From 1990 to 1997, June, 296 patients (156 males and 140 females), aged 16 to 45 years, admitted in the Neurology Department of the University Hospital of Nancy (F) for ischemic stroke, were prospectively evaluated according to a standardized analysis of anamnestic and clinical data, angiography (90 p. 100 of cases), TEE (78 p. 100), hemostasis. Women were younger (mean age = 34.82 y) than men (36.87 y; p = 0.003), with a peak in the 4th decade. Clinical event was a TIA in 14.2 p. 100, a stroke in 51.7 p. 100; it concerned the anterior circulation in 64.5 p. 100, posterior circulation in 25 p. 100, multiple territories in 10.5 p. 100. History of TIA, cervical-cranial pain or Horner syndrome suggestive of dissection, pregnancy or post-partum were found respectively in 60 (20.3 p. 100), 34 (11.1 p. 100) and 13 (9.3 p. 100) cases. Risk factors concerned 87.2 p. 100 of patients, mainly smoking (55.1 p. 100), oral contraceptive (53 p. 100), hyperlipemia (35 p. 100), and were more frequent in case of atheroma and lacunar stroke (p < 0.0000). Etiology, according to TOAST classification, was: atheroma (8.4 p. 100), cardioembolism (8.7 p. 100), small-artery disease (7.1 p. 100), dissection (15.5 p. 100), other determined causes (11.1 p. 100), multiple causes (5.7 p. 100), undetermined cause (34.8 p. 100). Septal pathology was found 34 times. Patients whose stroke remained unexplained were younger (33.7 y vs 37.7, p = 0.002), had less risk factors (p < 0.0000), had more TIA (p = 0.005), more often in the carotid territory (p = 0.008), had a better prognosis (p = 0.01), and showed more often emboli at angiography (p = 0.001). During a mean follow-up of 33 months (median = 19), 21 recurrent strokes occurred and 6 patients died. 134 (46 p. 100) patients had no sequelae, 101 (34.7 p. 100) minor disability, 42 (14.4 p. 100) major sequelae. These results, compared to the main studies of the literature, suggest the interest of common definition criteria and classification of etiologies. In practice, hierarchisation of investigations may be proposed, and vascular risk factors should be tracked in young patients. In patients whose stroke remains unexplained, further studies, as atrial vulnerability, are needed.