1. The recent development of novel oral anticoagulants (NOACs), including direct thrombin inhibitor (dabigatran) and factor Xa inhibitors (rivaroxaban, apixaban, and edoxaban), could potentially overcome many drawbacks of warfarin, and might provide a safer, and even more effective and convenient alternative approach to warfarin in non-valvular atrial fibrillation (NVAF), especially in Asians.
2. According to the results of a meta-analysis comparing Asians and non-Asians, NOACs are preferentially indicated in Asians in terms of both efficacy and safety.
3. There is no randomized controlled trial with sufficient power to directly compare the efficacy and safety among NOACs in NVAF, not to speak of Asians and Chinese.
4. Indirect comparisons are only based on observation with a lot of limitations such as heterogeneous background characteristics, difference in study design, and diversity in time within therapeutic range in control group. The findings from indirect comparisons are not conclusive but only hypothesis-generating.
5. This investigator-initiated prospective randomized open blinded end-point clinical trial will directly compare the efficacy and safety among 3 NOACs in patients with NVAF in Taiwan. We hypothesize that rivaroxaban or apixaban is non-inferior to dabigatran in terms of the efficacy.

1. participants

   a. eligible participants are randomly assigned to dabigatran, rivaroxaban, or apixaban with allocation ratio of 1:1:1

   • Patients are randomly assigned to receive dabigatran (110 or 150 mg twice daily), rivaroxaban (15 or 20 mg daily), or apixaban (5 mg twice daily) with dosage and frequency approved by the Ministry of Health and Welfare, Taiwan. Reduced doses (dabigatran 110 mg twice daily, rivaroxaban 10 or 15 mg daily, or apixaban 2.5 mg twice daily) are allowed in a subset of patients with one or more of the following criteria: an age of at least 80 years, a body weight of no more than 60 kg, a serum creatinine level ≥1.5 mg per deciliter (133 μmol per liter) or creatinine clearance around 30 to 49 ml per minute)

2. blood sampling, genotyping, and measurement of biomarkers

   a. bood samples (13 mL) from peripheral veins in all study subjects at baseline and 10 mL 3 months later, and stored for enzyme-linked immunosorbent assay as well as genotyping

3. outcome follow-up a. clinical follow-up is performed and clinical outcomes are obtained by clinic visit, telephone call or direct contact with participants or subjects' family quarterly after treatment for 2 times, then every 6 months

Inclusion Criteria: Known AF (paroxysmal or persistent/ permanent) who are suitable and ready for NOAC treatment plus at least one of the following criteria

• Prior ischemic stroke, transient ischemic accident or systemic embolism
• Left ventricular ejection fraction ≤40% (documented by echocardiography or contrast ventriculography)
• Symptomatic congestive heart failure (≥ New York Heart Association Functional Class 2) within 6 months before screening
• Age ≥75 years
• Age ≥65 but <75 years with diabetes mellitus, hypertension or coronary artery disease

Exclusion Criteria: Subjects are excluded if they have at least one of the following situations before screening:

• Known severe (i.e. hemodynamically significant) mitral stenosis regardless of having received operation
• Time elapsed from the onset of stroke ≤7 days
• Bleeding tendency
• Creatinine clearance rate ≤30 mL/min
• Known active liver disease (persistent elevation of alanine aminotransferase, aspartate transaminase or alkaline phosphatase ≥3 × upper normal limit; or advanced liver cirrhosis ≥Pugh B)
• Pregnancy
• Recent documented active malignancy or radiation therapy (≤6 months) and not expected to survive 3 years
• Unwilling to give informed consent
• Conditions other than AF that required anticoagulation
• Anemia (hemoglobin level <90 g/L) or thrombocytopenia (platelet count <100 × 109/L)
• Persistent uncontrolled hypertension (systolic blood pressure >180 mmHg or diastolic blood pressure >100 mmHg)
• Active infective endocarditis
• Patients considered unreliable by the investigator or have a life expectancy less than the expected duration of the trial because of concomitant disease, or has any condition which in the opinion of the investigator, would not allow safe participation in the study