Objective: To identify risk factors for operative stroke and death from carotid endarterectomy.
Design: Systematic review of all studies published since 1980 which related risk of stroke and death to various preoperative clinical and angiographic characteristics, including unpublished data on 1729 patients from the European carotid surgery trial.
Main outcome measure: Operative risk of stroke and death.
Results: Thirty six published studies fulfilled our criteria. The effect of 14 potential risk factors was examined. The odds of stroke and death were decreased in patients with ocular ischaemia alone (amaurosis fugax or retinal artery occlusion) compared with those with cerebral transient ischaemic attack or stroke (seven studies; odds ratio 0.49; 95% confidence interval 0.37 to 0.66; P < 0.00001). The odds were increased in women (seven studies; 1.44; 1.14 to 1.83; P < 0.005), subjects aged > or = 75 years (10 studies: 1.36; 1.09 to 1.71; P < 0.01), and with systolic blood pressure > 180 mm Hg (four studies; 1.82; 1.37 to 2.41; P < 0.0001), peripheral vascular disease (one study; 2.19; 1.40 to 3.60; P < 0.0005), occlusion of the contralateral internal carotid artery (14 studies; 1.91; 1.35 to 2.69; P < 0.0001), stenosis of the ipsilateral internal carotid siphon (five studies; 1.56; 1.03 to 2.36; P = 0.02), and stenosis of the ipsilateral external carotid artery (one study; 1.61; 1.05 to 2.47; P = 0.03). Operative risk was not significantly related to presentation with cerebral transient ischaemic attack versus stroke, diabetes, angina, recent myocardial infarction, current cigarette smoking, or plaque surface irregularity at angiography. Multiple regression analysis of data from the European carotid surgery trial identified cerebral versus ocular events at presentation, female sex, systolic hypertension, and peripheral vascular disease as independent risk factors.
Conclusions: The risk of stroke and death from carotid endarterectomy is related to several clinical and angiographic characteristics. These observations may help clinicians to estimate operative risks for individual patients and will also facilitate more meaningful comparison of the operative risks of different surgeons or at different institutions by allowing some adjustment for differences in case mix.