Last month, the American Academy of Neurology released the evidence-based guideline “Stroke Prevention in Patients with Nonvalvular Atrial Fibrillation” (http://www, neurology, org/content/82/8/716, full), which is an update of the Academy’s 1998 guideline. The guideline focuses on two questions: 1) For patients with cryptogenic stroke, what proportion of patients with nonvalvular disease atrial fibrillation (NVAF) can each be detected by the various different techniques? and 2) For patients with NVAF, what antithrombotic regimen reduces the risk and severity of stroke compared with no treatment or other treatments while minimizing the risk of bleeding? For patients with cryptogenic stroke, what percentage of patients with NVAF are identified by each of the different techniques? In patients with recent cryptogenic stroke, the odds of detecting NVAF with cardiac rhythm monitoring range from 0% to 23%. The detection rate may be related to the duration of monitoring. In patients with NVAF, what antithrombotic regimen reduces the risk and severity of stroke compared with no treatment or other treatments while minimizing the risk of bleeding? Effect of Warfarin INR Levels: In patients with NVAF, anticoagulation intensities of INR 2, 0-3, and 0 were associated with reduced incidence and severity of ischemic stroke compared with lower INR levels. Other antithrombotic drugs compared with warfarin or its derivatives: In patients with NVAF, dabigatran may be more effective in reducing the risk of stroke or systemic embolism compared with warfarin (150 mg twice daily, relative risk ratio [RR] 0, 66; RRR 34%). The risk of bleeding was similar, but the application of dabigatran 150 mg twice daily was associated with a low risk of intracranial hemorrhage (RR 0, 40) and a high risk of gastrointestinal bleeding (1, 51% and 1, 02% per year, respectively). In NVAF patients at high risk of cerebral or systemic embolism, rivaroxaban was similarly effective as warfarin in preventing embolism and there was no difference in the risk of major bleeding episodes, however, the incidence of intracranial and fatal bleeding was lower with rivaroxaban than with warfarin (RRR 22%). Apixaban 5 mg twice daily may be more effective than warfarin in NVAF patients at intermediate risk of embolization (RRR 20, 3%). The superiority of apixaban was limited to a decreased risk of hemorrhage (including intracranial hemorrhage) and decreased mortality, and it was not superior to warfarin in reducing the risk of cerebral embolism and systemic embolism. For stroke prevention in patients at risk of stroke with NVAF, oral anticoagulation may be more effective than clopidogrel plus aspirin. However, the risk of intracranial hemorrhage is also higher with the former. In patients at intermediate risk of stroke with NVAF, triflusal (an antiplatelet drug, similar in structure to aspirin, used in Europe, Latin America, and Southeast Asia) plus acenocoumarol (a coumarin derivative, used mainly in European countries) with reduced anticoagulant intensity (INR target value of 1,25-2,0) might be more effective than Conventional anticoagulation intensity (INR target value 2, 0-3, 0) with acenocoumarol alone is more effective (RRR 61%) for stroke risk reduction (only one study and smaller than recent novel oral anticoagulants) In patients with NVAF, low-dose aspirin plus vitamin K antagonist therapy may increase the risk of bleeding complications. There is insufficient evidence that the combination of aspirin and vitamin K antagonists reduces the risk of ischemic stroke or other thromboembolic events. Other Antithrombotic Drugs Compared With Aspirin: In patients at moderate risk of embolism with NVAF and not suitable for warfarin, the effect of apixaban 5 mg twice daily in reducing the risk of stroke or systemic embolism may be due to aspirin (RRR 55, 1%), whereas the risk of bleeding is similar. In patients with NVAF who are not candidates for vitamin K antagonists, the combination of clopidogrel and aspirin reduces the risk of serious vascular events, especially stroke, compared with aspirin alone (RR 0, 72), but increases the risk of serious bleeding (RR 1, 57), including intracranial hemorrhage (RR 1, 87). Anticoagulation in special populations: Anticoagulation may also be effective in elderly patients and in patients with CKD. there is an increased risk of bleeding in