The purpose of this study was to determine optimal velocities for detecting ≥50% and ≥80% restenosis prior to considering carotid intervention/carotid artery stenting (CAS) after carotid endarterectomy (CEA) with patching in symptomatic and asymptomatic patients. Two hundred CEA patients with 195 pairs of imaging (duplex ultrasound versus computed tomography angiography [CTA]/carotid arteriography) were analyzed. Peak systolic velocities (PSVs), end diastolic velocity (EDV) and internal carotid artery/common carotid artery (ICA/CCA) ratios were correlated to angiography. Receiver operator characteristic (ROC) curves determined optimal velocity criteria in detecting ≥50% and ≥80% restenosis. The mean PSVs for ≥50% and ≥80% restenosis were 248 and 404 c/s, respectively (P < 0.001). A PSV of ≥213 c/s was optimal for ≥50% restenosis with sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and overall accuracy (OA) of 99%, 100%, 100%, 98% and 99%, respectively. An ICA PSV of 274 c/s was optimal for ≥80% restenosis with sensitivity, specificity, PPV, NPV and OA of 100%, 91%, 99%, 100% and 99%, respectively. ROC analysis showed that PSVs were significantly better than EDVs and ICA/CCA ratios in detecting ≥50% restenosis. Standard duplex velocity criteria should be revised after CEA using patching. Specific carotid duplex velocities can be used to detect ≥50% and ≥80% restenosis after CEA with patch closure prior to carotid intervention/CAS.