Issues in Comparative High-Risk Datasets
Roy K. Greenberg, MD, explains the Cleveland Clinic's experiences with high-risk aneurysm patients, the limitations of current comparative analyses, and the search for predictors of mortality.
Endovascular Today: How does the Cleveland Clinic's high-risk endovascular aneurysm repair dataset compare to that of the EVAR-2 study, as well as to the US IDE dataset?
Dr. Greenberg: It is difficult to compare these trials because we don't know if the patients are the same types of patients. We restricted our dataset to patients we considered "high risk" for open surgery. Many of those patients were on home oxygen, had low ejection fractions, preoperative renal insufficiency, and they may have qualified as "surgically unfit" classification in EVAR-2 as well. But there are likely others who would not qualify for EVAR-2, so it is a mixed dataset. At present, we are working on a method of developing a means to predict which patients would benefit from prophylactic AAA repair.
Endovascular Today: How do the data from the Medicare dataset compare?
Dr. Greenberg: The current Medicare dataset is very fundamentally different from our dataset, EVAR-2, and the SVS Outcomes Committee's high-risk dataset. The Medicare data, which were presented at the Vascular Annual Meeting by K. Craig Kent, MD, relate to 30-day outcome data, which is in the perioperative range rather than long-term survival. In my opinion, the question about surgically unfit or high-risk patients should not be whether or not we can get them through the procedure. From the Cleveland Clinic dataset, I think we have set the bar pretty high. Very sick patients can be operated on with an endovascular procedure with a very low perioperative risk. To achieve this requires a team approach to handle the cardiac issues, the anesthesia issues, the surgical issues, etc.
If we minimize perioperative problems, then the question that we really need to ask ourselves is, which patients are going to die from something else before they benefit from the aneurysm repair? This question has not yet been answered by any trials or datasets and is fundamental to the treatment of aneurysmal disease.
Endovascular Today: How do you think the Outcomes Committee's release of the US IDE data will impact current clinical trends regarding high-risk patients?
Dr. Greenberg: I don't think it will impact them. These data are important, but none of the studies currently being discussed answer the question of which patients should not be treated. That is the fundamental question, and it has yet to be answered by a major study, so it is left to individual physician judgment.
Endovascular Today: Should a study be conducted to answer this question, or can past data be further reviewed to make a determination?
Dr. Greenberg: I think we need to look at the data that we have, and analyze it in terms of predicting which patients are going to benefit from an endovascular repair. That has not really been done in great detail at this point. Overall, I do think there is a subtle but statistically important point that a lot of people are missing: There is a difference between associating things with mortality and actually predicting mortality. A lot of things can become associated with mortality. For example, if a patient's creatinine is high, that is associated with early death, but it is statistically different from something that is predictive of death.
Endovascular Today: Are there patient characteristics that are predictive of mortality?
Dr. Greenberg: That is the question that no one has answered. And in the Cleveland Clinic study of 412 patients (with more than 140 at 4-year follow-up), we stated that at 1 or 2 years after endograft placement, there was no single individual variable that was predictive of death. In light of this, when we look at the IDE studies, most of them involved an inclusion criteria that required the patient to have an expected 2-year survival, it seems as though we are in a Catch-22, because a 2-year survival is expected, but there is no variable that can predict that.
Roy K. Greenberg, MD, is Director of Endovascular Research, The Cleveland Clinic Foundation, Cleveland, Ohio. Dr. Greenberg may be reached at (216) 445-5306; firstname.lastname@example.org.