Patients who have recently had an ischemic stroke with no clear cause might have undetected atrial fibrillation (AF) that isn't caught during their initial hospital stay. After discharge, these patients are typically monitored for AF using devices like Holter monitors or implantable loop recorders. Treatment options during this period include anticoagulants or aspirin. Anticoagulants are more effective in preventing recurrent strokes if AF is present, offering an 80% risk reduction compared to aspirin's 20%. If AF is detected, anticoagulant treatment continues; if not, patients may switch to aspirin after 6-12 months. Despite the clinical rationale for using anticoagulants during this search period, their benefit-risk ratio compared to aspirin has not been fully evaluated.

Patients with recent ischemic stroke with an unknown cause may have a clinical, radiological and echocardiographic pattern of covert atrial fibrillation which is yet undetected by history, ECG monitoring or telemetry during the initial hospital stay. These patients are usually discharged with a prescription of detection of atrial fibrillation in the next weeks by either Holter ECG monitoring, or loop recorder such as wearing device or implantable device. Current standard treatment for these patients is either using anticoagulant treatment during the time of atrial fibrillation search or simply aspirin with the risk that the patient has an undetected paroxysmal atrial fibrillation with a high risk of recurrent stroke that is less prevented by aspirin (20% RRR) than by anticoagulant (80% RRR). Anticoagulant treatment in this situation is clinically justified by the fact that if the patient actually has an atrial fibrillation, an anticoagulant treatment during the search period of AF will protect the patient from a recurrent stroke, rather than treating the patient after the discovery of AF with the risk that an AF-related recurrent stroke occurs in-between. Anticoagulant treatment is then continued if atrial fibrillation is demonstrated by monitory devices, or is stopped with a switch to aspirin after 6-12 months of negative AF search, since investigators know, after NAVIGATE-ESUS and RESPECT-ESUS trials, that long-term aspirin is safer and as effective as anticoagulant in case of cryptogenic stroke.

Although anticoagulant treatment during the search period for AF makes clinically sense, and is widely used in clinical practice, its benefit/risk ratio compared to aspirin therapy has not been evaluated

Inclusion Criteria:

Included patients must fulfill the following 4 criteria:

1. patient aged ≥65 years with:

   1. recent (<15 days) cerebral infarction
   2. with cerebral ischemia proven on MRI or head-CT

2. with no known atrial fibrillation before stroke and no atrial fibrillation detected during hospital stay (monitoring or telemetry) and no mural thrombus.

3. but with suspected atrial fibrillation:

   1. multiple territorial (i.e., in the territory of a cerebral artery or one of its branches) cerebral infarctions in several arterial territories involving both hemispheres, or in the same hemisphere, or in both anterior and posterior circulation, symptomatic or not
   2. or a single cerebral infarction and systemic emboli (e.g., renal, splenic, hepatic or mesenteric infarction, peripheral emboli in arm or leg), symptomatic or not
   3. or any ischemic stroke with dilation of atrium (>34 mL/m²) or left atrial spontaneous echocardiographic contrast or LAA velocities < 40 cm/sec or pro BNP > 400 pg/mL or left ventricular ejection fraction (LVEF) < 40% or supraventricular extrasystole ≥ 400/24 h or longest "atrial run" ≥ 20 beats on telemetry
   4. or age ≥80 year-old and a single infarction

4. and a plan to detect atrial fibrillation with, long term Holter ECG, wearing device or implantable loop recorder

5. with a Rankin score equal or less than 4

6. patient has signed an informed consent

7. Patient is affiliated to a social security.

Exclusion Criteria:

1. Patients with a known cause of stroke, using ASCOD classification A1, C1, S1, O1, D1
2. Symptomatic brain hemorrhage (the mere presence of microbleeds on gradient echo imaging is not an exclusion criteria)
3. Uncontrolled hypertension (following the judgment of the investigator)
4. Clear indication to anticoagulant or antiplatelet therapy
5. Contra-indication to anticoagulant or antiplatelet therapy
6. Intercurrent disease that may interfere with evaluation of the primary end-point or that may prevent follow-up study visits
7. Participation in another interventional clinical trial.
8. Under contraception in case of childbearing potential
9. Patient under guardianship or curatorship

Outcomes will be evaluated blindly by the members of the critical event committee. They will receive medical files without the patient's name, surname, phone number, or address.