Author + information
- Received March 5, 2016
- Revision received June 21, 2016
- Accepted July 22, 2016
- Published online November 1, 2016.
- Zhenjiang Li, MD,
- Qingsheng Lu, MD,
- Rui Feng, MD,
- Jian Zhou, MD∗∗ (, )
- Zhiqing Zhao, MD,
- Junmin Bao, MD,
- Xiang Feng, MD,
- Jiaxuan Feng, MD,
- Yifei Pei, MD,
- Chao Song, MD and
- Zaiping Jing, MD, PhD∗ ()
- Department of Vascular Surgery, Changhai Hospital, Second Military Medical University, Shanghai, China
Background Stent grafting is a therapeutic option for patients who are unable to undergo urgent surgical repair of ascending aortic dissections. However, follow-up regarding outcomes is limited.
Objectives This study reports mid-term outcomes with endovascular repair for ascending aortic dissections in patients deemed high risk for open repair.
Methods Between May 1, 2009 and January 31, 2011, 15 ascending aortic dissection patients (ages 45 to 78 years) ineligible for direct surgical repair underwent endovascular repair (1 acute dissection, 7 subacute dissections, and 7 chronic dissections) and were closely followed up for a median of 72 months (range 61 to 81 months).
Results The mean interval between aortic dissection onset and treatment was 25.5 (range 6 to 353) days. Technical success was achieved in all patients. No major morbidity or deaths occurred perioperatively. During the follow-up period, there were no deaths, 8 complications occurred, and there were 4 reinterventions. A new dissection in the aortic arch was treated with a branched endograft. One patient developed retrograde aortic dissection and a left ventricular pseudoaneurysm was successfully treated with open surgery. One cardiovascular ischemia was treated with stenting and 1 supraventricular tachycardia was treated with radiofrequency ablation. Other morbidities included perigraft endoleak, a bird-beak sign, a temporary pericardial effusion, and a left kidney atrophy. Significant enlargements of true lumens and shrinkage of false lumens and overall thoracic aorta were observed at 12 months. No significant changes were detected subsequently. Minimal impact on aortic valve function was recorded over time.
Conclusions Our results with the novel endovascular procedure appear acceptable. Additional evidence and studies with larger sample size and longer follow-up are needed to support the durability of this new technique.
Surgical replacement of the ascending aorta has been the gold standard treatment for patients with ascending aortic dissection (AAD). Despite increased understanding of the pathophysiology, improved surgical techniques and devices, and advances in anesthetic management, improvement in overall surgical outcomes remains modest, with a mortality of 10% to 35% (1–3). Furthermore, approximately 10% to 30% of patients with AAD are deemed inoperable due to advanced age, severe comorbidity, or hemodynamic instability (4,5). These patients were deemed high-risk candidates for conventional open repair and only received medical treatment with a high in-hospital mortality of 58% (6).
Minimally invasive endovascular techniques have been explored in an attempt to improve the prognosis of patients at high surgical risk. Initial findings from sporadic reports showed acceptable early results, indicating the feasibility and preliminary efficacy of endovascular repair despite complications associated with procedural cerebral ischemia, progressive aortic valvular insufficiency, and cardiac insufficiency (7–9). One computed tomography-based study (10) indicated that about 30% of patients with type A dissection could be appropriate candidates for endovascular repair. A larger number of patients might benefit from this less invasive approach if the safety, efficacy, and durability are established.
However, limited outcomes prompted better evaluation of this emerging therapeutic modality (9). We offered endovascular treatment to 15 AAD patients unsuitable for surgery in our center between May 1, 2009 and January 31, 2011 and demonstrated acceptable early outcomes (11). To gain better insight into the safety and durability of the endovascular procedure, we followed the initial patient cohort continuously and analyzed the clinical outcomes and impact on aortic remodeling and aortic valve function over a mean follow-up period of 72 (range 61 to 81) months.
