BACKGROUND: Carotid endarterectomy (CEA) benefits patients with symptomatic carotid stenoses. However, given the perioperative morbidity, mortality, and cost of the procedure, it is unclear whether patients with carotid stenosis and no symptoms benefit. The authors of this meta-analysis reviewed trials on the topic to look for an overall answer.
POPULATION STUDIED: The patients studied had carotid stenosis of any degree but had never had symptoms such as transient ischemic attack (TIA) or cerebrovascular accident. Stenosis was diagnosed using a wide variety of modalities. Of the 2206 patients identified in all studies, 73% were men, and the average age was 66 years.
STUDY DESIGN AND VALIDITY: This was a Cochrane meta-analysis of randomized controlled trials of CEA versus medical management for patients with asymptomatic carotid stenosis. Studies were identified by searching MEDLINE, Current Contents, and the Cochrane Stroke Group Trials Registry. Experts in the field were also contacted to identify any additional published or unpublished trials. The diagnostic criteria for carotid stenosis varied widely among the studies and included angiography, oculoplethysmography, and clinical carotid auscultation. All surgical approaches to CEA were grouped together, while medical management was essentially watchful waiting. Depending on the study, aspirin was given to no patients, to only the medical management patients, or to both the medical and surgical patients. Follow-up varied from 24 to 48 months. Two reviewers independently evaluated the quality of the 7 identified trials and then extracted the data. Three trials were excluded: One because it is still in progress and without data; one because of randomization flaws; and one because the methodology was unclear and was suspicious for flaws.
OUTCOMES MEASURED: The primary outcomes were stroke or death in the CEA perioperative period (from the time of randomization till 30 days after) and stroke or death during long-term follow-up. TIA was not included as an outcome. Differences in the severity of stroke outcome and quality of life could not be assessed.
RESULTS: The CEA patients incurred more stroke or death in the perioperative period (event rate=3.1% vs 0.4% in medical patients; relative risk=6.5; 95% confidence interval, 2.6-16.0), yielding a number needed to harm of 37. For every 37 patients who underwent CEA, one had a stroke or died during the perioperative period who would not have died if medically managed. However at 3-year follow-up, CEA patients did slightly better in terms of stroke and death than medical management: The number needed to treat to prevent one ipsilateral stroke or perioperative death during 3 years was 53; to prevent any stroke or perioperative death, 43; and to prevent any stroke or any death, 33. However, none of these results achieved statistical significance. When trials comparing surgery to medical management with aspirin were removed from analysis statistical significance was reached, but clinical impact was essentially unchanged.
When compared with watchful medical management, carotid endarterectomy incurs clear perioperative harms yet yields only a small long-term benefit (a slightly reduced risk of stroke and death during 3 years). Furthermore, aspirin had to be factored out to reach statistically significant differences, yet we would routinely give aspirin as part of medical management. If your local surgeons have complication rates significantly lower than 3.1%, it might make intervention more favorable. It is also unclear whether the degree of asymptomatic stenosis would change outcomes. But for the average patient give aspirin, and advise against endarterectomy for asymptomatic carotid stenosis.