The aim of the study was to determine if total serum cholesterol (and its fractions) is related to the severity of angiographically defined extracranial carotid stenosis in patients with prior stroke or transient ischemic attack (TIA). We analyzed the data from 1,614 patients who underwent conventional cerebral arteriography and had serum cholesterol levels measured at entry to the North American Symptomatic Carotid Endarterectomy Trial (NASCET). Patients were classified into one of three ordered categories, according to their linear measurement of severity of extracranial carotid stenosis: mild (<30%), moderate (30–69%), and severe (70–99%). Using ordinal regression analysis, the level of total cholesterol was observed to be significantly related to the severity of carotid stenosis (adjusted OR = 1.17/2 mmol/1; 95% CI: 1.01–1.37). The only other variable that remained statistically significant after adjusting for all other potential confounders was the degree of contralateral stenosis. Our study demonstrates a relationship between the severity of angiographically defined carotid stenosis and total serum cholesterol level in symptomatic patients in whom their presenting stroke or TIA was related to their carotid stenosis. Noting that reductions in total serum cholesterol levels decrease the risk of future coronary events in cardiac patients, further study is required to determine if an analogous response occurs in patients with stroke or TIA. Nevertheless, it seems prudent at the present time to have patients with hypercholesterolemia receive appropriate medical attention for their condition.