Recommendation 1 – Stratification of patients using a predictive index for stroke risk (2014)
We recommend that all patients with AF or AFL (paroxysmal, persistent or permanent), should be stratified using a predictive index for stroke risk (for example, the “CCS algorithm” based on the CHADS2 model) (Strong Recommendation, High Quality Evidence).

Values and preferences (2014)
Use of a modified version of the CHADS2 schema (the “CCS algorithm”) is recommended to facilitate the choice of appropriate antithrombotic therapy by incorporating the substantial risk of stroke conferred by age 65-74 to the well validated CHADS2 risk stratification scheme. However, it excludes female sex or vascular disease alone for the reasons detailed above.

Recommendation 2 – OAC therapy for patients ≥ 65 years or CHADS2 ≥ 1 (2014)
We recommend that OAC therapy be prescribed for most patients with age ≥ 65 years or CHADS2 ≥ 1 (the “CCS algorithm”) – see Figure 1. (Strong Recommendation, Moderate Quality Evidence) Values and preferences (2014) This recommendation places relatively greater weight on the absolute reduction of stroke risk with OACs compared to aspirin in patients aged >65 or with CHADS2 ≥ 1 and less weight on the increased risk of major hemorrhage with OACs compared to aspirin.

Recommendation 3 - No OAC therapy for patients < 65 years with no CHADS2 risk factors and antiplatelet therapy for those patients with coronary or arterial vascular disease (2018)
For patients with non-valvular AF/AFL aged < 65 years with no CHADS2 risk factors, we suggest no antithrombotic therapy for stroke prevention (Weak Recommendation, Moderate Quality Evidence), with management of their coronary or arterial vascular disease as directed by the 2018 CCS/CAIC Focused Update of the Guidelines for the Use of Antiplatelet Therapy.

Practical tip (2018)
For patients with non-valvular AF/AFL aged < 65 years with no CHADS-65 risk factors, the risk of stroke associated with AF is not sufficiently elevated to justify OAC therapy. For this group treatment should be directed at the underlying coronary/arterial vascular disease (peripheral vascular disease or aortic plaque) as outlined in the 2018 CCS/CAIC Focused Update of the Guidelines for the Use of Antiplatelet Therapy. Therapeutic options include ASA 81 mg daily alone; or in combination with either clopidogrel 75 mg daily, ticagrelor 60 mg bid, or rivaroxaban 2.5 mg bid (when approved).

Recommendation 4 – Most patients should receive NOAC (2014; updated 2018)
We recommend that when OAC-therapy is indicated for patients with non-valvular AF, most patients should receive dabigatran, rivaroxaban, apixaban or edoxaban in preference to warfarin (Strong Recommendation, High Quality Evidence).

Values and preferences (2014)
This recommendation places a relatively high value on the greater ease of use of the NOACs in comparison to warfarin, and the results of large RCTs showing that the NOACs are either non-inferior or superior to warfarin in stroke prevention; the drugs have no more major bleeding or less bleeding vs warfarin and especially less intracranial hemorrhage. The recommendation places less value on the shorter clinical experience, lack of a specific antidote, and lack of a simple test for intensity of anticoagulant effect with the NOACs. The preference for one of the NOACs over warfarin is less marked among patients already receiving warfarin with stable therapeutic INRs, no bleeding complications, and who are not requesting a change in OAC therapy.

Recommendation 5 – Warfarin when mechanical valve, mitral stenosis or renal dysfunction (2014)
We recommend that when OAC is indicated, warfarin be used rather than one of the NOACs for those patients with a mechanical prosthetic valve, those with rheumatic mitral stenosis and those with a CrCl of 15 - 30 mL/min (Strong Recommendation, Moderate Quality Evidence).

Values and preferences (2014)
This recommendation places high value on the evidence from one RCT of the inferiority of dabigatran compared to warfarin for the prevention of thromboemboli in patients with a mechanical prosthetic valve. It places relatively high value on the long experience and clinical reports of the use of warfarin in patients with rheumatic mitral stenosis and patients with CrCl 15 – 30 mL/min and the absence of such information for NOACs.

Recommendation 6 – Patients who refuse OAC should receive ASA plus clopidogrel (2014)
We recommend that patients whose risk of stroke warrants OAC therapy, but who refuse any OAC, should receive ASA 81 mg/ day plus clopidogrel 75 mg/ day (Strong Recommendation, High Quality Evidence).

Values and preferences (2014)
This recommendation places high value on the superiority of the combination of ASA and clopidogrel over ASA alone in the ACTIVE-A trial. However, bleeding risk of combined antiplatelet therapy may not be very different from OAC monotherapy.

Recommendation 7 – Annual renal function assessment (2012)
We recommend that patients with AF who are receiving OAC should have their renal function assessed at least annually by measuring serum creatinine and calculating CrCl (Strong Recommendation, Moderate Quality Evidence) and should be regularly considered for the need for alteration of OAC drug and/or dose changes based on CrCl (Strong Recommendation, Moderate Quality Evidence).

Recommendation 8 – Antithrombotic therapy should relate to CrCl (2012)
For antithrombotic therapy of CKD patients, therapy should relate to CrCl as follows:

CrCl >30 mL/min: We recommend that such patients receive antithrombotic therapy according to their risk as determined by the “CCS algorithm” as detailed in recommendations for patients for patients with normal renal function (Strong Recommendation, High Quality Evidence).

