Improving Trial Design
Mark H. Wholey, MD, explains why patient selection and operator experience remain critical to CAS studies and the future of the procedure.
Endovascular Today: What are your thoughts on the EVA-3S trial?
Dr. Wholey: It is an interesting French trial with a major bias in favor of CEA. The major limitation to the trial was that France has been very resistant to minimally invasive CAS as a replacement for surgery. As a result, the surgeons performing the stenting had very minimal and limited training, which created a trial comparing experienced surgeons performing CEA with inexperienced operators performing CAS.
We know that with CAS, there are two critical ways to avoid stroke: patient selection and operator experience. The trial never had a chance because it included operators who had performed a minimum of three carotid stent placements in their lifetime. You cannot have someone who has done three or five carotid stenting procedures enter into a randomized trial against experienced surgeons. Quite frankly, I do not discuss the trial. I was startled that EVA-3S was accepted by peer-review in the NEJM. I do not think they had experienced reviewers, and it never should have been accepted.
Endovascular Today: How do you view the SPACE trial?
Dr. Wholey: The SPACE trial is a different situation. I thought the SPACE trial was good. It was randomized on 1,200 symptomatic patients, which is not a small number. I think one of the reasons that that trial was terminated prematurely was financing. They claimed that CAS did not meet noninferiority compared to surgery, but the bottom line is that it was an equivalence trial, and at the time that the trial was discontinued, the MAE rate for surgery was 6.34%, and stenting was 6.84%. Basically, in the symptomatic patients, both arms were statistically equivalent. To the best of my knowledge, I think those investigators were well-qualified.
Endovascular Today: What are your thoughts on the SAPPHIRE trial?
Dr. Wholey: I think the SAPPHIRE trial was a landmark trial; it was the first randomized trial that had any significance comparing high-risk CAS and CEA patients. The stenting arm, in all parameters, showed to be equivalent or better than surgery.
The subsets in SAPPHIRE were the most impressive. The diabetic subset had an impressive statistical difference in favor of CAS. Although there were only 85 patients in that subset, it was clear that stenting was so superior in all parameters—periprocedural strokes, MIs, and bleeding event rates. If you look at the endpoints, they were all clearly more established in the CAS arm than the CEA arm.
Endovascular Today: What criticisms of SAPPHIRE did you observe?
Dr. Wholey: I think there was a lot of resistance to the trial at this year's VEITH meeting based on several issues, the primary argument being that the inclusion criteria were not high-risk. However, I think they were very high-risk criteria. The patients had class 3 congestive failure, ejection fractions of 20% to 30%, and contralateral occlusions. Surgeons have a 14% stroke rate in the NASCET trial for contralateral occlusions. This tells me that these were high-risk patients. I think that chapter can be closed: carotid stenting in the high-risk asymptomatic patients is as good as surgery.
Another criticism of SAPPHIRE is that it was underpowered. I think the randomized component of it was especially good. The registry showed 7% periprocedural stroke, death, or MI. Non–Q-wave MIs were a predictor of MI for the future. The endpoints were good, but surgeons could not handle MI as an endpoint. If you look at the SAPPHIRE randomized trial, there were no major strokes. I think it was a well-run trial, and I believe the data are accurate.
Endovascular Today: Do you think the lack of major strokes in SAPPHIRE was because of the experience of the operators?
Dr. Wholey: I think SAPPHIRE had experienced operators, but CABERNET had very experienced operators. The MAE rate was 3.4% all procedural stroke, death, and MI. There is no question that if you want to include octogenarians and symptomatic patients within the trial, the number of investigational sites should be limited. For example, in CAPTURE, 144 sites were chosen for the enrollment of 3,500 patients. Obviously this number of sites resulted in operators with minimal experience. This could account for the 16.7% stroke rate in the octogenarians and 11% in the nonoctogenarian symptomatic patients.
If we want to eliminate strokes, we should do a trial of the 10 most-respected hospitals and investigators in the country. You do 400 patients, all comers: octogenarians, symptomatics, and asymptomatics. Octogenarians should not be a separate subset. We should establish which patients are too high risk for stenting. We have designated patients too high risk for surgery, but we need that designation for stenting. Then we would not have 16% procedural morbidity; we would have 5% in the symptomatic and 3% in the asymptomatic.
