Abstract
Thoracic aortic aneurysms represent a major health problem. Untreated thoracic aortic aneurysms may rupture, which has a dismal outcome. The standard treatment for thoracic aneurysms is open surgical repair, but it is associated with high mortality and morbidity. Endovascular repair provides a less invasive and safer alternative.
A systematic review was performed of all published literature on the above subject. Our primary objective was to measure 30-day mortality for nonruptured thoracic aortic aneurysms. Studies describing other pathologies, such as aortic dissection, mycotic aneurysms, penetrating ulcers, traumatic transactions, and pseudoaneurysms, and studies from which independent data for thoracic aortic aneurysm could not be separately extracted were excluded. Case series describing less than 10 patients and all case series describing ruptures or concealed ruptures were excluded as well.
Twenty-six case series and one comparative study were identified. This formed a cohort of 1,038 patients. Technical success was possible in more than 97% of patients. The 30-day mortality was calculated to be 5.1% even though the group under study was mostly those who were refused surgery by a surgeon or had a higher surgical risk. The incidence of paraplegia and stroke was 3.1% and 4.7%, respectively. Early endoleak was seen in 16.7% of patients, whereas 11.7% of patients developed late endoleak, but most did not require any additional procedure. The rate of reintervention was 14.9%. The 12-month mortality rate was 14.2%.
Endovascular repair shows encouraging short-term results. It is associated with significantly less mortality and morbidity, but long-term results need to be further investigated.
Thoracic aortic aneurysms (TAAs) represent a major health problem. Like abdominal aortic aneurysms (AAAs), most TAAs are asymptomatic and are diagnosed incidentally. Although less common than AAAs, aneurysms of the thoracic aorta may be associated with a worse prognosis. Following diagnosis, untreated patients with TAA have a 2-year survival of < 30%, with half of deaths resulting from aneurysm rupture. This dismal survival is significantly improved by operative intervention. 1 Although open repair has become a refined surgical process, with extracorporeal circulation for peripheral organ preservation and multiple techniques for spinal cord protection, it has nevertheless been associated with significant mortality, and cumulative morbidity in high-risk population frequently exceeds 50 to 70%. The highly invasive nature of this procedure necessitates a prolonged recovery period, with return to well-being frequently delayed to 4 to 6 months postoperatively. Additionally, high-risk patients previously denied surgical repair might become surgical candidates if less invasive endovascular option were possible. 2 Endovascular treatment of aortic aneurysms stemmed from Parodi's initial studies, 3 which were reported in 1991. Since then, endovascular repair of that infrarenal abdominal aorta has become routine, with most large centers offering endovascular treatment as part of established clinical practice. As a logical sequel to endovascular repair of abdominal aortic aneurysms (EVAR), many investigators studied the possibility of using endovascular techniques to treat lesions of the thoracic aorta. Initial studies have shown favorable results. 4
TAAs are different from AAAs because they have quite varied pathology, such as true aneurysms, dissections, false aneurysms, mycotic aneurysms, and traumatic rupture. As they have varied etiology and prognosis as well, each should be studied separately.
The purpose of this study was to assess the evidence for efficacy and safety of EVAR in TAAs only.
Method
Search Strategy
The following databases were searched: BIOSIS, CINAHL, Medline-Central, CDSR, CRD, EMBASE, HTA, NHS EED, Medline in Process, Science Citation Index, and Social Sciences Citation. The keywords were “thoracic aortic aneurysms and aortic dissection,” “aortic rupture,” “stents,” “endograft,” “endovascular aneurysm repair,” “EVAR,” “intravascular grafts,” and “endoluminal repair.” The search was limited to reports published in English in the last 10 years. The data were collected up to May 2007.
Inclusion Criteria
Inclusion criteria were as follows:
Studies comparing endovascular repair with open surgical repair for thoracic aortic aneurysms only
Studies reporting on the safety of endovascular repair (ie, 30-day mortality for TAAs only)
Exclusion Criteria
Exclusion criteria were as follows:
Results
Twenty-six case series and one comparative study were identified. The data extracted from 26 case series were for nonruptured TAA alone.
In all studies combined, 1,038 patients underwent endovascular repair. The mean age of patients calculated from the data available in 469 patients was 71.7 years. For 295 patients, gender was mentioned; 62.3% of those patients were male and 37.6% were female. The vast majority of patients who underwent endovascular repair were at high risk for surgery because of existing comorbid conditions. Some of the above studies reported the results of endovascular grafting in patients who were unfit for surgery or were refused surgery by surgeons.
Procedure
Technical success was defined as complete exclusion of the aneurysmal sac without any additional procedure and without any perioperative death or complication. 7 Technical success was possible in 97.4% of patients (cohort n = 588). The conversion to conventional surgery was not mentioned in all studies, but it was quite low when reported: 0.82% in 342 patients. We were able to extract a reintervention rate from 498 patients, which was 14.9% (74 of 498) and mostly due to persistent endoleak.
