Dr. Kim J. Hodgson (Springfield, Ill.).

The authors present a retrospective review of 267 patients who underwent superficial femoral artery balloon angioplasty, concentrating on the outcomes of a subset of 58 patients who were treated with intravascular stents for suboptimal initial technical results of balloon angioplasty alone. Complete occlusions were present in 89% of the treated limbs, and 50% of the lesions were in excess of 15 cm in length, the longest being 35 cm, with an average length of 16.5 cm. In our own endovascular practice, we long ago abandoned dilatation of most lesions longer than 10 cm, especially complete occlusions, because of similarly poor long-term results to those reported today.

This gives rise to my first set of questions. Why were patients with such severe disease selected for balloon angioplasty rather than surgical reconstruction, which can be performed under local anesthesia, if need be, to minimize operative risk? Was it realistic to expect to successfully treat such lesions without supplemental stenting, a procedure known to be thrombogenic in this location?

To ameliorate the thrombogenicity of the stents, you adopted a policy of long-term anticoagulation therapy in all of your stented patients. We have observed late stent thrombosis in patients who had superficial femoral artery stents who have discontinued their warfarin therapy or whose prothrombin times fall below therapeutic levels. Were all of your patients therapeutically anticoagulated at the time their interventions thrombosed?

One of your stent implantation criteria was a pressure gradient across the lesion in excess of 15 mm Hg after balloon angioplasty alone. In at least 90% of your cases, the measurement of arterial pressure downstream of the dilated lesion was achieved with a catheter that traversed the area of questionable residual stenosis from the upstream side, possibly partially obturating the vessel and creating an artificial pressure gradient. We have found the absence of a pressure gradient measured in this way to be reassuring, but have been unwilling to trust a positive determination for this reason. Can you tell us how many limbs were judged to have had suboptimal additional dilatation on the basis of a pressure gradient measurement alone? Do you think that the pressure gradient in these limbs may have been artificially induced and that they may have fared better without having been stented?

Acknowledging that dilatation and stenting of vessels with such severe disease is not likely to produce long-term benefit, there is a flip side of the argument to consider. In your manuscript you report that 17 limbs were censored because of the death of the patients, only one of which appeared to be directly related to the procedure. Depending on whether the patients who died underwent single-limb or dual-limb procedures, this translates into a mortality rate of 25% to 31% over the average follow-up period of 13.8 months; a rate so high that long-term success may not be a critical issue.

Nonetheless, at the mean follow-up period of 13.8 months, 59% of the limbs were able to maintain an increase of at least 0.15 in their ankle-brachial index, and 56% had persistence of at least one clinical category of improvement. How do you reconcile these observations with the dismal patency results you have reported? Could it be that your duplex scan determinations of a >50% restenosis are flawed, or might a 50% restenosis, as determined by duplex scanning, be clinically or hemodynamically insignificant?

In the final analysis, the data presented appear to support our long-standing belief that even stents cannot substitute for clinical judgment when it comes to determining which atherosclerotic lesions are suitable for presently available endovascular interventions.

Dr. Bruce H. Gray.

Thank you for your comments. As you know, we have excellent surgical coverage at the Cleveland Clinic. These patients had profound levels of ischemia, and because of multiple medical problems our first option was to pursue percutaneous treatment. If successful, we could obviate the need of a surgical procedure. Unfortunately, what we have noticed through the course of the study period is that these patients frequently have recurrent disease. The purpose of these data is to show that the percutaneous alternative to long-segment superficial femoral artery disease is not a durable procedure.

Our perception with these procedures is that the morbidity and mortality rates were low. These data certainly cause us to stop and reflect on actually how tough these procedures are in these patients. Most of them had poor underlying cardiac status and truly were not good surgical candidates.

Your comment about an artificial pressure gradient being induced by placing a 5F catheter across the lesion after angioplasty is performed is a good one. After successful dilatation of the superficial femoral artery to greater than 6 mm, one should be able to reduce the pressure gradient below 15 mm Hg even with a 5F catheter across it. If one uses the more liberal criteria for stent placement, that being a 5 mm residual pressure gradient, that is hard to achieve with a 5F catheter across the lesion. Therefore, this is less of an issue because of the 15 mm Hg gradient threshold.

The use of anticoagulation therapy may be a moot point in these patients because the underlying problem is predominantly that of intimal hyperplasia going on to significant restenosis and reocclusion. Warfarin does not impact the process of intimal hyperplasia. There are other agents that may be more beneficial, such as glycoprotein IIB/IIIA inhibitors. I think that it is going to take something dramatic to be able to halt this substantial intimal hyperplastic response of this active artery to the trauma that we induce.

Also, your statement that 17 grafts were censored because the patients died is inaccurate. “Censored” means that the patient did not reach an end point at the time of latest follow-up.

Dr. Richard L. McCann (Durham, N.C.).

My question relates to the use of warfarin. Our cardiologists have found in the coronary circulation that warfarin is dramatically inferior to the use of aspirin and Ticlid in terms of long-term patency of stented patients. You used warfarin in the majority of your patients. Do you have any evidence that it is either beneficial or harmful in this vascular bed?

