative survival at 100 months follow-up.

At 100 months follow-up, complication- and reintervention-free cumulative survival were 61.7% and 79.5%, respectively (Figure 3). Procedure- or aorta-related cumulative mortality was 4.5% (only 1 in-hospital death), and total cumulative mortality from other causes was 38.8%.

Figure 3. Kaplan-Meier survival curves: complication- and reintervention-free cumulative survival during follow-up (61.7% and 79.5%, respectively, at 100 months).

We analyzed factors possibly associated with poor prognosis (complications, reinterventions, or procedure- or aorta-related mortality) following treatment of aortic ulcers: age, acute aortic syndrome, emergency presentation, presence of associated intramural hematoma,4 and aortic diameter.18 We found no statistically significant relationship in any case (P>.05).

Penetrating ulcers of the thoracic aorta present in elderly patients with many comorbidities, and high blood pressure is the most frequent risk factor.2, 19 Advanced age and the frequency of comorbidities have particularly encouraged many groups to treat these lesions by using endovascular surgery, which provides better results, both in early morbidity and mortality, than does open surgery.9, 10, 12, 13 In fact, the only death in our series was a consequence of bronchoaspiration in an octogenarian, with no direct aortic complication.

Just how malignant the clinical course of penetrating aortic ulcers actually is remains controversial. While some publications report a benign clinical course and low risk of progress, aortic rupture or other serious complications,6, 19 several prospective series find high risk of progress and aortic rupture (≤38% early rupture, above all in symptomatic patients) and even ≤40% of urgent procedures for aortic rupture in patients receiving medical treatment.4, 5.

While initial treatment of aortic ulcers can be considered medical (monitoring, control of blood pressure, and aortic imaging), there is no consensus about indications for surgical or endovascular treatment. Our group indicates endovascular intervention in patients with penetrating aortic ulcer with infection of the aortic arch or descending thoracic aorta who present favorable anatomy and when associated with symptoms (acute aortic syndrome) or in asymptomatic patients associated with rapidly progressing >60mm aortic diameter in the ulcerated zone, or with >20mm ulcer diameter and depth.3.

Results of endovascular treatment in this type of patient have been promising in several series that have achieved high rates of technical success (92%-95%). Moreover, given the small extent of the lesions and the good condition of the proximal and distal aorta, in >90% of cases only 1 device is used,14, 20 results that closely coincide with our series.

Given that many ulcers are located in or near the aortic arch, many previous supraaortic trunk debranching procedures were needed (27.3%), as reported elsewhere (5.3%-29.7%).4, 9 Differences in prevalence may be due to the more aggressive treatment of proximal lesions and the increasing use of carotid-left subclavian artery bypass before occlusion of the latter, in order to minimize the risk of ischemia of the upper extremity, stroke, or spinal chord ischemia.15, 21, 22, 23 In fact, our attitude on this issue has changed over time. Before 2006, we performed selective revascularization (dominant left vertebral artery, previous coronary bypass with internal mammary artery, left arm vascular access); since 2006, we have performed systematic revascularization. Moreover, by associating the bypass to ligation or embolization of the proximal subclavian artery, we eliminate the risk of type II leak described in the earlier cases.

As well as supraaortic trunk debranching, dorsal artery occlusion by the endograft can also increase risk of spinal chord ischemia, above all in patients with a deteriorated collateral network (occlusion of subclavian or internal iliac arteries, previous abdominal aortic or thoracic surgery)15, 24, 25 and extensive aortic coverage. This can be minimized through intra- and postoperative control of cerebrospinal fluid drainage.26 However, specifically in aortic ulcers—due to their limited extent and the use of single, usually short devices—we predicted low risk of spinal chord ischemia. Consequently, cerebrospinal fluid control was only used in patients with extensive aortic coverage or previous aortic surgery, 9% of cases.

In our series, the only instance of spinal chord ischemia was in a patient with no left subclavian artery coverage, who underwent short aortic coverage; this patient had not had preoperative cerebrospinal fluid drainage.

As well as low initial and 30-day mortality, several series report short-term mean follow-up (usually <3 years 9 , 10 , 13 , 20), and larger registries are the exception. 3 Most authors coincide in describing low postoperative and in-hospital mortality (from 0% 13 to 14.6% 9 , 27), usually in symptomatic, emergent patients with high comorbidity. Although complication rates during follow-up are significant (from 18% 4 to 30% 14), these are usually resolved with endovascular techniques, achieving low or nil mortality related to the aorta or conversions to open surgery during follow-up. 9 , 10 , 27 The series we present confirms the continued achievement of the good mid-term results described to date. Even though it is not complication- and reintervention-free (38.3% and 20.5%, respectively, at 100 months follow-up), in the mid-term, endovascular surgery associates with high cumulative survival free from aorta- or procedure-related death (95.5% at 100 months).

Endovascular treatment of penetrating ulcers of the aortic arch and descending thoracic aorta is possible and effective despite high patient comorbidity. It associates with a considerable rate of complications and reinterventions (38.3% and 20.5% respectively, at 100 months follow-up), above all in-hospital. Nonetheless, long-term aorta- or procedure-related mortality is low (4.5%).

Dr. Riambau receives consultancy fees from Aptus Endosystems Inc., Bolton Medical, Cordis Corporation, Medtronic Inc., TriVascular Inc., and W.L. Gore & Associates.

Received 6 May 2011
Accepted 27 July 2011

Corresponding author: Sección de Cirugía Vascular, Servicio de Cirugía Cardiovascular, Instituto Clínico del Tórax, Hospital Clínic, Villarroel 170, 08036 Barcelona, Spain. gasparmestres@gmail.com