Endovascular Today: Why was the CEA major adverse event rate so high in SAPPHIRE?
Dr. Wholey: That rate was high because they included MI as an endpoint, with 7% incidence of non–Q-wave MIs. That is why the 30-day event rate was 7% stenting and 12% for surgery, and the 7% difference was made up of the MIs. The stroke rates for surgery and stenting were very similar, except in the diabetic subset where stenting was clearly superior.
Endovascular Today: How would you suggest determining which patients are too high-risk for stenting?
Dr. Wholey: We would say "This asymptomatic patient with a 95% lesion is too high-risk for stenting. Why? Because the patient has a type 3 or 4 aortic arch, and we were 20 minutes in the diagnostic study just getting the catheters in the complex innominate, or the complex carotid, or the bovine arch." Therefore, if it takes you 20 minutes to do the diagnostic procedure with much difficulty, you should not do the stenting. Also, octogenarians should be staged with the diagnosis done one day and stenting performed the next day.
In another scenario, you have a symptomatic patient with an ulcerated lesion, and on Volcano intravascular ultrasound (Volcano Therapeutics, Inc., Laguna Hills, CA) virtual histology you note vulnerable plaque and thrombotically active plaque, or a thin fibrous cap that could rupture when stented or ballooned. On lesion analysis of plaque characteristics via Volcano virtual histology, you can determine that the patient is not a candidate for stenting.
Endovascular Today: What effect will the trials have on the future of CAS?
Dr. Wholey: In the beginning, we thought we could stent everyone. These trials, in a certain sense, are more harmful than good because if we continue with an event rate of 10% in octogenarians, and 12% in CREST, CMS will not allow reimbursement. This will result in CAS becoming a niche procedure. That is not satisfactory.
Endovascular Today: What would be an ideal CAS trial design?
Dr. Wholey: I am convinced that we need to do a trial of ultra-experienced operators who know what patients not to stent, who know they can stent the easy patients in 20 or 30 minutes, and they know patient selection. They will not stent the patients who are risk predictors because we know from our data in Pittsburgh that patients with lesions 2 cm long or greater are definitely at higher risk than every other factor. That was our predictor. It is imperative that we draw up predictors that say, "You are too high risk for stenting."
We must also include octogenarians and symptomatic patients. I say that age should not be a risk factor if the aortic arch and the lesions are satisfactory; I think we could stent those patients. We are collecting data from five high-volume experienced centers: Lenox Hill, UPMC Shadyside, Hoag Memorial Hospital, Leipzig Heart Center, and Dortmund, Germany. We have looked at that collective data in both octogenarians and nonoctogenarians, and the all-stroke and death rates varied from 1.3% to 3.6%. There was little difference between octogenarians and nonoctogenarians in those experienced centers.
Endovascular Today: What are your thoughts on CREST?
Dr. Wholey: I think CREST is a solid trial. We desperately need CREST because that is the only way CMS is going to listen. CREST is a long way from finishing, and unfortunately, CREST has a 12% octogenarian stroke rate. They pulled octogenarians from the lead-in phase of the trial, which is not good!
Endovascular Today: What is the difference between CREST and SAPPHIRE?
Dr. Wholey: CREST is a much lower-risk trial than SAPPHIRE, and its criteria for entry are similar to the NASCET enrollment. CREST has several enrollment criteria exclusions that did not exist for SAPPHIRE. For example, CREST would not enroll patients with either ejection fractions less than 30% or recent MI.
If there are other trials running when CREST is running, CREST may not finish. CREST is gaining momentum and is doing well–they have 1,500 patients enrolled. When they started enrolling the asymptomatic patients, it helped the trial. We need CREST because it will give us an indication of whether we can do low-risk symptomatic and asymptomatic patients with a 5% to 6% periprocedural event rate in the symptomatic group and 3% in the asymptomatic population.
Endovascular Today: When can we expect data from CREST to affect procedure?
