Introduction. In order to select patients for carotid angioplasty, some studies have set criteria for defining high-risk groups. Aim. To analyse and compare the outcomes of carotid endarterectomies (CEA) according to risk criteria. Patients and methods. We conducted a retrospective review of 669 CEA carried out between 1991 and 2004. Risk was differentiated according to the NASCET and SAPPHIRE criteria. Group A (high risk) was made up of 184 (27.5%) CEAs: 32 patients over the age of 80, 81 contralateral occlusions, eight with severe lung disease -forced expiratory volume in the first second (FEV1) < 1 L-; six with unfavourable anatomical factors (restenosis, radiotherapy or cervical surgery); nine with chronic renal failure (plasma creatinine > 3 mg/dL); 61 with severe heart disease (unstable angina, recent infarction, class III/IV heart failure, atrial fibrillation or combined heart surgery). Group B (absence of criteria) was made up of 485 (72.5%) CEAs. Perioperative mortality and cardiac and neurological morbidity were analysed. The statistical analysis was performed by means of the chi-squared test. Results. For groups A and B, neurological morbidity rates were 5.9% and 2.4% -p = 0.02; relative risk (RR): 2.41; 95% confidence interval (CI 95%): 1.08-5.37-; cardiac morbidity rate 1.6 and 0.4% (NS) and mortality rate 0.5 and 0.6% (NS), respectively. Overall morbidity and mortality (stroke, myocardial infarct or death) for group A was 8.1% and for group B it was 3.5% (p = 0.013; RR: 2.32; CI 95%: 1,18-4,56). Of all the criteria that were used, there was a correlation between contralateral occlusion and increased neurological morbidity (p = 0.001; RR: 4.66; CI 95%: 2,08-10,41). Conclusion. Our series confirms that contralateral occlusion increases the risk of a stroke.