patients with CKD on warfarin. RECOMMENDATIONS: Detection of patients with cryptogenic NVAF: A1. For patients with cryptogenic stroke in whom no NVAF is detected, clinicians can perform cardiac rhythm testing on an outpatient basis to detect patients with cryptogenic NVAF (Class C). A2. For patients with cryptogenic stroke without NVAF, clinicians may perform a prolonged cardiac rhythm study (eg, 1 week or several weeks) to identify patients with occult NVAF (grade C). Patient selection for antithrombotic therapy: B1. Clinicians should inform patients with NVAF of their increased risk of stroke and that this risk can be reduced by antithrombotic drugs. Patients should also be informed that the use of antithrombotic drugs increases the risk of severe bleeding (Grade B). B2: Clinicians should inform patients with NVAF to consider antithrombotic medications only when the benefit of a decreased risk of stroke outweighs the harm of the risk of severe bleeding (Grade B). B3: In patients with NVAF who have a history of TIA or stroke, clinicians should routinely administer anticoagulation to reduce the risk of subsequent ischemic stroke (Grade B). B4: In patients with NVAF who have no additional risk factors (“isolated” NVAF patients), clinicians should not administer anticoagulation. Clinicians may or may not use aspirin for antithrombotic therapy (Grade C). B5. To determine which patients with NVAF may benefit from anticoagulation, clinicians should apply a risk stratification scheme to identify patients with NVAF who are at high risk of stroke or who are not at clinically significant risk (level B). Selection of Specific Anticoagulants: C1. To reduce the risk of stroke or recurrent stroke in patients with NVAF, clinicians should choose one of the following regimens (level B): warfarin, target INR 2, 0-3, 0 dabigatran 150 mg twice daily (if CrCl >30 mL/min) rivaroxaban 15 mg/d (if CrCl 30-49 mL/min) or 20 mg/d apixaban 5 mg twice daily (if serum creatinine <1, 5 mg/dL) or 2, 5 mg twice daily (if serum creatinine 1, 5-2, 5 mg/dL and weight <60 kg or age at least 80 years [or both]) trifluromelane 600 mg plus vinpocetine coumadinol with a target INR of 1, 25-2, 0 (patients with intermediate risk of stroke, most often in Patients already taking warfarin: C2, Clinicians may recommend that patients already taking warfarin who are well controlled continue warfarin therapy rather than switching to a newer anticoagulant (Class C). Risk of Intracranial Bleeding: C3. Clinicians should recommend dabigatran, rivaroxaban, or apixaban for NVAF patients requiring anticoagulants when the risk of intracranial bleeding is high (Grade B). Gastrointestinal Bleeding Risk: C4. Clinicians may give apixaban (Level C) when anticoagulants need to be applied to NVAF patients at high risk of gastrointestinal bleeding. Other Factors Affecting Novel Oral Anticoagulants: C5. Clinicians should give dabigatran, rivaroxaban, or apixaban in patients who are unwilling or unable to monitor their INR levels on frequent cycles (Level B). C6. For patients who are unsuitable or unwilling to apply warfarin, clinicians should give apixaban (Level B). C7. If apixaban is not available, clinicians may give dabigatran or rivaroxaban (Grade C). C8. If anticoagulants are not available, clinicians may give a combination of aspirin and clopidogrel (Grade C). C9, If trifluralin is available and the patient is unable or unwilling to take a new oral anticoagulant (mainly in developing countries), clinicians should give vinpocetine (target INR 1,25-2,0) and trifluralin if the risk of bleeding is high in patients with NVAF who are at intermediate risk of stroke (Grade B). SPECIAL POPULATIONS: D1, In elderly patients (age >75 years) with NVAF without recent spontaneous bleeding or intracranial hemorrhage, clinicians should routinely administer oral anticoagulants (Level B). D2. For NVAF patients with dementia or occasional falls, clinicians may give oral anticoagulants. However, in patients with NVAF who have moderate-to-severe dementia or frequent falls, clinicians should explain to the patient or family that the risk-benefit ratio of oral anticoagulation is uncertain (Grade B). D3. Because the risk-benefit ratio of oral anticoagulation in NVAF patients with end-stage renal disease is unknown, there is insufficient evidence to make a recommendation for clinical practice (Grade U).