Between May 1, 2009 and January 31, 2011, 183 patients with type A aortic dissection were admitted to our hospital. An established clinical treatment process was implemented, and open surgery was considered first. One hundred forty-two patients entered the cardiothoracic center for direct surgical repair. For patients with multiple high surgical risk factors, a multidisciplinary consultation board consisting of cardiac surgeons, cardiologists, vascular surgeons, and anesthetists reviewed all the cases to evaluate the operating risk. A set of criteria was proposed as an integrated strategy to identify poor candidates for open surgery, including advanced age (>70 years), an American Society of Anesthesiologists Classification of IV or more, New York Heart Association functional class III or worse, previous sternotomy, or dysfunction of other important organ systems (i.e., severe chronic obstructive pulmonary disease, renal or hepatic insufficiency, stroke). Of note, a patient was not judged high risk and disqualified from a surgical procedure on the sole basis of advanced age or other single risk factor. We considered all the items comprehensively and identified high-risk patients on an individual basis knowing details of the risk features and specific conditions for each patient. Patients deemed inoperable (n = 41) were transferred to our vascular center for further assessment and possible endovascular treatment. Inclusion criteria were entry tear located in the mid-third of ascending aorta and length of proximal and distal aortic landing zone >20 mm. Exclusion criteria were involvement of coronary artery orifice, acute myocardial infarction, signs of cardiac tamponade, or severe aortic regurgitation (grade 3 or 4) documented by echocardiography, history of potentially life-threatening ventricular arrhythmia, and connective tissue disorders (Marfan or Ehlers Danlos syndrome). Patients who refused or were excluded for surgical or endovascular repair were treated medically.
Fifteen patients were treated using endovascular techniques. The incidence of endovascular procedures was 8.2% (15 of 183) for all type A patients referred to our hospital and 36.6% (15 of 41) for all patients unfit for open surgery. All endovascular procedures were approved by our internal review board, and all patients provided informed consent.
The endovascular procedures have been described previously (11). Cook Zenith TX2 Pro-Form extension stent grafts (Cook, Bloomington, Indiana), commercially available thoracic endografts, were used. They are made of full-thickness woven polyester fabric sewn to self-expanding stainless steel Cook-Z stents with braided polyester and monofilament polypropylene suture. This graft has the available dimensions matched with that of the ascending aorta (diameter varying from 28 mm to 42 mm and length ranging from 77 to 81 mm), which is the key factor. The Zenith TX2 Pro-Form is a modified version of TX2, with improved device and delivery system. The stent graft retained the key features of the original TX2 and provides adequate radial strength and force with appropriate oversize, although it does not manifest best compliance. Furthermore, the design of the stent graft without proximal bare springs was recommended at our center for cases for which there was concern of compromising the aortic valve. The modified deployment system of this new version holding the proximal end closed in a trifold fashion with trigger wires maintains the proximal stent parallel to the aortic inner curvature during deployment and allows micro-readjustments when the stent graft is deployed partially. This design facilitates accurate deployment, alignment, and apposition of the graft in the ascending aorta. This graft conforms adequately in curved aorta (12). Size selection of endograft was based on the ascending aortic diameter proximal to the tear, with about 20% enlargement in the true lumen in case of chronic pathologies to decrease the risk of late migration, and about 10% for acute lesions with fragile aortic walls. Diameters of major axis in the cross section perpendicular to centerline of flow on 3-dimensional reconstructed image (13,14) were selected as the estimated value of aortic diameter.
General anesthesia was used in all cases. Endografts were advanced into the ascending aorta over the extra stiff guidewire through femoral access. In all cases, the endograft was placed in the very proximal ascending aorta just distal to the orifices of the coronary arteries for exclusion of the entry tear and ensuring patency of the coronary and brachiocephalic arteries. Complete details of the procedure including position and deployment of endograft, removal of the delivery system, have been described previously (11).
Follow-up consisted of physical examination, echocardiography, computed tomography angiography (CTA) scans and documentation of mortality and morbidity at 6 months post-procedure and annually thereafter. Three-dimensional TeraRecon reconstructions based on follow-up CTA were used for measurements. Overall aortic, true lumen, and false lumen diameters were measured at the level of sinotubular junction, the ostium of the innominate artery, the point of maximum ascending aortic diameter, and the ostium of the left subclavian artery. Status of patency or thrombosis of the false lumen was assessed using delayed phase imaging. Echocardiography was used to evaluate cardiac and valvular function. Additional CTA examinations were conducted if patients manifested any new symptoms or signs of adverse events.
