February 24, 2010
European ICSS Study Concludes CAS Not as Effective as CEA in the Short-Term
February 25, 2010—Findings from the International Carotid Stenting Study (ICSS) suggest that in the short-term, carotid artery stenting (CAS) was neither as safe nor as effective at preventing stroke as carotid endarterectomy (CEA). The Lancet and The Lancet Neurology published the ICSS study and a substudy, respectively, online ahead of print that concluded that CEA is safer than CAS as a treatment for carotid blockages that can lead to stroke. Detailed abstracts of the studies are available at the above links to the journals’ Web pages.
The studies showed that CEA reduces the risk of stroke and death approximately in half within 30 days of treatment compared to stenting and significantly reduces the risk of stroke, death, and treatment-related myocardial infarction (MI) in the first few months after diagnosis. ICSS was funded by the United Kingdom’s Medical Research Council, the Stroke Association, the European Union, and Sanofi-Synthélabo (now Sanofi-Aventis, Paris, France). Professor Martin M. Brown, MD, was the chief investigator of ICSS.
The ICSS investigators assessed the effectiveness of CAS and CEA for preventing stroke, death, and procedure-related MI in 1,713 patients with recently symptomatic carotid stenoses. Patients were randomly assigned to stenting (n = 855) or surgery (n = 858) and were followed for up to 120 days after randomization.
The investigators reported that patients in the stent group had a significantly greater risk of stroke, death, or procedure-related MI within 120 days of randomization—the absolute risk was 8.5% compared with 5.2% in the surgery group. Additionally, within 30 days of treatment, the rate of stroke or death in the stent group was nearly twice the rate recorded in the surgery group. This difference was mainly due to a higher number of nondisabling strokes recorded in the stent group (36 vs 11 within 30 days of treatment), whereas the number of disabling strokes or deaths did not differ significantly (26 vs 18).
The ICSS investigators concluded, "Completion of long-term follow-up is needed to establish the efficacy of treatment with CAS compared to CEA. In the meantime, CEA should remain the treatment of choice for symptomatic patients suitable for surgery."
The Lancet noted that because investigators who assessed patients during follow-up were not blinded to treatment allocation, there was the possibility of ascertainment bias that might explain the higher incidence of nondisabling strokes found in the stent group. Therefore, in a substudy of ICSS published in The Lancet Neurology, Leo Bonati, MD, et al conducted a blinded analysis of the ICSS trial to assess the rate of ischemic brain injury on magnetic resonance imaging (MRI) after treatment in the two groups. Of the 50 centers in ICSS, seven took part in the MRI substudy, which included 124 CAS patients and 107 CEA patients.
Three times as many patients in the CAS group had new ischemic lesions on diffusion-weighted imaging on posttreatment scans than in the CEA group; therefore, the investigators concluded that the difference in clinical stroke risk in ICSS is unlikely to have been caused by ascertainment bias. Protection devices did not seem to be effective in preventing cerebral ischemia during stenting. Diffusion-weighted imaging might serve as a surrogate outcome measure in future trials of carotid interventions.
The ICSS MRI investigators stated that these neuroimaging data confirm that CEA is safer than CAS and that the increased risk of nondisabling stroke after stent treatment is unlikely to have been caused by ascertainment bias. Most of the new ischemic lesions were not associated with symptoms of stroke at the time of stenting, and the investigators suggest that the most likely explanation is that particles of atheromatous plaque are released during implantation of the stent, which are too small to cause any noticeable symptoms at the time, but nevertheless lead to small areas of brain damage.
Leading carotid interventionists discussed the contrasting findings from CREST and ICSS with Endovascular Today. According to interventional radiologist Barry T. Katzen, MD, who is a CREST investigator, one of the major differences between CREST and ICSS was the training and qualification of the interventional operators. "CAS is an operator- and learning curve-dependent procedure, which was demonstrated in the difference in outcomes between the two trials. The CREST qualification process was rigorous and outcomes based, with many operators being excluded and others being required to increase their experience and demonstrate proven low morbidity before being accepted as randomizing physicians. Additionally, only 72% of patients with CAS in ICSS had embolic protection. Clearly, less experienced operators performing CAS with low rates of embolic protection will produce outcomes that are poorer than those produced in CREST, which had the lowest event rates in both arms of any published trials. European sites should have a greater emphasis on training and experience." Dr. Katzen is co-chief medical editor of Endovascular Today.
Interventional cardiologist Christopher J. White, MD, noted that both CREST and ICSS showed no differences for their primary endpoints between surgery and stenting. In more than 800 patients, ICSS surgeons reported 0.5% MI, whereas CREST surgeons reported 2.3% MI in more 1,200 patients. These data suggest that ICSS does have an ascertainment bias, Dr. White stated to Endovascular Today.
Furthermore, ICSS considered interventionists to be experienced if they had placed a total of 10 carotid stents in their careers, whereas surgeons were required to have done 50 operations. Dr. White pointed out that two inexperienced interventionists who were being tutored caused five major strokes in 11 patients, which were counted in the final tally. By contrast, three-fourths of CREST investigators qualified for that study through a roll-in program that made sure that learning curve issues had been dealt with.
"At the end of the day, both ICSS and CREST support clinical equipoise between surgery and stenting," concluded Dr. White. "Patients and their doctors should make informed decisions about whether a stent or surgery is best for them."
Interventional neurosurgeon L. Nelson Hopkins, MD, a CREST investigator, stated that although ICSS was well run by experienced investigators, credentialing was significantly less rigorous in ICSS than in CREST. Dr. Hopkins also observed that embolic protection was only recommended and used in 72% of procedures in ICSS compared with CREST, which mandated embolic protection with forced lead-in experience.
"CREST shows that CAS has reached parity with CEA, that the two procedures are complimentary, and that further study of the CREST data set will suggest how to improve patient selection," concluded Dr. Hopkins. "So hopefully, the Centers for Medicare & Medicaid Services will give us the opportunity to offer both procedures."