In many stroke patients it is not possible to establish the etiology of stroke. However, in the last two decades, the use of transesophageal echocardiography in patients with stroke of uncertain etiology reveals atherosclerotic plaques in the aortic arch, which often protrude into the lumen and have mobile components in a high percentage of cases. Several autopsy series and retrospective studies of cases and controls have shown an association between aortic arch atheroma and arterial embolism, which was later confirmed by prospectively designed studies. The association with ischemic stroke was particularly strong when atheromas were located proximal to the ostium of the left subclavian artery, when the plaque was ≥4 mm thick and particularly when mobile components are present. In these cases, aspirin might not prevent adequately new arterial ischemic events especially stroke. Here we review the evidence of aortic arch atheroma as an independent risk factor for stroke and arterial embolism, including clinical and pathological data on atherosclerosis of the thoracic aorta as an embolic source. In addition, the impact of complex plaques (≥4 mm thick, or with mobile components) on increasing the risk of stroke is also reviewed. In non-randomized retrospective studies anticoagulation was superior to antiplatelet therapy in patients with stroke and aortic arch plaques with mobile components. In a retrospective case-control study, statins significantly reduced the relative risk of new vascular events. However, given the limited data available and its retrospective nature, randomized prospective studies are needed to establish the optimal secondary prevention therapeutic regimens in these high risk patients.