Dr. Wholey: It will be a couple of years, minimum. With our regulatory policy, once CREST completes, it will take an additional year to compile the 1-year data. It is a 2,400 patient trial, so it will be at least another 2 years. It will then require submission, FDA panel review, and ultimately a CMS approval process. Unfortunately, we are terribly over-regulated by the FDA, CMS, IRB, and considering that CAS represents a high-risk procedure, there is always the threat of malpractice. We are not getting the choice of devices that are available in Europe. We have distal protection devices in this country, but we do not have flow reversal or proximal flow control, and we should have both of these. We do not have a choice of stents and, if a promising one is introduced, it is 3 or 4 years before it is available for clinical application. The entire process moves very slowly.
Endovascular Today: How does the protocol for ACT differ from the other trials?
Dr. Wholey: ACT is a low-risk, asymptomatic-only trial with 3:1 randomization, which means that one patient is randomized for every three who are stented, and it is enrolling quite slowly–at least here in Pittsburgh–and it is totally dependent on the Xact stent and the Emboshield filter (Abbott Vascular, Santa Clara, CA). We are bound to that system. ACT-1 basically competes with CREST, yet they are ongoing because the companies want approval for a low-risk reimbursement. I have not yet seen any preliminary data yet.
Endovascular Today: What is the best way to compare these trials?
Dr. Wholey: Variations in the enrollment criteria existed among the trials. ARCHeR, for example, did not include contralateral strokes and death at 1 year, but SAPPHIRE was somewhat more inclusive and did include all strokes and MI at 1 year. There were variations in the endpoints of the trial. Again, although these were not overwhelming, they may have had an influence on the overall outcomes. The 30-day endpoints were quite similar, but the 1-year endpoints were different. Questions to ask are, "What was the periprocedural stroke and death rate? What was the major stroke event rate?" Certainly minor strokes are important, but most minor strokes will return to some degree of normal functionality. Unfortunately, major strokes are disabling and rarely ever return to baseline.
Endovascular Today: How do you perceive the surgeon's perspective regarding CAS?
Dr. Wholey: Surgeons who extensively perform CEA do a very good job, so they are fairly satisfied. Surgeons are not going to participate in randomized trials unless they can also do stenting. However, surgeons who are not involved in endovascular stenting have no incentive to enroll because they are getting good results with surgery, and they will not enroll their patients with stenting when they look at the data. The data show them that stenting is no better than surgery and may be worse in the symptomatic patient. Furthermore, those of us who support CAS would like better data from the symptomatic and the octogenarian patients.
Endovascular Today: What does the future hold for CAS?
Dr. Wholey: We have to be patient in the learning process of carotid stenting. We must remember that the first aortocoronary bypass procedures were not that smooth either. The technology is evolving, and the filters, recovery systems, flow reversal, and operator skills will improve with time. I think it is a matter of time before we work out the problems with octogenarians, but we have to be patient; it is a new procedure, in only its second generation. It will take time before the experienced operators lay out the criteria and then even more time to get regulatory approval.
I think there is a lull in CAS because those of us who have done more than 1,000 cases are tired of talking about it, and we are tired of the continual battles with CMS and the FDA, as well as sending patients home with critical lesions that are preocclusive but who unfortunately do not meet the entry criteria for the category B IDE trials. Consequently, these patients are being discharged with a stroke warrant. It is not that we are losing interest so much as saying that we are becoming somewhat disenchanted with the overall process.
Endovascular Today: What can operators do to improve their CAS results?
Dr. Wholey: If these trials keep producing the data that they do with symptomatics and octogenarians, it is more harmful than good. Opening these trials to 100 sites and 4,000 to 5,000 patients is not a good idea. Initially, you would like 10 high-volume, experienced operators to participate in a 500-patient trial that includes all levels–symptomatic, asymptomatic and octogenarians–and come in with a 6% or better all-stroke event rate. I think we can show that the best operators can achieve these good numbers, which will help us conclude that others can do the same after responsible credentialing. Inexperienced operators should be willing to spend more time in the training process. The problem may be that single operators are not yet good enough in the early stages of CAS; maybe it takes dual operators. We have to tighten up the whole process to include training and credentialing.
Mark H. Wholey, MD, is a Clinical Professor of Radiology at the University of Pittsburgh School of Medicine, UPMC Shadyside Hospital, and Chairman of the Pittsburgh Vascular Institute in Pennsylvania. Dr. Wholey may be reached at (412) 623-2083; firstname.lastname@example.org.