The accuracy of ophthalmodynamometry (ODM) was retrospectively analyzed in 36 patients who had ophthalmic examinations prior to carotid arteriography. In 75% of those patients, ODM correctly predicted the presence of absence of significant stenosis (ie, 50% or greater) or occlusion of the internal carotid artery. A 30% false-negative rate and 19% false-positive rate was found in this study. The accuracy was not improved when the examiner had more experience with the test. The accuracy of ODM with staff members who regularly perform the examination was 68% as compared to the ophthalmology residents' accuracy of 76%. Causes of false-negative and positive results include subjectivity of the examination, stenosis of arteries other than the internal carotids, and development of collateral circulation as was demonstrated arteriographically in 2 of our patients.