CrCl 15-30 mL/min and not on dialysis: We suggest that such patients receive antithrombotic therapy according to their risk as determined by the “CCS algorithm” as for patients with normal renal function. The preferred agent for these patients is warfarin (Conditional Recommendation, Low Quality Evidence).

CrCl<15mL/min (on dialysis): We suggest that such patients not routinely receive either OAC (Conditional Recommendation, Low Quality Evidence) or ASA for stroke prevention in AF (Conditional Recommendation, Low Quality Evidence).

Values and preferences (2016, updated from 2012)
Recommendation 11 places a relatively higher value on prevention of ischemic stroke than on bleeding complications associated with antithrombotic therapy, as well as the limited data available for new OACs in CKD patients. They also place a relatively higher weight on observational data linking warfarin and ASA use with mortality in patients on dialysis, and relatively lower weight on the potential for these agents to prevent ischemic stroke. Patients on renal dialysis who have atrial fibrillation continue to be at high risk of both stroke and major bleeding complications. This population has been largely excluded from clinical trials evaluating stroke prevention therapies, and there have been no substantial new advances in the management of these individuals. Such studies are being planned, but until they can be completed, clinicians must continue to balance the risks of stroke against the risk of bleeding complications.

Practical tip (2012)
No antithrombotic therapy may be appropriate for some patients with CrCl 15-30 mL/min (not on dialysis), with a stronger preference for avoiding bleeding complications than preventing ischemic stroke.

Practical tip (2016, updated from 2010)
Therapy with OACs or antiplatelet drugs may be appropriate for some patients with AF and CrCl <15 mL/min (on dialysis) in whom there is a stronger preference to avoid ischemic stroke despite uncertain benefit and likely greater bleeding risk.

Recommendation 9 – LAA closure devices to be used only in research and, exceptional cases (2014)
We suggest these non-approved LAA closure devices not be used, except in research protocols or in systematically documented use protocols in patients who are at high risk of stroke (CHADS2 ≥ 2) and yet antithrombotic therapy is precluded (Conditional Recommendation, Low Quality Evidence).

Values and preferences (2014)
This recommendation places relatively great weight on the absence of RCTs showing clear benefit to risk in favor of these devices and on the need for further research and careful case series.

Recommendation 10 – Acute management of stroke patients as per AHA/ASA guidelines (2010)
We recommend that patients with AF or AFL who experience a stroke be managed acutely according to the published guidelines of the American Heart and American Stroke Associations (Strong Recommendation, Moderate Quality Evidence).

Recommendation 11 – Hemorrhage on OAC to be managed per AACP guidelines (2010)
We suggest that patients with AF or AFL who experience hemorrhage while on OAC be managed according to the published practice guidelines of the American College of Chest Physicians (Conditional Recommendation, Low Quality Evidence).

Practical tip (2012)
No antithrombotic therapy may be appropriate for some patients with CrCl 15-30 mL/min (not on dialysis), with a stronger preference for avoiding bleeding complications than preventing ischemic stroke.

Practical tip (2016, updated from 2010)
Therapy with OACs or antiplatelet drugs may be appropriate for some patients with AF and CrCl<15 mL/min (on dialysis) in whom there is a stronger preference to avoid ischemic stroke despite uncertain benefit and likely greater bleeding risk.

Recommendation 12 - Idarucizimab for emergency reversal of dabigatran's anticoagulant effect (2018, updated from 2016)
We recommend administering idarucizimab for emergency reversal of dabigatran's anticoagulant effect in patients with uncontrollable or potentially life-threatening bleeding and/or in patients who require urgent surgery for which normal hemostasis is necessary (Strong Recommendation, Moderate Quality Evidence).

Values and preferences (2016)
This recommendation places relatively greater value on the ability of idarucizimab to reverse coagulation parameters indicative of dabigatran’s effect, its potential to decrease bleeding-related outcomes and risks of urgent surgery, and its safety and tolerability profile, and less value on the absence of a control group in the RE-VERSE AD trial and on the cost of the drug.

Practical tip (2018, updated from 2016)
In acute, life-threatening bleeding situations in which standard resuscitation (such as local measures, transfusion, etc) is anticipated to be insufficient (eg, intracranial hemorrhage), or in situations in which standard resuscitation has not stabilized the patient, idarucizumab 5g IV should be administered as soon as possible. Activated partial thromboplastin time (aPTT) and thrombin time (TT) may be used to qualitatively identify the presence of active dabigatran at baseline in a patient, although they are less sensitive than dilute thrombin time (DTT) and ecarin clotting time (ECT; 92% of patients in the REVERSE-AD trial had an elevated DTT or ECT,whereas only 74% had an elevated aPTT). However, obtaining these measures should not delay the administration of idarucizumab. In many instances of life-threatening bleeding, clinicians have to make a treatment decision on the basis of a history of dabigatran use rather than laboratory evidence. Renal function and timing of the last dose of dabigatran provide key information regarding the likely extent of remaining dabigatran effect.

Practical tip (2018)
"Urgent" surgery as defined in the REVERSE-AD trial is surgery that cannot be delayed beyond 8 hours (amended from 4 hours in the initial version of the protocol). The timing of surgery should be based on the clinical indication and stability of the patient. In instances where delayed surgery is appropriate, clinicians may obtain coagulation parameters (e.g.TT or aPTT) to identify patients who would be unlikely to benefit from idarucizumab. (see Practical Tip 1 above).

Practical tip (2018)
Reversing dabigatran therapy exposes patients to the thrombotic risk of their underlying disease. Oral anticoagulation should be reintroduced as soon as medically appropriate.