Comparisons with intraoperative findings suggest that Duplex scanning may be of value in predicting the morphology of carotid artery stenoses. Such studies, however, are based on the results obtained by experienced investigators. To clarify whether sonographic criteria can be standardized, 6 investigators from different hospitals each documented 30 carotid artery stenoses of greater than or equal to 40% diameter reduction which could be imaged in at least 2 planes. Three sonographic criteria (plaque surface, echo density, echo structure), each with 8 categories, were assessed from the image documentations by the 6 as well as by 2 independent further investigators. Depending on the experience and on the device used 29-81% of all stenoses greater than or equal to 40% examined in the different laboratories could be included in the study. The +/- 1 category concordance between the different investigators averaged 50-60% for all sonographic criteria independent from the degree of stenosis and from the image quality. Because of lack of sufficient reproducibility the subjective assessment of sonographic criteria is found to be not suitable for use in multicentre studies.