Patients presenting with symptoms suggestive of amaurosis fugax, or with findings of Hollenhorst plaques on fundoscopy are frequently referred for duplex evaluation to detect possible carotid artery disease. To better determine the reliability of monocular visual loss and the presence of Hollenhorst plaques for predicting the presence or significance of carotid artery stenosis, we prospectively studied 66 patients with these ocular signs and symptoms. After evaluation, the patients were categorized as follows: 34 of 66 (52%) patients had amaurosis fugax, 23 (35%) had asymptomatic Hollenhorst plaques, 7 (11%) had retinal artery occlusion, and 2 (3%) had venous stasis retinopathy. All patients were evaluated ophthalmologically, with carotid duplex scanning and spectral analysis. A stenosis of greater than 60% was regarded as significant. The presence of risk factors including hypertension, diabetes, a history of CVA or TIA's, tobacco use and hyperlipidemia was recorded. There were no statistically significant differences (p greater than 0.05) in the incidence of atherosclerotic risk factors between the four groups. Patients with amaurosis fugax were more likely to have a significant carotid artery stenosis than those with asymptomatic Hollenhorst plaques or retinal artery occlusion (53% vs 9% vs 0% respectively) (p less than 0.006). We conclude that routine carotid duplex scanning is indicated in all patients with amaurosis fugax in view of the frequent association with significant carotid stenosis (53%). However, the presence of Hollenhorst plaques in the absence of visual symptoms appears not to have a significant association with carotid disease and may not necessarily require routine screening unless other risk factors for carotid stenosis are present.(ABSTRACT TRUNCATED AT 250 WORDS)