Discussion

Article Outline

• Copyright

Dr Jeffrey P. Carpenter (Philadelphia, Pa). Drs Dillavou and Makaroun have enlarged our understanding of thoracic aortic repair by identifying some preoperative risk factors for morbidity and mortality after both open and TEVAR [thoracic endovascular aneurysm repair] approaches using the TAG [W. L. Gore and Associates, Flagstaff, Ariz] database. They have identified a number of risk factors, which unfortunately cannot be modified, that predict bad outcomes: gender, height, renal function, presentation, and age, among those.

I would ask, therefore, what is our take-home message regarding patient selection for open or endovascular repair from these data? How would you reduce these findings to actual practice? Can you tell us a little more about the endoleak rate of 10.6%? Are these mostly type II leaks, or are they attachment or junction leaks? Also, could you comment on that sac expansion rate of 19%. Is that endotension as we saw with the abdominal excluder, or is this a different problem with the TAG? What have your investigators done about that issue, and what is recommended?

Also, I am fascinated by your finding in the manuscript, which you provided to me, that the short proximal neck length predicted perioperative morbidity. Perhaps that is a modifiable risk factor. What are the practical implications of this short neck finding related to morbidity?

And then, finally, in light of the finding of prospective randomized trials of abdominal aneurysm repair vs open or endovascular means, specifically that there is no long-term mortality benefit to EVAR when compared with open aneurysm repair. Payment has been denied for EVAR in parts of Europe. EVAR is under very careful scrutiny by payers in this country as well.

Now, I know the TAG database is not a randomized study of open vs endovascular repair of thoracic aneurysms, but it remains the best available data set we have for comparison between open and endovascular groups. Do you have any glimpse of the long-term comparison of mortality between those groups? Does the mortality benefit in the perioperative period that you noted extend to the longer term?

Dr Dillavou. Thank you, Dr Carpenter. Going through the questions one by one, you ask an excellent question about what is the take-home message from these data. I agree that some factors cannot be changed. However, I believe that we can use this information to educate patients and to enhance patient selection. Some proactive measures can also be taken. For instance, if we know that women have a higher chance of vascular complications after TAG endografting, perhaps planning a conduit is the better choice rather than just deciding to forge through.

You ask about the endoleak rate of 10.6% at 5 years. We found that most of these were attachment site leaks, and they were fixed relatively easily over time. Only three patients have had to undergo conversion, two were for aortoenteric fistulas and one was for proximal migration and dilation of the proximal neck zone. None of the conversions were for endoleaks. All were able to be treated percutaneously.

The sac expansion rate of 19% at 5 years is something that deserves attention. A few patients had endoleaks, but the vast majority of them did not. We attribute sac expansion to the same forces and porosity that we see in the abdominal group. All of these TAG grafts that were put in were the standard expanded PTFE [polytetrafluoroethylene] grafts.

Comparing these grafts to the low-porosity TAGs showed differences that did not achieve statistical significance; at 2 years in the low-porosity group, we saw a 3% sac expansion vs a 13% sac expansion in the regular group. We hope that the low-porosity graft will decrease sac expansion. Thus far, no action has been taken on these patients if no endoleak is found.

You ask about short proximal neck length related to perioperative morbidity. We have found that with carotid–subclavian bypass, there is a statistically higher incidence of stroke after endografting. This is true whether or not the carotid–subclavian bypass was done preoperatively or immediately postoperatively. We hypothesize that this is due to a more diseased arch and more manipulation. All strokes were perioperative events. This is something to take into consideration when planning the operation and educating the patient.

Finally, you ask about our results and compared them to the EVAR I and II trials done in the United Kingdom. We found that there was a significant aneurysm-related survival difference, with a higher aneurysm-related survival in the thoracic endograft group compared to the open surgical group. However, looking at overall mortality, there was no difference between the groups at 5 years. So there is a difference in aneurysm mortality, and it persists to 5 years.

Dr Keith D. Calligaro (Philadelphia, Pa). Dr Dillavou, very nice presentation. I know this was retrospective and you can't really match up the patients. The question is whether patients who underwent open surgical repair did so because they did not have an adequate neck proximal or distal? If there wasn't an adequate landing zone for an endograft, more of the patients undergoing open repair might require clamping above the subclavian or below the viscerals and that alone would increase the morbidity.

Dr Dillavou. No. The clamp site limits were basically the same for both groups, that the proximal clamp could not have been placed more proximal then the left carotid artery and the distal clamp could not be placed more distally than the celiac trunk. So the clamp sites were the same. Some of these were historic controls that were done prior to the availability of the endovascular graft.

Dr Maciej L. Dryjski (Buffalo, NY). I have a question regarding the mortality. How would you explain the higher mortality in the symptomatic group of patients? Is it perhaps that the symptoms were not related to the aneurysm itself, but rather some other medical condition?

Dr Dillavou. To make sure I understand your question, you are asking if I believe that the increased mortality in the symptomatic aneurysm group was because of hemodynamic instability and presentation?

Dr Dryjski. Were these patients hemodynamically unstable, or were the symptoms related to some other condition, not necessarily an expanding aneurysm?

Dr Dillavou. These were not hemodynamically unstable patients. There was no one with a rupture or a leak or anything like that. These were patients who presented, whom the surgeon felt were having pain or other symptoms due to the aneurysm, and in most cases they were done more urgently than a traditional planned aneurysm repair. It is my hypothesis that part of the reason that these patients have an increased mortality is because that their operations probably were less electively planned. We did not find any difference in operating room blood loss or other intraoperative factors, hematocrit values, things like that.

Procedure times were slightly longer for symptomatic vs nonsymptomatic aneurysms, but there again, that just may be related to the more difficult nature of the anatomy. There is no concrete reason why the symptomatic aneurysm group had such a dramatic increase in mortality, but my hypothesis is just that the patients were not optimized prior to their repairs. Also, some of these patients may not have been offered elective surgery due to anatomic or medical factors, but because they presented symptomatically they underwent a repair.

Dr Marat Goldenberg (Reading, Pa). I have a two-part question. You said that in women the operative time was longer. Question number 1, do you think this is due to the smaller access vessels or unique characteristics of the aneurysm? Do you think they have higher level of complication because of the longer time under anesthesia or the degree of difficulty?

Dr Dillavou. That's an excellent question. I believe that the longer procedure times were probably related to smaller access vessels and also to complications that were encountered in the operating room. More than half of the vascular complications that we saw were actually dealt with at the time of the procedure—iliac artery ruptures, femoral artery occlusions, et cetera—and so this obviously adds to procedure time. It is my belief that this, the difficult access, is the cause and the long procedure time is the effect, although it is impossible to know for sure.

That is something that we struggled with statistically in this database, and that is that long procedure times, no matter what we were analyzing, long procedure times predicted bad outcomes. But we also saw predictably longer procedure times with open surgery and in patients with complications, which all sort of goes together. And so it is my hypothesis that the long procedure times were an effect rather than the cause of the morbidities.