Dr. Gray.

The restenosis rate seen in this series far surpasses any restenosis rate that is reported in the coronary literature, as far as I know. This diffuse, long intimal hyperplasia segment disease is far more substantial than what aspirin and Ticlid or warfarin can avoid.

Dr. Geoffrey H. White (Sydney, Australia).

Your accumulated experience of femoral stenting procedures internationally suggests that they are not of use in patients who have long-segment disease, in whom multiple stents are used with poor runoff, limb-threatened patients, high-risk patients, and that the use of warfarin only serves to increase the complication rate. The data that you have presented suggests that this is precisely the patient population in whom you are using this technique. So first, I would like to ask why you have reserved recanalization and stenting precisely for those patients in whom the information so far suggests that it’s not of use? As a corollary, I would like to ask what treatment you use for the patients who are good risks in whom you might expect this treatment to actually work?

Secondly, I would also like to ask a question regarding restenosis. We have found that restenosis as seen on duplex scans often remains relatively asymptomatic and that the asymptomatic status of these patients may persist for several years despite duplex findings that show apparent restenosis >50% or even >75% within the stent. So I would be interested to hear whether your experience was similar to our experience or whether restenosis was always associated with recurrent symptoms in your study.

Dr. Gray.

Our clinical patency rate at latest follow-up was 56%, as compared with our anatomic primary patency rate of 22%. Our criteria of restenosis can account for our high prevalence of reaching the end points that were defined. That may attest to the fact that restenosis may or may not be clinically significant in all cases.

There are previous data, as you know, that anticoagulation therapy may not be necessary for performing superficial femoral artery stenting; however, that is in much shorter disease segments. Long-segment disease throughout the superficial femoral artery is much more prone to restenosis and failure of our endovascular procedures. This sets the stage for the next generation of percutaneous procedures, such as stent-grafts or coated stents.

Dr. Samuel S. Ahn (Los Angeles, Calif.).

I rise to congratulate and compliment the authors for their willingness to bring these negative results to our attention, and also to congratulate the Program Committee for allowing them to do so. I think such reports are very important. However, I also rise to criticize the methods used in this study, the fact that it was done in a retrospective fashion using off-label products in a nonapproved fashion.

The literature is replete with data that show that PTA for stenosis greater than 7 to 10 cm has very poor results; yet the authors treated such unfavorable lesions, and then stented them secondarily in attempt to bail out or improve their situation. In the future, such studies should be conducted in a prospective fashion under an FDA-approved IDE. Such behavior will provide us much more responsible studies, better scientific data, and better care of our patients.

Having said that, I would like to ask some questions. I’d like to find out more about the mechanism of the restenosis. I think 18% of your patients died; do you have any autopsy specimens of the treated lesions? Also, what were the locations of the restenosis, were they in the body of the stent or at the distal or proximal ends?

Secondly, I would like to ask you about the technique that you use. When you performed balloon-dilation before the stent went in, how many dilatations did you perform, and at what pressures? In my own experience I have found that if you keep ballooning multiple times or at higher pressures, or for a longer duration, you can often get rid of the gradient. Did you do that, or did you stent after only one or two inflations at a fairly low pressure?

Dr. Gray.

Technically, we used high inflation pressures up to 18 or 20 atmospheres. Preprocedural anticoagulation with heparin allows for prolonged balloon inflation (as long as 5 minutes) without much risk of procedural thrombosis.

Restenosis is seen not only at sites of stent overlap but also throughout the body of the stent. Perhaps it may be a result of a gap present between the stent and the actual vessel wall. With the deployment of a long stent in the superficial femoral artery followed by subsequent dilation, these stents may not become fully adherent to the vessel wall, which would predispose to restenosis.

Dr. James C. Stanley (Ann Arbor, Mich.).

Dr. Gray, I have two comments. One, I have not seen a paper that has damned a technology like this since a paper was presented at these meetings that dealt with bovine carotid grafts. I doubt that very many people returned home and used that conduit for arterial reconstructions after that, other than for arteriovenous fistulas.

The second is, just a response to Dr. Ahn, I agree that if someone had this technology at hand and could have done a prospective study, there may have been a different outcome in the number of patients in many centers that were treated with this technique. But it reminds me of a past president before Frank Veith, in a discussion of the rather nonspecific nature of trying to identify people at coronary risk undergoing vascular surgery, and someone stood up and talked about multiple gated acquisition scans at one of the meetings, and he said, “You’ve got very interesting data. I’m not sure why you did this.” He said, “If you drive your car up to the parking structure and you’ve got 20 of them behind you and you all drive off the edge, you’ve got a mess down in the bottom. You don’t need to do a statistical analysis of it.”

I must say, I have not seen dismal results worse than this, and I’m not sure a prospective randomized study, if one has preliminary efficacy data, or safety data, would allow this to go to a randomized study. Many of us who sit on the Joint Council of these two societies have great certainty about similar technology being applied to carotid stenting with or without balloon angioplasty at this time. So I think there has to be some safety data. I would say that if someone did a preliminary study with this particular procedure, the safety would not be there to justify a prospective study.