Most patients were followed up with computed tomographic (CT) angiography. They tended to have CT angiography at day 30, 6 months, and 12 months and then yearly after the first year. The mean follow-up calculated from available data was 38 months.
Mortality
The primary aim of the review was to calculate early mortality with endovascular repair in unruptured TAAs. The 30-day mortality calculated from all studies was 5.1% (n = 53) and ranged from 0 to 25%. The individual mortality rates for each study are given in Table 1.
30-Day Mortality of All Studies
Data were available for 528 and 262 patients for 12-month and 24-month mortality, respectively. Most deaths were not related to aneurysm but rather to other comorbid problems as this group has significantly high comorbidity compared with the general population. The 12-month mortality was 23.0% (123 of 528), and the 24-month mortality was 17.5% (46 of 262).
Paraplegia and Stroke
As with open surgery, paraplegia is a devastating complication of endovascular repair. Endovascular repair of TAA tends to decrease the incidence of spinal cord injury when compared with open surgery. 4 The incidence of spinal cord ischemia was 3.1% in a cohort of 19 studies (n = 745) in this review. The incidence rate varied from 0 to 11.1%.
Another main postoperative complication of endovascular repair of aneurysm of the thoracic aorta was cerebrovascular accident. 7 Our study showed that the rate of stroke varied from 0 to 13.3% in a group of 15 studies (n = 579). The mean incidence rate was 4.7%. There were not enough data to determine the incidence of acute and delayed paraplegia or stroke.
Endoleak
An endoleak is defined as the persistence of blood flow within the aneurysm sac, outside the lumen of the stent graft. Endoleaks following AAA repair are classified into five types. Type I involves a leak around the stent graft at proximal or distal fixation sites. Type II endoleak occurs from retrograde perfusion of the aneurysm sac. Type III is due to leak through a fabric defect or through the graft-to-graft attachment site. Types IV and V refer to graft porosity or endotension, respectively. High-pressure leaks (types I and III) have been implicated in aneurysm expansion and rupture postprocedure. These leaks are often due to incomplete fixation of endograft or migration of the graft following deployment. 8 The same classification is followed in thoracic aneurysm repair.
It was possible to extract data on the incidence of early endoleak (occurring within 30 days of the procedure) from 14 studies, which was 16.8% (75 of 447). From this cohort, 405 endoleaks were further subclassified into different types of endoleak: type I, 10.4% (n = 42); type II, 3.2% (n = 13); and type III, 3.0% (n = 12); no type IV or V endoleak was observed.
The incidence of late endoleak (occurring after 30 days of the procedure) was described in 13 studies (n = 436) and was 11.7% (51 of 436). Of this cohort, 337 endoleaks were subclassified as follows: type I, 34 (10.1%); type II, 9 (2.7%); and type III, 5 (1.5%); no type IV or V endoleak was seen.
The total number of endoleaks in our study was 136 (30.4%). However, most of the endoleaks did not require any additional procedure.
The included studies did not consistently report the extent of aortic coverage and location of endografts; hence, we are unable to make any conclusions regarding them.
Other Complications
It was possible to separate the complication of stent graft repair of unruptured thoracic aorta in 13 studies (n = 497). Rupture of the aorta (3.8%) and graft migration (2.4%) were the most frequently encountered complications specific to stenting, and 10.6% of patients had developed systemic complications as a result of the procedure. All complications are listed in Table 2.
Other Complications (from a Cohort of 497 Patients)
Freedom of Reintervention
The actual freedom of reintervention was not given in all studies but was calculated from four (n = 268) to be 75.4% at 1 year.
Hospital Stay, Intensive Care Unit Stay, and Average Blood Loss
It was not possible to extract data from the studies that were statistically significant, but it was clear that hospital stay, intensive care unit stay, and average blood loss per procedure are markedly less when compared with conventional open surgery.
Open versus Endovascular Surgery
Only one comparative study was identified that compared head-to-head results of both interventions. Bavaria and colleagues compared 140 patients who underwent endovascular repair for descending thoracic aorta with an open surgical cohort of 94 patients. 2 There was no statistically significant difference between the two groups in age, sex, and perioperative risk factors. Of 140 patients, technical success was possible in 137 in an endovascular cohort (success rate 97.8%). Perioperative mortality (ie, death within 30 days of the procedure or within the first hospital admission) was 2.1%(n = 3) in the endograft group, whereas it was 11.7% (n = 11) in an open surgical cohort (p = .04). The incidence of spinal cord ischemia was 2.9% in the endograft group (4 of 140), whereas the surgical cohort significantly raised the incidence to 13.8% (13 of 94). The incidence of cerebrovascular accident was similar in the endograft and surgical cohorts, 3.6% (n = 5) versus 4.3% (n = 4), respectively.
