This study evaluate the impact of early personalized nutritional management on 1-month mortality after acute stroke. In the randomised centres of the interventional group (early personalized nutritional management), each patient will be assessed by a dietician within 2 days of stroke. Any swallowing problems are screened, and any barriers to eating are identified before dietary support. Patient's food intakes are recorded until the 7th day post-acute to assess their energy intake. Depending on patient's nutritional objectives, a reinforced feeding strategy comprising 2 levels is put in place, in order to prevent malnutrition: in level 1, the dietician tries to meet nutritional requirements orally, while in level 2 he can recourse to artificial nutrition in the event of failure or inability to meet requirements with oral intake alone. This strategy is started immediately after the initial assessment. Intakes are reassessed every 24 to 48 hours by the dietician in order to adapt the nutritional strategy as quickly as possible in order to cover the patient's personnalized nutritional needs. No change in practice was required of the randomised centres in the control group. All patients will be contacted by telephone 30 days after stroke to collect parts of the assessment criteria, while the other criteria will be collected directly from the patient's computerised medical record. Three months after stroke, an evaluation of quality of life and modified Rankin test will be performed. Our hypothesis is that the implementation of an individualized management of early nutritional support, aimed at compensating for the decrease in intake associated with the disease, would reduce mortality at 30 days in patients hospitalised for stroke.