Thromboembolic events after carotid endarterectomy are not prevented by aspirin, but are due to the platelet response to adenosine 5′‐diphosphate

Background: Aspirin therapy fails to prevent a number of postprocedural thrombotic events, yet it still remains the standard antiplatelet regimen in most vascular surgical centres. After carotid endarterectomy (CEA), thrombosis of the endarterectomized vessel is preceded by increasing numbers of microemboli that can be detected with transcranial Doppler (TCD). The number and rate of emboli is highly predictive of thrombotic stroke. It was hypothesized that a preoperative test of platelet function might identify the mechanism(s) underlying post‐CEA thrombosis. Methods: Blood was taken from 120 patients using a standardized phlebotomy technique. Platelet fibrinogen binding was measured by whole blood flow cytometry, in unstimulated samples, and in response to adenosine 5′‐diphosphate (ADP) (10 −5 −10 −7 mol l −1 ) and thrombin (0·02–0·16 units ml −1 ). Platelet aggregation was measured using ADP (4–20 × 10 −7 mol l −1 ). The ability of aspirin to inhibit platelets was assessed by the aggregation induced by arachidonic acid. For the first 3 h after operation, the number of emboli occurring was quantified using TCD. Results: Of the 120 patients studied, 110 were monitored by TCD. These were divided into patients with more than 25 postoperative emboli ( n = 22) and those with fewer than 25 emboli ( n = 88). The degree of platelet inhibition induced by aspirin was not significantly different between the two groups ( P = 0·89). However, platelets from the group with high rates of embolization bound 58 per cent more fibrinogen on flow cytometry in response to stimulation with a physiological dose of ADP (10 −7 mol l −1 ) ( P = 0·006). Aggregation of platelets from this group was also increased in response to ADP (35 per cent) relative to the group with few emboli ( P = 0·001). ADP also induced more rapid aggregation in the patients with more than 25 emboli ( P = 0·04). There was no difference in the activity of resting platelets ( P = 0·4) or platelets stimulated by thrombin ( P = 0·43), between the two groups of patients. Conclusion: These data suggest that it is the platelet response to ADP which is important in arterial thrombotic complications rather than products of the cyclo‐oxygenase pathway. This observation could have significant therapeutic implications for other vascular or interventional procedures in which the endothelium is disrupted. © 2000 British Journal of Surgery Society Ltd