The endovascular cohort had a lower incidence of respiratory and renal complications but a higher incidence of peripheral vascular complications when compared with the surgical cohort. However, there was no significant difference between the endograft group and the surgical group in terms of 2-year survival, which was 78% and 76%, respectively. 2
Discussion
Treatment of thoracic aneurysmal disease is challenging. Nonoperative survival is dismal. 2 The standard technique for treatment of TAAs is elective open surgical repair with graft interposition. The standard approach involves a major surgical procedure with lateral thoracotomy, use of cardiopulmonary bypass, deep hypothermic circulatory arrest, a long procedure time, and a variety of perioperative and postoperative complications. The peri- and postoperative complications vary from coagulation disorders to paraparesis and paraplegia. 9 Surgical mortality ranges from 8 to 26%. 10
Catheter-based techniques, with their theoretic advantages, have changed the trends of vascular surgery. The minimally invasive character of stent grafting, with the obvious avoidance of major surgery and aortic cross-clamping in case of a diseased thoracic aorta, is challenging, although the long-term results are unknown. For certain groups of patients at high risk with severe concomitant diseases, for patients with previous multioperations, and for patients with multiple traumas, the endoluminal method offers an alternative solution instead of a frustrating open procedure, although, again, the successful outcome would by no means be totally ensured because, as with every surgical procedure, endoluminal interventions are also accompanied by complications. 11
It is noteworthy, however, that the mortality and morbidity rates of the entire data collected during this review are similar to those of the largest studies. 12 The mortality rate in our study for the first 30 days was 5.1%. Because of the incomparability of patients, these early mortality rates cannot simply be compared with those of the open surgical group. 10 However, we can safely say that this review has shown that the early results of endovascular repair are significantly better than those of the open approach, although the population who received endovascular intervention is in a much higher surgical risk group. Despite encouraging early term results, only one study, Demers and colleagues, looks at longer term follow up (ie, 5-year outcome). 13 In their study, the 5-year survival for the endovascular group was 49%, which was considerably lower than the expected survival for a population matched for age and gender. However, late survival in good surgical candidates (78% at 5 years) was satisfactory and comparable to contemporary results after conventional open treatment.
Apart from mortality, the success of EVAR for TAA is determined above all by the rate of rupture and conversion to conventional open repair. 11 Our study showed encouraging results in all of the above aspects as technical success was more than 97% and the rate of conversion was below 1%. Aortic rupture occurred in only 3.8% of cases. To be an acceptable alternative to open repair, EVAR for TAA also needs to be a durable method of excluding the aneurysm as late rupture or late conversion to open repair would significantly undermine the mortality and morbidity benefits of the primary procedure.
The Achilles heel in stent graft therapy is incomplete sealing with an early endoleak. Of course, most endoleaks appear at either the proximal or distal neck or in the attachment zone between the two stent grafts. 9 These type I or type III endoleaks, if required, can be easily and successfully treated with another graft. Type II endoleaks are much less common in thoracic the aorta.
Late endoleak owing to migration or dislocation at the graft ends or modular junctions may result from inadequate grip or seal but also from inability of a relatively inflexible device to resist or adjust to distorting forces applied by changing aortic dimensions after successful exclusion. 14 The incidence of endoleak will be seen less and less in the future owing to constantly improving grafts and technology. The rate of endoleak is lower in the new generation of endografts compared with the older version, as evident in the study by Appoo and colleagues. 15 The endoleak rate was 14% compared with the older graft, for which it was 45.7%. The rate of endoleak in our review was 30.4%, and most of the endoleaks did not require any further intervention.
Paraplegia seems to be one of the most clearly decreased complications with the endovascular approach. 16 Our review showed that the overall incidence of spinal cord ischemia was 3.1% (paraplegia rate for open repair varies between 2 and 11%). 17 This lower complication rate compared with open surgery can be explained by the shorter procedural time, the persistent blood circulation throughout the procedure, and the absence of aortic cross-clamping, so the stent graft probably does not produce a steal phenomenon in the perfusion of spinal cord. Concomitant or previous AAA repair and long segment thoracic aortic exclusion appear to be important risk factors for spinal cord ischemia. Spinal cord protective measures such as cerebrospinal fluid drainage, use of steroids, and prevention of hypotension should be used for patients with the aforementioned risk factors. 14
Strokes are caused by manipulation of wires and catheter in an atherosclerotic aortic arch. 14 To keep the embolic risk low, the tip of the delivery system should be as short and flexible as possible; no manipulation of the artificial aortic valve should be done where clots could be present, and good anticoagulation is mandatory. Finally, it is imperative to flush completely the delivery system to avoid gas emboli. 14
As would be expected from a minimally invasive technique, the incidence of systemic complications such as myocardial ischemia, renal failure, and respiratory complications appears to be reduced compared with the open surgical approach.
Conclusions
Although no randomized trials have been performed, early results of endovascular repair demonstrate a number of advantages over open surgical repair when feasible. However, few long-term data are available to confirm the advantages of endovascular repair. More detailed long-term studies are required to clearly show the durability of EVAR for treatment of aneurysmal disease in the thoracic aorta.
Footnotes
Acknowledgment
Financial disclosure of authors: Umar Sadat is funded by a Royal College of Surgeons and Medical Research Council (UK) clinical research training fellowship.
Financial disclosure of reviewers: None reported.