Continuous variables were reported as mean ± SD. Skewed variables were summarized as median and range. The diameters were compared using 2-sided paired sample Student t tests. A p value <0.05 was considered significant. SPSS software version 19.0 (SPSS, Inc., Chicago, Illinois) was used for all analyses.
The baseline demographics, clinical characteristics, and risks for open surgery in the 15 patients undergoing endovascular repair are summarized in Table 1. The mean age of the patients was 65.0 ± 12.1 years (range 45 to 78 years). Eight patients presented with multiple severe comorbidities or previous sternotomy, although they were younger than 70 years and were considered at high surgical risk and treated with endovascular procedures. Men constituted 80% of all the patients. The median time interval from initial dissection to treatment was 25.5 days (range 6 to 353 days). Based on temporal IRAD (International Registry of Acute Aortic Dissection) classification, 1 patient was found in acute phase (2 to 7 days), 7 patients in subacute stage (8 to 30 days), and 7 patients in chronic condition (>30 days), whereas no patient was in hyperacute phase (0 to 24 h) (15). It is important to note that these 15 patients were treated medically for a variety of reasons specific to the hospital where they initially sought treatment and survived the initial phase of acute aortic dissection. They came because of persistent symptoms or anatomic findings that warranted therapeutic intervention.
All the entry tears were located in the middle portion of the ascending aorta according to pre-operative CT scans and intraoperative measurement. Among these 15 patients, 14 patients were DeBakey type I dissections and one was a type II dissection. Three patients were diagnosed with aortic dissection that proximally extended partially into the aortic root but without coronary artery involvement. Despite that stent grafting should be initiated at the level of “normal” aorta, endograft repair was still attempted in these 3 patients to avoid the progress of the vital pathologies, which will not heal without mechanical exclusion. Although satisfactory outcomes may not be expected, endovascular therapy at least serves as a bridge to definitive open repair, if needed, in later stages. The dissection lesion was confined to the ascending aorta in 1 patient. Antegrade extension to descending aorta was found in 1 patient and propagation to abdominal aorta occurred in 13 patients (Table 1).
The intraoperative details are listed in Table 2. The mean endovascular procedure time was 128.6 ± 26.2 min. The mean ratio of the diameter of the stent graft to that of the aorta (proximal to the tear) was 118.7 ± 6.3%. Average stent graft coverage of the ascending aorta was 84.1 ± 9.3%.
Table 2 summarizes the perioperative and follow-up outcomes. Patients were followed for a median of 72 months (range 61 to 81 months). No deaths occurred during the follow-up, but 8 major morbidities were recorded. A single new dissection occurred in the aortic arch distal to the ascending aortic endograft at 3 months and was successfully treated using a customized branched stent graft (Figure 1). The other patient developed retrograde type A aortic dissection (RTAD) and potentially procedure-related left ventricular (LV) pseudoaneurysm at 29 months post-operatively. He was converted to surgical repair and successfully recovered following replacement of ascending aorta and proximal arch and linear closure of the LV pseudoaneurysm.
One patient manifested severe stenosis of left anterior descending artery with myocardial ischemia symptoms and was relieved after coronary stenting. One patient experienced supraventricular tachycardia at 37 months and was treated with percutaneous radiofrequency ablation. Another patient with temporary pericardial effusion recovered spontaneously with conservative treatment. The CT scan showed the stent graft in the right place. Left kidney atrophy was detected in 1 patient at 59 months with slightly elevated serum creatinine of 143 μmol/l. The patient had undergone left renal artery stenting to restore the blood flow compromised by dissection.
One patient developed perigraft endoleak without dilation of aortic segment at 71 months. The limited perigraft flow was traced to the gap between the distal portion of the endograft and lesser curvature of proximal aortic arch leading to unsatisfactory sealing. In the absence of any contrast material in the false lumen and absence of symptoms, medical treatment and vigilant monitor was prescribed. Another patient developed a bird-beak configuration at the distal landing site at 36 months (Figure 2) without any symptoms. The retrograde flow into the wedge-shaped space was limited without any adverse events. The configuration was stable until the last follow-up (Figure 2).
