ARTICLE IN BRIEF
Investigators reported that about 22.7 percent of patients treated with medicine alone experienced an ipsilateral stroke within two years of being enrolled in the study, compared with 21 percent who received the bypass surgery in addition to best medical care — a difference that was not statistically significant.
LOS ANGELES—Everything about the trial assessing the controversial extracranial to intracranial (EC/IC) bypass surgery worked well: The operation was performed with minimal technical problems; the grafts used in the bypass procedure remained patent; the morbidity expected for the surgery was just about what was expected.
The problem: Patients on medical therapy alone did much better than expected – so much better, in fact, that the differences in two-year ipsilateral strokes between those who had surgery in addition to the best medical treatment were indistinguishable from patients who were treated medically alone.
The EC-IC bypass aims to improve outcomes by increasing cerebral blood flow. The procedure — similar to cardiac bypass surgery in which blocked heart arteries are bypassed — entails connecting a branch of the external carotid artery to a branch of the internal carotid artery, either directly or via a vein graft.
“We had anticipated that, based on historical data, the event rate among patients on medical therapy would be about 40 percent,” said William J. Powers, MD, the H. Houston Merritt Distinguished Professor and Chairman of Adult Neurology at the University of North Carolina in Chapel Hill. Dr. Powers described the outcomes data of the Carotid Occlusion Surgery Study (COSS) here at the American Stroke Association International Stroke Conference in February. The researchers had anticipated an event rate of 24 percent in the surgically-treated group.
COSS was a randomized controlled, blindly adjudicated clinical trial to test whether EC/IC bypass surgery, when added to best medical therapy, could reduce subsequent ipsilateral ischemic stroke at two years. The patients who were enrolled had recent symptomatic internal carotid artery occlusion and ipsilateral increased oxygen extraction fraction measured by PET.
As it turned out, about 22.7 percent of patients treated with medicine alone experienced an ipsilateral stroke within two years of being enrolled in the study. Among the patients who received the bypass surgery in addition to best medical care, 21 percent experienced an ipsilateral stroke, Dr. Powers said in his oral presentation. The difference between the groups was not significant (p=0.88), he said.
To be eligible for the study, patients had to demonstrate ipsilateral hemisphere carotid territory transient ischemic attack or mild-to-moderate ischemic stroke within 120 days, as measured by the Barthel Index assessing daily living skills.
The primary endpoint of the study in the medical group was all stroke and death from randomization through 30 days plus ipsilateral stroke within two years, and the primary endpoint for the group that had the bypass was all stroke and death from surgery through 30 days postsurgery plus ipsalateral stroke within two years. Dr. Powers said that almost all the events were ipsalateral strokes in two years.
The study began recruiting patients in July 2002. The Data Monitoring and Safety Board stopped the trial in June 2010 when the pre-specified boundary for declaring futility had been crossed based on an analysis of the 139 subjects who had been followed for a full two years. Based on previous studies, the St. Louis Carotid Occlusion Study and the EC/IC Bypass Trial, the researchers calculated that 372 randomized patients were necessary to show the benefits of the surgery.
The monitoring board noted that, “given the unexpected low rate of primary endpoints in the medically treated group, a clinically meaningful difference in favor of surgery would not be detectable without a substantial increase in sample size.”
To perform the study, researchers screened 4,967 patients and enrolled 705 of them for the PET scans that were required for entry into the study. Eventually, Dr. Powers noted, 195 patients were randomized: 98 to receive medical care and 97 to undergo surgery on top of best medial treatment.
In the medical group, three of the patients crossed over to receive surgery; four individuals in the surgery group decided against having the procedure.
Among the 93 patients who underwent the surgical procedure, the operation was performed within two weeks of randomization. No strokes occurred in the interval. Fourteen strokes occurred within 15 days of the surgery, six additional strokes occurred within two years of follow-up. The 30-day graft patency was 98 percent and the grafts were still patent at the time of last observation in 96 percent of those who had surgery. The oxygen extraction fraction PET ratio improved as well, Dr. Powers said.
“In spite of excellent graft patency and improved cerebral hemodynamics, EC/IC bypass surgery failed to provide an overall benefit on two-year stroke recurrence due to the much better than expected recurrence rate in the non-surgical group,” Dr. Powers said.
After Dr. Powers concluded his oral report, doctors in the audience suggested that the reason the medical group did better was due to improvement in medical therapies over time, rendering the hypothesized rate of 40 percent obsolete. But when Dr. Powers was asked to provide data regarding specific medical treatment such as how many of the patients in the medical group were taking statin therapy, he demurred. He said those types of data were collected but would be presented at later conferences.
Commenting on the study for Neurology Today, Larry B. Goldstein, MD, professor of medicine and director of the Duke Stroke Center in Durham, NC, said: “What we have learned from this trial is that historical data may be an improper surrogate marker for the basis of clinical trials.”
In the discussion of Dr. Powers' presentation, audience members related that some of the most serious patients were not eligible for the study because of the requirement for PET scanning. Dr. Goldstein, however, said: “The study was designed to treat the sicker patients and the treatment failed to help them. I don't know that surgery would have helped these patients that were discussed in the audience.” He suggested that the sickest patients who might have benefited from surgery might not have survived the procedure.
Dr. Goldstein said it would be difficult to find a population of patients who might ultimately benefit from the surgery. “We now have two trials that have shown that this type of surgery is not effective in reducing events when compared with best medical care,” he said.
He noted that the study had achieved its surgical goals. “The procedure improved perfusion but that did not translate to better outcomes. This points out the danger of using surrogate markers,” Dr. Goldstein said.
The study, which was funded by the NIH, was conducted in 48 medical centers in the United States, and one in Canada. •