There is controversy over the choice of anaesthesia for carotid surgery. The aim of this retrospective study was to report the early results of carotid surgery performed with loco-regional anaesthesia by blockade of the cervical plexus. From 1987 to 1992, 405 consecutive operations on the carotid bifurcation were performed using this technique. Among these patients, 202 (50%) were asymptomatic, the indication being carotid narrowing of 80% or more, while clinical signs were observed in the other 203 patients (50%). According to the Sundt classification, 360 patients (89%) had a medical risk and 45 (11%) had a neurological risk. Occlusion of the controllateral internal carotid was present in 47 patients (12%). The deep blocade affected roots C2-C3-C4 followed by a superficial block using 0.5% Bupivacaine. An endarteriectomy was performed in 96% of the cases. In 10 (2.5%), there were complications or the cervical plexus blocade was insufficient: 6 caused the operation to be postponed and/or use of general anaesthesia, but none of these led to postoperative complication. In 35 patients (8.6%) neurological events during clamping required establishing a shunt (6.1% and 27.6% in cases with permeable or occluded controlateral carotids respectively). Neurological recovery was rapid and complete except in 4 cases. There were 8 central neurological complications which persisted or appeared postoperatively: 5 regressive ischaemic events, 2 persistent ischaemic events and one which led to the patient's death. No cardiac complications were seen. We conclude that blocade of the cervical plexus is a simple and effective technique for surgery of the carotid bifurcation. With this method, detection of clamp intolerance is more reliable and it gives enough time for endarteriectomy. The risk of coronarian complications is low due to good haemodynamic stability. This method is a low-cost technique and is better adapted to understanding the mechanisms of neurological complications.