LOS ANGELES - Carotid artery bypass surgery was no better than medical therapy in the secondary prevention of ischemic stroke, despite excellent graft patency at up to 2 years, in the Carotid Occlusion Surgery Study.
The rate of subsequent ipsilateral stroke was almost exactly the same at 2 years, regardless of whether patients were randomized to the surgical or medical arm of the Carotid Occlusion Surgery Study (COSS), Dr. William J. Powers said at a late-breaking science session during the International Stroke Conference.
"In spite of excellent graft patency and a significant improvement in cerebral hemodynamics, the bypass surgery failed to provide an overall benefit on 2-year stroke outcomes," said Dr. Powers, chairman of the department of neurology at the University of North Carolina at Chapel Hill.
The failure of surgery to improve outcomes relative to medical treatment was probably due to a much lower than expected rate of subsequent stroke among patients in the medical group (23%), compared with those in the surgical group (21%), Dr. Powers said. The investigators had projected a subsequent stroke rate of 40% for the medical therapy group and 24% for the surgical group, based on results of the St. Louis Carotid Occlusion Study and the Extracranial-Intracranial (EC-IC) Bypass study.
Investigators began COSS in 2003 but terminated it in June 2010 when it became obvious that surgery would confer no additional benefit over medical therapy. "The Data Safety Monitoring Board found that the prespecified futility analysis point was reached when we had full 2-year follow-up data on 139 of our patients," Dr. Powers said. "At that point, it was highly unlikely that the study would show any significant difference between the groups."
The COSS cohort included patients at very high risk for a second stroke. Patients had to have experienced a transient ischemic attack or a mild to moderate stroke within 3 months of enrollment and symptomatic occlusion of the internal carotid artery. Patients also were required to have evidence of hemodynamic cerebral ischemia, as shown by an ipsilateral increased oxygen extraction fraction (OEF) ratio of greater than 1.130 on PET. "This [ratio] has been determined previously to put patients at a very high risk of subsequent stroke," Dr. Powers noted.
The trial compared outcomes in 98 patients who were treated by best available medical treatment with antithrombolytic agents and risk-factor intervention vs. 97 patients who received the same medical therapy with the addition of superficial temporal artery to middle cerebral artery bypass. The primary end points for the surgical group were any stroke or death from surgery within the first 30 days after randomization and ipsilateral ischemic stroke by 2 years. The medical therapy group had the same 2-year end point, with a 30-day end point of all stroke and death during the first month after randomization.
Dr. Powers did not provide the baseline characteristics of the group, except to say that the systolic blood pressure was slightly lower in the surgical group.
In the surgical group, 93 were operated on within the first 10 days of randomization. None experienced a stroke during that intervening time. In all, 14 strokes (15%) occurred within 30 days of the procedure. "This was not significantly different [from what] was seen in the EC-IC Bypass trial," Dr. Powers noted. He did not provide information on surgical complications, except to say that "the perisurgical complications were what we expected and in line with the EC-IC Bypass trial."
At 30 days, graft patency was 98% and the mean OEF ratio had improved from 1.258 to 1.109. "So the surgery did what we thought it would do: It dropped the OEF ratio," he said.
At 2 years, graft patency was 96%. By this time, subsequent ipsilateral stroke had occurred in 21% of the surgical cohort, which was similar to the expected 24% rate.
Dr. Powers did not present 30-day outcomes in the medical therapy group. However, by 2 years, subsequent ipsilateral stroke had occurred in 23% of the group, "far lower than 40% rate we projected based on historical controls, and not significantly different from the surgical group," he said.
Following Dr. Powers’ presentation, several audience members asked for additional information, including any possible explanation of the unexpectedly good results in the medical therapy group. "Was it because they were all on statins, or because their hypertension was better controlled?" asked Dr. Gary Steinberg of Stanford (Calif.) University. "A theme we’ve heard over and over at this stroke meeting is that we can’t rely on historical data anymore because the medical treatments are getting so much better."