Atherosclerosis is a common finding in the vertebrobasilar junction and in the basilar artery. Several theories try to link the process of atherogenesis with the forces exerted by the flowing blood. An attractive relation has been found between the locations in vessels at which atherosclerotic plaques are often present and the locations in models where complicated flow patterns exist. Most of the studies provided data on bifurcations. Finding a similar relation in an arterial confluence would certainly add to the credibility of the (causal) relationship between hemodynamics and atherosclerosis. Further support can be provided if variations of the geometry result in changes of the location of the atherosclerotic lesions, corresponding to the changes of the flow force distribution. In our previous numerical and experimental work, the influence of geometric and hemodynamic parameters, such as asymmetrical inflow, confluence angle, and blunting of the apex, on the flow in vertebrobasilar junction models has been investigated in detail. Recirculation areas and distribution of the wall shear stress have been computed. In this anatomic study, the effect of modulation of these geometric and hemodynamic parameters on the flow pattern is compared with the size and location of plaques in human vertebrobasilar junctions and basilar arteries. In addition, a comparison is made between the preferential areas of atherosclerotic plaques in junctions and bifurcations to demonstrate the localizing role of hemodynamics in atherogenesis. The apex of the vertebrobasilar junction and the lateral walls of the basilar artery appeared to be prone to atherosclerosis. In 43 of 85 vertebrobasilar junctions, a plaque was found at the apex. Furthermore, the summed plaque thickness at both lateral walls differs significantly (paired t test, P=.03) from that at the walls facing the pons and the skull base. In contrast, several authors found that the lateral walls of the mother vessel and the apex in bifurcations are often spared. Modulation of the various parameters in the models changed the size of the regions with low wall shear stress and/or recirculation areas dramatically. A comparable effect was found in the occurrence of plaques in the human vertebrobasilar junction; eg, for an atherosclerotic plaque at the apex, a predicted probability larger than 0.5 was computed for blunted apexes and for sharp-edged apexes with a confluence angle exceeding 90 degrees. Apparently, two geometric risk factors for an atherosclerotic plaque at the apex can be distinguished: a blunted apex and a large confluence angle.