During follow-up, no migration of the stent grafts or compromised coronary arteries and supra-arch vessels were detected. No patient suffered from neurological complication such as transient ischemic attack, spinal cord ischemia, or stroke.
Complete thrombosis of the false lumen of the ascending aorta was noted in all the 14 DeBakey type I patients and remained stable during the follow-up. Complete thrombosis of the false lumen of the descending aorta was noted in 11 patients (11 of 14, 78.6%), and partial thrombosis was noted in 3 patients (3 of 14, 21.4%). The patient with DeBakey type II dissection was converted to open surgery repair with ascending aorta replacement for RTAD and LV pseudoaneurysm, representing an aortic remodeling failure.
Changes in the mean diameter of the total aorta, true lumen, and false lumen were measured on the basis of follow-up imaging, as shown in Table 3 and Figure 3. Notably, the DeBakey type II patient was excluded when analyzing the changes of ascending aortic dimension due to replacement of dissected ascending aorta and was excluded when analyzing the changes of descending aortic dimension with the lesions confined to ascending aorta. The total aortic diameter significantly decreased at the levels of maximal ascending aorta and distal ascending aorta within 12 months post-procedure. Trends in decreased diameter were also seen at the level of proximal ascending aorta and proximal descending aorta, although insignificant. During the subsequent follow-up, the average diameters of total aorta remained constant. Significant changes in size were observed in both true lumen and false lumen before endografting and at 12 months, indicating clear expansion of the true lumen and shrinkage of the false lumen as a result of tear exclusion. During subsequent follow-up, the diameters of true lumen and false lumen at all the levels measured did not change significantly.
The pre-operative LV end-systolic diameter was 38.9 ± 10.6 mm. At the current follow-up, the LV end-systolic diameter was 42.1 ± 5.4 mm (p = 0.324). The LV ejection fraction was 53.6 ± 2.5% pre-operatively and 52.5 ± 2.1% (p = 0.219) at the end of follow-up. The diameter of the sinotubular junction was 26.8 ± 3.5 mm before procedure and the diameter was 27.6 ± 3.7 mm (p = 0.562) at the end of follow-up, indicating stability of the diameters. As shown in Table 1, 13 patients suffered from mild aortic regurgitation pre-operatively, whereas moderate aortic insufficiency was detected in 2 patients with dissection extending down the root. Improvement in aortic insufficiency from moderate to mild degree was seen in the 2 patients before discharge. During follow-up, all the patients manifested mild aortic insufficiency, despite the slight reflux during diastolic period of echocardiography, revealing the relatively normal function of the aortic valve. In 1 patient with pre-existing LV enlargement and moderate mitral valve regurgitation before procedure, the cardiac function remained stable without continuous dilation of the LV and the reflux volume of mitral regurgitation remained at a mild level of 7 ml.
Dorros et al. (16) first treated AAD with endovascular technique via transseptal access in 2000. Endovascular repair of AAD was reported first in 2003 through femoral access (17). Since then, a number of isolated studies described stent grafting in the ascending aorta in an emergency setting as a bridge or in an elective setting as an alternative for patients who had been deemed inoperable. Current studies showed encouraging early outcomes and confirmed the feasibility in selected patient cohorts despite major adverse events (7–9,18). However, long-term outcomes of this novel therapeutic paradigm have not been reported adequately, and most available studies reported results of follow-up <12 months (19,20).
Early mortality associated with endovascular repair for AAD was acceptable ranging from 0% to 14%, as most cases were treated in their acute phase (8,9,18,21), underlining the minimal physiologic impact of endovascular procedure and indicating the potential value of reducing the mortality in this high-risk patient cohort. In our study, the mortality of perioperative and follow-up period was 0%. This alternative therapeutic method appears to increase the survival rate in selected patients and maintain the mortality benefit for a long-term follow-up. The less-invasive nature of endovascular repair and rapid mobilization and recovery may contribute to this survival benefit, which is associated with relatively low rates of major procedural complications. In addition, the thrombosed false lumen in type A aortic dissection was reported to be an indicator for survival (22). Therefore, early closure of the dissection entry tears using this approach and induced false lumen thrombosis may also help improve the long-term survival, as compared with medical therapy (Central Illustration). However, several biases may contribute to these good outcomes. All these patients were those who survived the initial phase with medicine alone, indicating the selection bias in the study. Meanwhile, the rate of acute AAD was rather low, with 7 chronic dissections in 15 patients, which showed an inherent survival bias.
