Carotid endarterectomy, one of the most frequently performed vascular surgeries, has been associated with significant reductions in rates of ischemic stroke in at least two large randomized controlled trials. Initial results indicated that endarterectomy was beneficial only in patients with stenosis of 70 percent or more. The surgery confers no benefit in patients with 30 to 49 percent stenosis and does more harm than benefit when stenosis is less than 30 percent. Rothwell and colleagues examined subgroups of patients to determine which factors, in addition to the degree of arterial stenosis, would indicate the possibility of benefit from endarterectomy.

They pooled data from the two largest randomized controlled trials, the European Carotid Surgery Trial and the North American Symptomatic Carotid Endarterectomy Trial, which together included 95 percent of patients ever randomized in trials of this surgery. The participants had experienced recent clinical events in the distribution of the carotid artery, and the symptomatic vessel was imaged by selective catheter angiography. Participants were randomized to medical or surgical treatment, and follow-up was conducted by neurologists or by subspecialists in stroke.

Follow-up data were obtained for 5,893 patients over an average of 66 months. Of the 3,157 patients treated with endarterectomy, 222 (7 percent) had operative deaths or strokes. The perioperative risk of death or stroke was increased in women, patients with diabetes, patients with occlusion of the other carotid artery, patients with ulcerated or irregular plaques, and patients who had hemispheric (rather than retinal) events preceding clinical events. Age, sex, and time since last symptomatic event greatly modified the outcome of surgery. When all variables were included in the analysis, the benefits from surgery were greatest in men, patients older than 75 years, and those who had surgery within two weeks of their last ischemic event.

The authors calculate that the number needed to treat (NNT) with carotid endarterectomy to avoid one ipsilateral stroke within five years was nine among men and 36 among women in patients with stenosis of 50 percent or greater. The comparable NNT was five among patients older than 75 years compared with 18 among those younger than 65 years, and five among those randomized within two weeks of an ischemic event compared with 125 among those randomized more than 12 weeks after the event. They stress that additional factors besides the degree of stenosis must be considered when advising patients and families about carotid endarterectomy. Among these factors, delay in surgery appears to be the most significant. Ideally, carotid endarterectomy should be performed within two weeks of the ischemic episode.