Discussion

Article Outline

• Copyright

Dr Quinones-Baldrich (Los Angeles, Calif). I would like to congratulate Dr Brown and the vascular service at Northwestern University for an excellent presentation of their 7-year experience with endovascular repair of complex thoracic aortic pathology. They have treated 111 high-risk patients with either complete endovascular or a hybrid alternative to surgical repair. Some of these interventions were not really minimally invasive. It is really a new day in vascular surgery when we call an aortic debranching or an elephant trunk part of a minimally invasive approach. They do represent an option to what is otherwise a formidable operative intervention. Although a comparison to standard surgical repair is not really appropriate, the overall results support their conclusion that endovascular repair of complex aortic pathology is a viable alternative in the management of these difficult patients. As an advocate of endovascular treatment to these complex cases, I remain concerned that the overall results are not a significant improvement over those accomplished by experienced surgeons performing open repair.

A combined stroke and paraplegia incidence of about 12% must be reduced if we are to continue to offer endovascular repair as a better alternative. Keep in mind that their series includes patients with penetrating aortic ulcer, intramural hematoma, traumatic injuries, and other pathology in which the risk of these complications is relatively low for open surgical repair. Spinal catheter drainage was used in 17 patients. In several of these patients the indication for spinal drainage was related to the extensive coverage of the descending thoracic aorta. In the rest, the indication was previously aortic surgery. There were 38 patients in their series that had prior aortic surgery, which leads to my first question: why was spinal catheter drainage not used in all patients with prior aortic surgery? What was the incidence of paraplegia in those patients with prior aortic surgery in whom spinal catheter drainage was not used?

The 30-day mortality of 5.4% as reported in this series is excellent, particularly when one considers that only about half of the cases were performed electively. On the other hand, the late mortality, which the authors report represents the 1-year mortality as delineated in the manuscript, was about 23% as calculated by nonactuarial methods. This is of concern. I have attempted to calculate their cumulative mortality based on the information in the manuscript, and by my estimate, the 1-year cumulative mortality approaches 30%. This does not take into account the additional interventions and other complications that required treatment. Keep in mind that a high 1-year mortality tends to improve the calculation of incidence of nonfatal events because survivors tend to have fewer of these events. Their follow-up ranges from 1 to 78 months, which leads to my next question: have the authors calculated actuarial survival for their entire series? What is the 1-year mortality as calculated by these actuarial methods?

Early and mid-term endograft infection was uniformly lethal. The authors have attempted to identify factors that would increase the risk of infection as a complication. Could the authors tell us which risk factors or findings on imaging they consider a contraindication for endovascular repair?

Several patients in their series underwent successful open surgical repair after a failed endovascular approach, which leads me to my final question: do the authors recommend primary open repair in certain cases? What are their criteria to recommend open repair as opposed to an endovascular treatment?

I would like to thank the society for the privilege of discussing this excellent experience and the authors for providing me with a copy of the manuscript prior to the meeting.

Dr Brown. Thank you again to the society for allowing us the opportunity to present our results for the treatment of thoracic aortic pathology. Many thanks as well to Dr Quinones for reviewing the manuscript and putting forth these excellent questions.

Our 30-day mortality rate of 5.4% is comparable to other reports for the endovascular treatment of thoracic aortic disease; however, it is true that we did see a high rate of death in our series within 1 year of their index procedure. The cause of death was related to the patient’s comorbidities or graft infection. We treated a high-risk group of patients with various types of thoracic aortic pathology. Those patients who were not part of an industry-sponsored trial were treated even though they may have had a limited life expectancy from other disease states. We managed these patients based upon the supposition that we could treat their thoracic aortic pathology in a less morbid manner than we could with open surgery. What is evident is that patients with thoracic aortic disease, especially those in need of an emergency operation, are amongst the highest risk patients that we care for. Even a minimally invasive operation is not without significant risks. We will include in our manuscript an analysis of our survival in the form of a life table.

With regards to our stroke rate, we acknowledge that 9% is, indeed, a high number when compared to other published series. We looked at the relationship of stroke to type of device deployed and also to the patient’s cerebral artery anatomy. We found a higher incidence of stroke in patients in which a custom-made graft was deployed. The large and inflexible delivery systems required for custom-made graft deployment cause excessive manipulation of the aortic arch, which may result in more embolic events from disruption of aortic atherosclerotic plaque and debris. This is supported by other reports that have noted similar rates of stroke when custom-made grafts are used. In our patients in whom the less bulky industry-designed devices were placed, our stroke rate was only 2.9%. We also found, not unexpectedly, that when patients had coverage of a left subclavian artery that gave origin to dominant left vertebral artery, strokes occurred. We now routinely transpose or bypass the left subclavian artery in almost all of our patients prior to intentional coverage, but especially in patients with a dominant left vertebral artery. This approach has resulted in a decrease in the incidence of posterior circulation stroke. As technology continues to improve we predict the incidence of stroke may decrease further.

In terms of paralysis, our results are on par with other reports of endovascular thoracic aortic repair. We put in spinal drains in approximately 16% of our patients. It is true that 34% of our patients had previous aortic surgery. The decision to place drains was based on surgeon preference and also depended on the emergency nature of the operation. Even though we do not use a standard algorithm for placement of spinal drains, we have turned toward placing prophylactic spinal drains in those patients in which extensive intercostal artery coverage is anticipated and/or in the patient with compromised pelvic and lumbar arterial flow from previous aortic repair. In the latter half of our series, spinal drains were placed more frequently than in the beginning of our experience.

To address your question about predicting endograft infection, we are still trying to sort this one out. Four out of the six patients who developed delayed infection had endovascular aortic repair under emergency conditions such as rupture or acute expansion with pain. In our series of patients, there were no clear-cut imaging findings that indicated primary aortic infection. In one patient, we did suspect an underlying remote infection from a prior open intervention that was present at the time of deployment. The remaining patients either had presumed device seeding from systemic infections that developed later or had unknown underlying primary aortic infections, which was what brought them to emergency repair in the first place. We caution against the use of these grafts in the face of a known infection since our experience has shown that delayed endograft infection is a highly morbid condition and in our series led to death, rupture, and/or type I endoleak in all cases. We administer antibiotic prophylaxis before every case and aggressively treat any postoperative infection in order to prevent possible delayed seeding of the endograft.

To address the final question regarding criteria for open repair vs endovascular repair, we choose the operation that is bet suited for the patient’s anatomy and clinical condition. Patients with anatomy suitable for treatment with an endograft and who were also considered relatively high risk for open repair were more likely to undergo endovascular repair. We concede that one of the limitations of this paper is the fact that we do not have an open surgical patient group for comparison. Many of our degenerative aneurysm patients were treated as part of a clinical trial, while other patients with other pathologies were treated with a minimally invasive approach because they were not thought to be candidates for open conventional repair either due to medical comorbidities or due to the clinical situation.

While there certainly are limits to the benefit gained with the endovascular management of thoracic aortic pathology and even though many patients require secondary interventions, the endovascular treatment of thoracic aortic pathology is efficacious and has decreased perioperative morbidity and mortality when compared to open surgical repair, especially in high-risk patients.