Endoleaks are considered the Achilles heel of endovascular procedures. In the endovascular treatment for type B aortic dissection, the mid-term endoleak rate was reported as high as 8.1% (22). On this basis, endografting in the ascending aortic segment can be more likely to cause endoleaks, predominantly type I (proximally or distally), which may potentially lead to false lumen dilation and rupture. During operation, good seals were achieved with ascending aortic stent graft in all patients, whereas a single limited endoleak appeared during follow-up, resulting in a late endoleak rate of 6.7% (1 of 15). Endoleak formation was associated with a small radius of curvature of arch and shortage of landing zone for stent graft (23). Therefore, the low rate of early endoleak in our study may benefit from strict patient selection with proximal or distal landing zone >20 mm and entry tear in the middle part of ascending aorta. However, the discrepancies of configuration between straight endograft and curved proximal aortic arch, the pulsatory motion of aorta and strike of pulsatory blood flow may deteriorate the early benefit and result in perigraft endoleak and even endograft migration in late phase. A bird-beak sign occurred at the distal end of stent-graft in another patient at 36 months, which may be attributed to dilation of distal ascending aorta and significant proximal arch angle transformation as a result of aortic remodeling (24), underscoring the risk of incomplete endograft apposition to the aortic wall in late phase. To achieve sustainable endograft conformance, assessment of the proximal arch angle was encouraged when confirming the candidacy for endovascular treatment. A new generation endograft matching the inner curvature of the distal ascending aorta and proximal arch was needed.
In the ascending aortic segment, close to the aortic valve, high aortic impulses cause significant pulsating motions of the aortic wall and inadvertent displacement of the stent graft. Compared with the descending aorta, stable fixing of stent graft in this aortic segment is a challenge. To achieve adequate fixation, a larger oversize rate was needed. However, the risk of stent graft–induced injury increased, which can result in new dissection. The appropriate oversize of stent graft is crucial to balance stable fixation to the aortic wall and less intimal injury. Furthermore, 1 RTAD observed in this series highlights the fragility of the ascending aorta in dissection and the need for selection of appropriate stent size and specific device design. To address these issues, further biomechanical and hemodynamic investigation focusing on ascending aorta, stent graft configuration, and their interaction may be inevitable. And in the future, an endostapling system should facilitate more favorable fixation, with minimal injury and better long-term performance (25). In addition, this RTAD patient underwent surgical replacement of dissected ascending aorta and recovered uneventfully, indicating that the endovascular approach can also be used as a bridge in emergent setting for high-risk patients and later converted to definitive open repair under improved patient conditions.
Manipulation of tracking wire or tapered cone of the delivery system in the LV may lead to injury of the intima of the ventricle, which results in adverse cardiac events, such as the LV aneurysm seen in our study. This complication was reported to occur in up to 1.9% of patients undergoing transfemoral transcatheter aortic valve implantation, in which intraventricular manipulation was also needed (26,27). The J-ended guidewires that are widely recommended lower the incidence of LV trauma. Our experience suggests that the grafts with shorter and soft tip or nose cone minimize the injury when operating the device in the ascending aorta and the ventricle.
As reported in previous studies on endovascular repair of descending aortic dissection (28), in most patients, the aortic remodeling was completed within 1 year after procedure, whereas false lumen thrombosis rates at the diaphragm level increased over 3 years. In our study, significant changes in the diameter of aorta were observed in the first year post-operatively. The diameters of total aorta, true lumen, and false lumen remained stable or changed slightly without significant difference at 3 years and during subsequent follow-up in the absence of any new dissection or endoleak. It can be explained as a common process during the aortic remodeling, which is consistent with long-term results of endovascular treatment for Stanford type B dissection. In terms of morphological changes associated with the ascending dissection, the therapeutic effect of endovascular exclusion can be sustained.
