Optimal surgical strategy in patients with concomitant coronary and carotid artery disease is debatable. We have analysed 15-years of experience (January 1981-August 1996) with 195 consecutive patients in whom we have used two different surgical approaches. Group A consisted of 48 patients who underwent a single-stage surgical procedure, and group B (147 patents) underwent a two-stage procedure, either as carotid endarterectomy followed by coronary artery bypass surgery (group B 1 , 97 patients), or as coronary artery bypass surgery followed by carotid endarterectomy (group B 2 , 50 patients). Overall, there were 40 (20.5%) patients with left main coronary artery disease, 49 (25.1%) with poor left ventricular function, 128 (65.6%) with previous myocardial infarction, 134 (68.7%) were in New York Health Authority (NYHA) functional class III or IV, and bilateral carotid involvement was present in 57 patients (29.2%). Unstable angina was more frequent in groups A and B 2 (P < 0.0001), NYHA class III-IV was more frequent in group A (versus B 1 , P = 0.001 and versus B 2 , P = 0.02), low ejection fraction (EF) was more frequent in groups A and B 2 (P < 0.0001 for both), bilateral carotid stenosis in groups A and B 1 (P = 0.02 and P = 0.0001, respectively) and ulcerated plaque in group B 1 (versus A, P = 0.0001). These differences dictated the surgical strategy, which resulted in different protocols for clinical and operative management. Results: Early mortality for the entire group was 4.6% (9/195-6.2% in group A, 6.2% in group B 1 and 0% in group B 2 , respectively P > 0.05). Serious morbidity occurred in 7.3% of patients (14195-8.3% in group A, 7.2% in group B 1 and 6% in group B, respectively P > 0.05). Univariate analysis revealed only bilateral carotid stenosis as a predictor of outcome (P = 0.04). Follow-up was completed for 156 patients (80.0%) and averaged 84.1 ± 13.3 months (range 1-180 months). Kaplan-Meier survival estimate for the entire group was 81% and event-free survival was 76% at 5 years. Actuarial and event-free survivals were similar for all groups. Early and late outcome in these patients were influenced more by their preoperative clinical status than by the surgical strategy itself. It is therefore concluded that surgical approach should be individualized for the majority of patients.