After graft implantation, the aorta may enlarge and stiffen, and subsequently vascular compliance may decrease, which could increase systolic pressure, decrease diastolic pressure, and consequently affect cardiovascular health and function resulting in aortic regurgitation and LV hypertrophy (29). We hypothesize that endografts with greater compliance would have better performance, with mechanical properties similar to those of human ascending aorta, simulating its contraction and expansion functions. Meanwhile, compliant and flexible devices may interact with the aortic wall with less invasion. Notably, although the Cook Zenith TX2 used in this series was not the option with best compliance (30), our study showed relatively low rates of device-related complication and cardiovascular events over time. As no severe aortic regurgitation occurred and the LV ejection fraction remained within normal range despite a slight decrease, it seems that endografting of the ascending aorta does not adversely affect long-term cardiac function or aortic valve competence. However, this benefit should be verified in further studies with larger sample size and longer follow-up. In the future, studies measuring the circumferential aortic strain and aortic deformation may provide insight into the mechanism.
Improvement in stent graft design is needed to achieve better outcomes. Devices with larger diameter in shorter length, conformability to the curvature of aorta and flexibility to decrease device-related injury should be designed. Advanced delivery systems should allow for stepwise deployment and repositioning for accurate deployment. A short and soft tip may decrease ventricular and valve trauma. In the future, a composite stent graft with a valve component may extend the proximal landing zone to the aortic valve annulus, and a specially designed branched stent graft may extend the distal landing zone to aortic arch.
The sample size of this study was small (n = 15) and the patients were those who survived the initial phase with medicine alone, indicating a highly selected cohort with inherent survival bias. Additional studies with patient stratification (e.g., subgroups of acute, subacute, and chronic dissection) based on larger numbers of patients and longer follow-up are needed to arrive at definitive conclusions. To verify the long-term effectiveness of endovascular procedure, comparison with similar patient cohorts treated with medications is clearly needed.
Currently, open surgery remains the standard therapy for ascending aorta dissection. The mid-term results of our novel endovascular intervention in high-risk patients suggest acceptable morbidity and mortality with positive aortic remodeling and minimal impact on aortic valve function. Longer-term follow-up and larger studies are required to confirm the effectiveness and durability of this new technique.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In patients with ascending aortic dissection at high surgical risk, endovascular repair can be performed with minimal adverse effect on aortic valve function and sustained survival.
TRANSLATIONAL OUTLOOK: Larger studies are needed to confirm the durability of endovascular therapy compared with medical treatment and surgical repair and to more clearly identify those most likely to benefit from this approach.
The authors thank Suming Zhang, BS, for his contribution to the data collection.
Supported by the National Natural Science Foundations of China (grants 81330034, 81270386, 81273522), and the Shanghai Medical Talents Training Plan (grants XYQ2013087). The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Li, Lu, and Feng contributed equally to this work.
- Abbreviations and Acronyms
- ascending aortic dissection
- computed tomography angiography
- left ventricle
- retrograde type A dissection
- Received March 5, 2016.
- Revision received June 21, 2016.
- Accepted July 22, 2016.
- American College of Cardiology Foundation
- Bonser R.S.,
- Ranasinghe A.M.,
- Loubani M.,
- et al.
- Zhang H.,
- Li M.,
- Jin W.,
- Wang Z.
- Lu Q.,
- Feng J.,
- Zhou J.,
- et al.
- Melissano G.,
- Civilini E.,
- Bertoglio L.,
- Logaldo D.,
- Chiesa R.
- Booher A.M.,
- Isselbacher E.M.,
- Nienaber C.A.,
- et al.,
- for the IRAD Investigators
- Dorros G.,
- Dorros A.M.,
- Planton S.,
- O'Hair D.,
- Zayed M.
- Pinaud F.,
- Daligault M.,
- Enon B.,
- de Brux J.L.
- Sze D.Y.,
- van den Bosch M.A.,
- Dake M.D.,
- et al.
- Stortecky S.,
- Buellesfeld L.,
- Wenaweser P.,
- Windecker S.
- Masson J.B.,
- Kovac J.,
- Schuler G.,
- et al.
- Bosman W.M.,
- Hinnen J.W.,
- Rixen D.J.,
- Hamming J.F.